“Each is liable to panic, which is exactly, the terror of ignorance surrendered to the imagination.”
– Ralph Waldo Emerson (1803-1882)

The words of this American poet and philosopher describe in poetic form, the power of the imagination to induce fear and panic in the absence of knowledge and rational thought. Primitive man was terrified by events he did not understand and could not explain – imagine the impact of thunder and lightning on earliest hominids that did not yet have even a rudimentary knowledge of physics. Such events and countless others would understandably have been terrifying – the terror of ignorance surrendered to the imagination.

As modern man has more than ample knowledge of the natural sciences, most individuals do not panic over storms, solar eclipses, or other natural phenomena of which we have the comfort of comprehension. Earthquakes, hurricanes, tornadoes or other “natural disasters” are clearly terrifying for those directly involved, but while terror may be a normal reaction panic is not, for panic is fear of a non-physical imagined threat. The chronic panic attacks that constitute Panic Disorder (PD) are ultimately an acute, irrational fear of the imagined dissolution of the self. The individual who experiences such attacks are not unlike primitive man with an imagination fueled by ignorance. In the case of PD it is not ignorance of the external world of natural science but ignorance of the inner world of cognitions, affects, and somatosensory experiences – the result of arrest at or regression to a primitive level of emotional development. Ignorance in this inner domain fuels imagination, contributes to catastrophic explanations of ambiguous internal events and sensations, and culminates in panic attacks. This lack of internal knowledge and self-awareness constitutes the deficits of alexithymia.

Etiological Theories of PD

Though an isolated panic attack is experienced by 7% to 34% of the general population (Norton, Harrison, Hauch & Rhodes, 1985; Telch, Lucas & Nelson, 1989), only a small percentage of these individuals develop overwhelming fear of another attack or the fear of fear cycle (Goldstein & Chambless, 1978) that is involved in the recurrent and unexpected panic attacks that define PD. Panic attacks may occur in other anxiety disorders but are situationally bound as in simple or social phobias. Barlow (1988) proposes a model for the etiology of PD based on the similarity between the physiological activation of a panic attack and the fight or flight response. A panic attack is a fight or flight response in the absence of real danger, and as such is a false alarm. Due to the powerful impact of a panic attack, an immediate association develops between the triggering stimuli and the intense physiological sensations activated during the attack, an example of classical conditioning. For individuals who develop PD, whenever a stimulus generates the sensations associated with the first attack, a false alarm of danger is sounded and a panic attack is triggered. When the false alarm becomes associated with the internal cues or sensations of fear, a learned alarm develops. Future panic attacks can then result from the learned alarm of those physiological sensations. Barlow believes that individuals with PD have an inborn susceptibility for somatic preoccupation and over reactivity to physiological cues.

PD is but one of several disorders characterized by prominent somatic features which, according to Barsky (1992), are exacerbated by somatosensory amplification or the “tendency to experience somatic and visceral sensation as intense, noxious, and disturbing” (p. 28). He describes amplification as the tendency for hypervigilance of bodily sensations, especially infrequent sensations, and the tendency to interpret them as pathological. All physiological sensations can be amplified including normal reactions such as increases in heart rate; minor dysfunctions such as an eye twitch or autonomic reactions to emotional arousal; or the symptoms of more serious medical diseases. He believes amplification may be a transitional state or an enduring trait acquired in childhood or even “hard wired” in from birth (p. 29).

According to Clark (1988), “catastrophic misinterpretation of certain bodily sensations is a necessary condition for the production of a panic attack” (p. 84). By this, he means there is a “cognitive disturbance in panic patients” (1986, p. 469) that causes faulty processing of bodily sensations and the belief that these sensations are more serious than they actually are. This can manifest in mistaking rapid pulse or respiration as an indication of a pending heart attack, or equating breathlessness with suffocation and imminent death. This faulty cognitive process leads to fear of the bodily sensations of fear, or a fear of fear cycle wherein affect becomes dysregulated.


Taylor (1992) and Taylor, Bagby and Parker (1997) consider PD as one of several disorders of affect regulation, describing “panic [PD] as a disorder involving deficits in the cognitive processing and regulation of emotion” (1997, p. 142). In the absence of dysfunction on both the cognitive and affective level, PD would not emerge from that first panic attack. Deficiencies in cognitive and affective regulation that predispose an individual to PD are in part a disruption in the internalization of comforting parental introjects due to bonding and attachment failures (Taylor, 1987). The central component of their preeminent theory of affect regulation is the concept of alexithymia, a multifaceted personality construct introduced by Nemiah and Sifneos (1970) that constitutes a deficiency in adequate affective identification, expression and regulation. It was Sifneos (1973) who coined the term from the Greek: a=lack, lexis=word, thymos=emotion. The alexithymia construct contains these features: (a) difficulty identifying feelings and distinguishing between emotions and corresponding bodily sensations, (b) difficulty describing feelings to others, (c) constricted imaginal life and fantasies, (d) externally oriented cognitive style (Nemiah, Freyberger & Sifneos, 1976; Taylor, 1984a, 1994; Taylor, Bagby & Parker, 1991).  Taylor et al. (1997) state that the deficits that constitute alexithymia are the result of disturbances in affect development in early childhood, consistent with the theories of Krystal, Lane and Schwartz, and Bowlby.

Etiological Theories of Alexithymia

Krystal’s theory of affect development (1988) describes infantile affects as preverbal, somatized and undifferentiated that progress to become verbalized, unsomatized and differentiated in the absence of childhood psychic trauma. This progression results in affect tolerance as well as the capacity for attenuation of affects and the ability to use affects as signals that assist the evaluation of situations and appropriate responses. Childhood psychic trauma can hinder this development, resulting in the dread of affects that represent a return of trauma. He attributes alexithymia to a developmental arrest following infantile or childhood psychic trauma or a regression in affective-cognitive function resulting from adulthood trauma. Alexithymia represents: “regression from verbalized, desomatized, and differentiated affects toward the resomatized and undifferentiated form [that] represents a predisposition to psychosomatic diseases” (p. 264). Similarly, he describes PD as “a perfect example of regression in affects to their infantile, somatic, undifferentiated form.” (1992, p. 407).

Lane and Schwartz (1987) propose a cognitive-developmental theory of emotional development that stresses the fundamental role of emotional disturbance in all mental disorders. Their central premise is that emotional awareness is a form of cognitive processing that develops upward through five levels of structural transformation. The emergence of symbolization and language assists in the child’s development of emotional awareness and capacity to regulate affects on both the intrapersonal as well as interpersonal level. With the development of language, the child can verbally represent and think about their emotions with decreasing dependence on the parent. The formation of language and symbolization allow for cognitive-emotional development with the emergence of cognitive schemas of affects that hierarchically elevate the experience of affect. Within their five level epigenetic model of affect development, emotion at the earliest level is experienced as a bodily sensation and a tendency toward action; later it is experienced both on the psychological and somatic level; and at the highest level it is a blend of different and differentiated feelings with an awareness of other’s feelings. Taylor et al. (1997) assert that disorders of affect regulation and alexithymia result from either a disruption in progression within this model or regression to an earlier level.

Bowlby’s seminal attachment theory (1969, 1973) has influenced the development of models of affect development in that it has much explanatory power for the development of the insecurity involved in the etiology of childhood and adult disorders. His theory challenges the earlier theory of secondary drive as a motivator for attachment behavior. The secondary drive theory states that animal and human desire for proximity to others results from a need for gratification of primary needs such as warmth, shelter, sex, but primarily food. Attachment theory proposes the desire for proximity, or attachment behavior, is driven by a need for protection from predators and the security that protection provides, and that attachment behavior is most highly elicited when threats to personal security cause alarm. When one lacks confidence that protective attachment figures will be available or responsive, the result is what Bowlby termed anxious or insecure attachment. Conversely, confidence in ones attachment figures results in secure attachment. The theory suggests childhood models of attachment persist and individuals with anxious or insecure attachment may therefore be susceptible to manifesting clinging behavior and anxiety symptoms, as “uncertainty about the accessibility and responsiveness of attachment figures is a principal condition for the development of unstable and anxious personality” (1973, p. 322). Bowlby believed that while many factors play a causal role, early life experiences that cause doubt in the availability of attachment figures when most needed contributes to disruption in the formation of secure attachment. In reference to Bowlby’s theory, Taylor (2000) states “alexithymia is associated with insecure attachment” (p. 136).

History of the Alexithymia Construct

The alexithymia construct was born out of studies by several researchers of patients who suffered from the so-called classical psychosomatic disorders. In the 1950s, Horney (1952) and Kelman (1952) noted poor response to psychoanalytic therapy by patients who evidenced psychosomatic symptoms. These patients were characterized by their limited emotional awareness, minimal inner experiences and concrete thinking styles. Similar traits were noted by Sifneos (1967) and Nemiah and Sifneos (1970) in their work with individuals with psychosomatic disorders, traits such as difficulty describing subjective feelings, limited fantasy life, and a communication style marked by over attention to minute, external detail. In that same time period, Krystal (1968) and Krystal and Raskin (1970) were conducting work with patients with drug addictions and patients with posttraumatic disorder, noting similar cognitive and affective characteristics in these patients that Sifneos and Nemiah were reporting. And in 1984, Nemiah described PD as “the prototypical psychosomatic disorder” (p. 134).

Integrated Theory of Affect Regulation

The Taylor (1992) and Taylor, Bagby and Parker (1997) theory of affect regulation proffers a comprehensive explanation of the etiology of PD by integrating elements of all these theories and emphasizing the role of alexithymia in the somatic process. Their theory allows for neurobiological involvement in alexithymia with “constitutional-inherited or acquired disturbances in the regulation of neurotransmitter systems that mediate affects” (1997, p.140) or a deficiency in the coordination of function between the two cerebral hemispheres. One study conducted thus far on the heritability of alexithymia has demonstrated a strong genetic involvement (Heiberg & Heiberg, 1977). Within the Taylor et al. model, the panic attacks of PD are “sudden overwhelming floods of undifferentiated affect” (1997, p. 141) the result of catastrophic misinterpretation of bodily sensations consistent with Clark’s model. Although central neurotransmitter systems are involved in regulating the biological reaction to affects such as separation anxiety, threats to security in important relationships can trigger panic attacks. Individuals with insecure attachment patterns (Bowlby’s model) remain excessively dependent on others, not having adequately separated/individuated from mother and integrating her anxiety-regulating capacity. “Alexithymic patients with panic disorder, however, are likely to show incomplete transitional object development . . . and may still rely on sensation objects for self-regulation” (Taylor et al., 1997, p. 147). Their theory considers the somatization of disorders such as PD as regression to or arrest at earlier preverbal and somatized levels of affect development in the Lane and Schwartz (1987) model, consistent with the Krystal model (1988).

Toronto Alexithymia Scale (TAS)

In 1985 Taylor, Ryan and Bagby developed a self-report instrument to measure for alexithymia, the 26 item Toronto Alexithymia Scale (TAS). This scale drew criticism for its high rate of false positives for alexithymia (Horton, Gewirtz & Kreutter, 1994a, 1994b) and in 1992, Taylor, Bagby and Parker published a revised version. The new Toronto Alexithymia Scale (TAS-20) was shortened to 20 items but demonstrated improved test-retest reliability and internal consistency. It has a three-factor structure theoretically consistent with the alexithymia construct. The three factors are: (a) Factor 1 – difficulty identifying feelings and distinguishing them from bodily sensations of emotion (b) Factor 2 – difficulty describing feelings, and (c) Factor 3 – externally-oriented thinking. This instrument demonstrates superiority of reliability and validity over other available measures such as the MMPI Alexithymia Scale, the Schalling Sifneos Personality Scale, and the Revised Schalling Sifneos Personality Scale, and has been demonstrated to be a psychometrically sound measure of alexithymia (Bagby, Parker & Taylor, 1994; Bagby, Taylor & Parker, 1994).

Three studies have been conducted using the original 26 item TAS to measure alexithymia in a PD sample compared to a control group. One study reported ‘true’ or ‘probable’ alexithymia in 60% of a PD sample and 12% in a healthy control group (Joukamaa & Lepola, 1994). Another study found alexithymia in 47% of patients with PD compared to 12% of patients with social phobia (Parker, Taylor, Bagby & Acklin, 1993), and a third reported alexithymia levels of 67% for a PD sample and 13% for an obsessive-compulsive disorder (OCD) sample (Zeitlin & McNally, 1993). Three similar studies have been conducted using the revised TAS-20. One measured alexithymia in PD and social phobia samples and found rates of 34% and 28% respectively, a nonsignificant difference (Cox, Swinson, Shulman & Bourdeau, 1995). Another reported alexithymia rates of 54% in a PD sample and 58% in a social phobia sample (a nonsignificant difference) compared to 15% in a healthy sample (Fukunishi, Kikuchi, Wogan & Takubo, 1997). The third reported a PD sample to be significantly associated with a lower total TAS-20 score compared to samples with: somatoform disorder, depression and obsessive-compulsive disorder (Bankier, Aigner & Bach, 2001).

Statement of the Problem

This current research explored a correlation between alexithymia and PD in adolescents and adults, and the relationship between PD and the three factors that comprise the TAS-20. Factor one (F1): difficulty identifying feelings and distinguishing them from physical sensations, appears to theoretically overlap both somatization and the amplification construct and would therefore seem to be strongly associated with PD. The covariables of age and gender are considered, as a weak relationship has been demonstrated between TAS-20 scores and age while a significant relationship between TAS-20 scores and gender has been demonstrated (Bagby, Parker & Taylor, 1994). Differences in levels of alexithymia in PD with agoraphobia were compared to levels in PD without agoraphobia. All levels of alexithymia were measured with the TAS-20.

The Hypotheses

This retrospective research study tested the following hypotheses:

  1. There is a significant relationship between alexithymia and PD.
  2. There is no significant relationship between alexithymia and other DSM-IV diagnostic groups that have been conceptually associated with the alexithymia construct.
  3. There is no significant difference in alexithymia between PD with Agoraphobia and PD without Agoraphobia.
  4. There is a significant relationship between the TAS-20 Factor 1 and PD.
The Assumptions
  1. The first assumption is that the TAS-20 is a valid and reliable measure of alexithymia.
  2. The second assumption is that subjects will be honest and accurate in their responses to all items on the TAS.
Study Rationale

This research project was conducted to measure significant differences in the prevalence of alexithymia among disorders that have been associated with alexithymia. This involved a comparison of alexithymia levels in those with PD and those with a disorder characterized by one of the salient symptoms theoretically associated with the construct such as anxiety, depression, phobia, substance abuse, and posttraumatic stress disorder. In conducting this research, I expected to find a significant difference in alexithymia between those with PD and those with other disorders. I did not expect to find significant variance between panic disordered individuals with, and those without, agoraphobia. These expectations are the result of first hand clinical experience with these groups that has strongly suggested high levels of alexithymic characteristics in a PD population, irrespective of agoraphobia, compared to other diagnostic groups. The degree of apparent alexithymia in PD has presented a tremendous therapeutic challenge in assisting these patients to develop the self-awareness necessary to diminish symptoms. Additionally, it was assumed that the TAS -20’s F1: difficulty differentiating between feelings and physical sensations, would be the factor most significantly associated with PD given the predominance of somatization in PD. Confirmation of any of these correlations could provide direction for the development of appropriately focused treatment interventions that target core alexithymic traits in PD as well as alexithymic traits that may present in the general mental health population.

The following literature review explores the relationship between alexithymia and PD. This includes the history of the identification of the alexithymia construct (Nemiah & Sifneos, 1970; Sifneos, 1973) as well as various theories of affect development that contribute explanatory theory to the construct (Bowlby, 1969, 1973; Krystal, 1988; Lane & Schwartz, 1987). This review particularly focuses on the comprehensive theory of affect regulation proffered by Taylor (1992) and Taylor, Bagby and Parker (1997) as well as the prominent tool for measuring alexithymia, the TAS-20 (Taylor, Ryan & Bagby, 1985; Taylor, Bagby & Parker, 1992). Also covered are the prominent theories of PD (Barlow, 1988; Clark, 1986, 1988; Klein, 1993; Krystal, 1988) and the amplification and somatization process involved in PD (Barsky, 1992).  Finally this review summarizes the research studies that have measured and compared levels of alexithymia in subjects with PD as compared to control samples using the original TAS (Joukamaa & Lepola, 1994; Parker, Taylor, Bagby & Acklin, 1993; Zeitlin & McNally, 1993) and the revised TAS-20 (Bankier, Aigner & Bach, 2001; Cox, Swinson, Shulman & Bourdeau, 1995; Fukunishi, Kikuchi, Wogan & Takubo, 1997).

Genesis of the Alexithymia Construct

The earliest description of alexithymic characteristics was reported by MacLean (1949) when he noted psychosomatic patient’s inability to verbalize their feelings. Ruesch (1948) noted similar limitations in symbolism and verbalization in post traumatic and psychosomatic patients, and considered them to have an ‘immature’ or ‘infantile personality’. Horney (1952) and Kelman (1952) observed similar characteristics in psychosomatic psychiatric patients who demonstrated very poor response to psychoanalytic treatment due to concrete thinking and limited emotional awareness. Around the same time, Krystal (1962, 1968) was describing a similar lack of differentiation of affects in drug addicts and in patients with severe post traumatic states. In the 1960s and 1970s, Sifneos (1967) and Nemiah and Sifneos (1970) were conducting studies into similar cognitive and affective traits. In the absence of a better term, Sifneos (1973) coined the term alexithymia to describe the deficits in affective identification and awareness that were evidenced by patients with psychosomatic disorders he interviewed from 1954 to 1968 while the director of the Psychiatric Clinic at Massachusetts General Hospital.

Characteristics of Alexithymia

The traits Sifneos (1973) noted in some, though not all, of these psychosomatic patients included: difficulty finding words to describe emotions, a tendency to act out to express emotion and avoid conflict, poverty of fantasy life, an “overall appearance of being dull . . . and a relative constriction in emotional functioning” (p.256). He also noted the presence of externally-oriented or operative thinking termed “pensee operatoire” (Marty & de M’Uzan, 1963) and he speculated that these deficits may result from developmental or biological factors.

Alexithymia is considered a dimensional trait that has a normal distribution in the general population, as opposed to a categorical trait that is either present or absent. It also does not appear to be related to intelligence, socioeconomic status, educational level or culture (Taylor, Bagby & Parker, 1997). Alexithymic individuals are prone to states of undifferentiated negative affects and limited positive affects. Krystal (1988) and Sifneos (1987) have described this limited capacity for positive emotional states like happiness and love as constituting anhedonia. Alexithymic individuals have a propensity to develop overly dependent relationships, or conversely, may prefer to be alone and avoid formation of relationships, an apparent adult manifestation of an insecure attachment style (Taylor et al., 1997).

Additional characteristics associated with the alexithymia construct include a tendency toward action to express emotions, avoidance of conflict, poor dream recollection, social conformity, and wooden posture with few facial expressions (Taylor et al., 1997). Sifneos (1967) noted frequent complaints of anxiety and depression with very limited ability to describe those emotions. Taylor (1984b) noted of individuals with alexithymia: “speech is lacking in nuance, meager in the use of metaphor and devoid of affect” (p. 218). As affects are not adequately verbalized, it is very difficult for alexithymic individuals to elicit support and comfort from others. They have difficulty imagining themselves in another’s situation which limits their empathy and support for others. A 1993 study (Schaffer) demonstrated that while non-alexithymic individuals tend to employ interpersonal contact or soothing fantasy activity to regulate affects, alexithymic individuals tend to rely on oral or somatic modes of affect regulation such as overeating or somatization. (Taylor et al., 1997).

Alexithymia: State or Trait

There is considerable debate whether alexithymia is an enduring trait or a state reaction to some external event, or both. Freyberger (1977) refers to constriction of emotions in medically ill patients or those in a life-threatening state as secondary alexithymia. In these situations, emotional numbing appears to be the result of the defense mechanism of denial. Sifneos (1988) differentiated primary alexithymia: the result of neurobiological deficits, from secondary alexithymia: the consequence of early or late psychological trauma or developmental arrest. Taylor et al. “distinguish between alexithymia as a stable personality trait that is independent of etiology and alexithymia that is state-dependent and disappears after the evoking stressful situation has changed.

. . . Strong empirical support for alexithymia being a stable personality trait, rather than just a consequence of psychological distress, has been provided by several prospective studies” (1997, p. 37).

Krystal (1988) states that alexithymia and anhedonia are separate but concomitant consequences of significant trauma, but the presence of anhedonia with alexithymia is a key indicator that the alexithymia is the result of trauma – a state reaction. “Posttraumatic alexithymics have such guilt about any pleasure and gratification that they conceal even minimal gratification” (p. 253).  Sifneos (1967, 1975) viewed the alexithymia phenomenon as a state of biological or developmental origin, proposing a possible inborn predisposition for alexithymia. He believed it should not be treated with anxiety-provoking therapy such as psychoanalysis which may only increase the psychosomatic symptoms that result from anxiety.

A 1997 study (Fukunishi, Kikuchi, Wogan & Takubo) in Tokyo conducted pre and post treatment measures of alexithymia among two sample groups. Using the TAS-20, they found pretreatment elevation in alexithymia levels among 26 patients with PD (54%) and 24 patients with social phobia (58%). After treatment, they conducted a 6-month follow-up study and found the prevalence rate of alexithymia was 30.8% for the PD group and 33.3% for the social phobia group, a significant decrease for both patient groups. Overall, there was a decrease in TAS-20 scores for 88.5% and 83.3% of the patients in the PD group and social phobia group, respectively. Based on these findings, the authors conclude that the change in alexithymia that results from psychotherapeutic interventions and reduced anxiety suggests that alexithymia is a state reaction to anxiety.

First Comparative Research Study of Alexithymic Traits

Sifneos (1972/1973) conducted a study to compare the rate of these alexithymic traits in a group of 25 psychosomatic patients with a control group of 25. The so-called psychosomatic group consisted of 9 patients with ulcerative colitis, 9 with asthma, 5 with peptic ulcer, and 2 with rheumatoid arthritis. The control group involved patients with diagnoses such as: borderline personality, depression, alcoholism and what was termed hysterical personality. Using the 17 item Psychosomatic Questionnaire of the Beth Israel Hospital Psychiatric Service, the experimental group endorsed twice as many positive responses to the key alexithymic questions as did the control group. His work convinced Sifneos that while these characteristics are evident in psychosomatic type disorders, there is an even wider distribution of some alexithymic traits in the general population than he had originally anticipated. In particular he noted alexithymic characteristics among drug addicts, alcoholics, paranoid and borderline personalities.

Theories of the Etiology of Alexithymia

Biological model
MacLean (1949) proposed that psychosomatic individuals have difficulty verbalizing feelings because of a lack of connections between the limbic system and the neocortex. He speculated that their emotions cannot be adequately processed by the neocortex to find expression in the symbolic use of words, but instead are immediately discharged through autonomic pathways of physiological arousal and expression. Nemiah (1975) and Nemiah, Freyberger and Sifneos (1976) also posit a neurological theory of a deficiency in the flow of information between the ‘visceral brain’ or the hypothalamus, and the language centers of the cerebral cortex. Nemiah posits the efficacy of antidepressant medication in the treatment of PD as support of a biological model of panic.

The ‘split brain’ studies by Hoppe and Bogen (1977) who researched 12 cases of commisurotomized or surgically divided hemispheres, provided solid anatomical evidence of severe resultant alexithymia. Their data reported the “left hemisphere is linguistic, analytic, logical, sequential, and constructive, whereas the right hemisphere is specialized in visual-spacial [sic], synthetic, and Gestalt perception. Without the right hemisphere we are not able to sing in a melodic way and we have difficulty in perceiving the whole Gestalt. The right hemisphere senses the forest, so to speak, while the left one cannot see the forest for the trees (Hoppe, 1981, p. 6)” (Krystal, 1988, p. 257).

Deficit model vs. defense model
Nemiah, Freyberger and Sifneos (1976) propose a deficit model, similar to that of Marty and de M’Uzan (1963), suggesting alexithymic traits result from deficits in personality organization as opposed to a defense model of alexithymia as a product of defense mechanisms. Nemiah (1984) suggested that psychosomatic symptoms, whether panic or physiological, result when a deficit in higher order processing leads to a “short-circuiting of arousal into autonomic channels” (p. 134).  Nemiah (1997) elaborated on the processes that occur in response to affect provoking events, describing two sets of reactions: perceptual-cognitive and affective. First there is a conscious perception and cognitive evaluation of the event and a secondary somatic/emotional arousal response. This secondary response undergoes a process of psychic evaluation of several components. This evaluation process involves: redefining the raw emotion into identifiable feelings such as anger, sadness, joy or fear; identifying descriptive words for the feeling; producing fantasies expressive of the feelings; and associating these feelings to related memories. Nemiah suggests alexithymia as a disturbance somewhere along this chain of processes in the affective response, due to either psychological defenses or deficits, or due to physiological disturbances in the neuronal structures and pathways that underlie psychological processing.

Treatment with psychosomatic individuals indicates they do not respond to a conflict model of treatment, as psychological exploration does not reveal the presence of unconscious feelings or fantasies that would constitute internal conflict and generate symptom formation. Instead, it appears that the psychosomatic process is a manifestation of “ego deficits in the capacity for fantasy production and for experiencing and differentiating affects … the result of disturbances in early growth and development or of regression from more mature forms of psychological functioning” (Nemiah, 1984, p. 133). Therefore a deficit model is essential for understanding the development of psychosomatic disorders. Nemiah notes the similarity of a deficit model to Freud’s early conceptualization of anxiety as the somatic manifestation of internal arousal without higher level cerebral processing.

Psychoanalytic perspective
Joyce McDougall, (1974) one of the most prominent psychoanalytic contributors to the body of theoretical knowledge of affective disturbances, postulates the problems of alexithymic and psychosomatic individuals are rooted in early preverbal mothering difficulties. The child perceives the mother’s ambivalence over the loss of the child via differentiation, separation and individuation, thereby diminishing the child’s comfort over becoming a separate individual. Denial of the child’s own identity results, eroding the connection between self-representations and love-object representations, and a lack of identification with a soothing, caretaking mother. This contributes to limited spontaneous fantasy and a robot like consequence: adapting well to reality but without a world of imagination or feeling. McDougall (1984) additionally states that: “Sexual perversions are probably the most alexithymic form of libidinal expression that exist!” (p. 392).

First Connection of Alexithymia to PD

Nemiah (1984) is credited for the first elucidation of a relationship between alexithymia and PD. In an article describing the differences between the symptoms of neurotic patients and those with psychosomatic disorders, he noted these traits of the psychosomatic population. “They could not associate like neurotic patients, they were unable to talk about feelings or fantasies, and they maintained a stubborn, unchanging inability to recognize internal psychological conflicts or problems despite the attempts of skilled therapists to uncover emotional factors” (p. 132). He further described their limited involvement in therapy, and premature termination with minimal degree of therapeutic progress. Nemiah cited the earlier emergence of the alexithymic construct whose salient features correlate with the psychosomatic process and symptomatology. He noted the presence of prominent alexithymic characteristics in his psychosomatic patients, such as difficulty describing and differentiating between feelings, limited production of fantasy and preoccupation with external events. Within this context, he states that one could consider “panic anxiety … as the prototypical psychosomatic disorder” (p. 134), putting forward the association between PD and alexithymia.

Krystal (1992) later supported the association of alexithymia and PD when he described panic as “an attack of mixed physiological elements of dysphoric affects …. a perfect example of regression in affects to their infantile, somatic, undifferentiated form” (p. 407) and described alexithymia as “regression from verbalized, desomatized, and differentiated affects toward the resomatized and undifferentiated form” (1988, p. 264). Taylor, Bagby and Parker (1997) consider panic to be a disorder of affect regulation in which alexithymia is central to its etiological cognitive and affective deficits.

Theories of Affect Development

John Bowlby
Bowlby’s attachment theory (1969, 1973) provides elucidation of the impact of early bonding disruptions on affect development and affect regulation – concepts central to the alexithymia construct. Bowlby purports attachment behavior manifests in most human infants within the first four to seven months of life, with many infants of four months of age already capable of distinguishing between mother and others as evidenced by different responses to each. Proximity-maintaining behavior is demonstrated by the infant’s cries when mother leaves the room with attempts to follow her as soon as they are able to crawl, and attachment behaviors intensify when the child is alarmed by environmental stimuli such as mother’s departure or the approach of a stranger. During the second and much of the third year, the child’s increasing awareness of the impending departure of the mother triggers attachment behaviors: crying, protesting, or attempting to cling and follow.

Bowlby notes a sudden development when the child approaches their third birthday – an increased capacity for temporary separations from the mother and the ability to feel secure in a strange place. This security is conditional upon: the presence of a familiar subordinate caregiver, the absence of any sense of alarm or illness, awareness of the mother’s whereabouts with knowledge of reunion within a short timeframe. Attachment behavior is demonstrated with less frequency and intensity after age three, decreasing further throughout adolescence as attachment to parents grows weaker while attachment and attraction to others increase. Throughout adulthood it remains a normal response, in illness or times of stress, for attachment behavior to manifest in seeking proximity to known or trusted individuals. Attachment behavior is accompanied by the most powerful of human affects: threat of loss causes anxiety, actual loss causes sorrow, and proximity to an attachment-figure creates a sense of security.

Bowlby’s theory challenges the earlier theory of secondary drive as an explanation for proximity-seeking behavior. Secondary drive theory, evolved from assumptions versus direct observations, suggests the attachment to caregivers derives from primary drives: food, liquid, warmth and sex. This theory was seriously challenged by the early work on imprinting by Lorenz (1935/1957). His work and that of subsequent researchers (Cairns, 1966; Cairns & Johnson, 1965; Scott, 1963; Shipley, 1963) demonstrated attachment behavior developing not out of primary drive gratification, but rather out of the need for contact comfort. Harlow’s groundbreaking work with rhesus monkeys in the early 1960’s corroborated the primacy of contact comfort over food. His experiments demonstrated monkeys consistently favoring a soft, cloth non-feeding surrogate mother figure over a wire model that included a feeding bottle. While some time was spent feeding on the wire model, most time was spent on the cloth model. Some of these infant monkeys would even cling to the soft figure while leaning over to suck from the feeding model. These monkeys would increase their attachment to the cloth model when alarmed, finding no reassurance from the wire model, and intensifying their clinging behavior when their sense of alarm increased (Harlow, 1961).

Harlow’s experiments provided explanation for the paradoxical clinging response to an abuser, be it parent or partner. The increased alarm and sense of threat posed by abuse triggers a proximity-seeking response to an attachment figure, even when the attachment figure is the source of the abuse. Bowlby elucidates the ultimate function of attachment behavior is survival of the species. Not only does attachment behavior provide protection from predators, but allows the infant to learn basic survival skills from the mother or primary caregiver.

Bowlby suggested that individuals, who demonstrate behaviors that can be described as dependent or over-dependent, are exhibiting attachment behavior more frequently and urgently than their age or circumstances warrant. He termed this condition anxious or insecure attachment, due to the lack of confidence that attachment figures will be accessible and/or adequately responsive to one’s needs. Bowlby believed the cause of this condition to be experiences that gave an individual reason to question the availability of their attachment figures – either the result of actual separations or threats of abandonment. He defined anxiety as the feeling precipitated when “attachment behavior is activated” (p. 405). This includes the feeling triggered by the unsuccessful seeking of an attachment figure or the uncertainty over the availability of an attachment figure. “When an insecure individual, uncertain whether his attachment figures are going to be accessible and responsive, or even alive, is faced with a potentially fear-arousing situation, he is more likely to respond to it with fear, and also more likely to respond with intense fear, than is an individual who feels secure and confident in his attachment figures” (p. 313). Adult anxiety disorders, including agoraphobia and panic attacks, may be adult sequelae of insecure attachment patterns laid down in childhood (1973).

Conversely, Bowlby cites confidence in ones attachment figures as the foundation for the stable and self-reliant personality he terms secure attachment. “Uncertainty about the accessibility and responsiveness of attachment figures is a principal condition for the development of unstable and anxious personality so is there [sic] a strong case for believing that an unthinking confidence in the unfailing accessibility and support of attachment figures is the bedrock on which stable and self-reliant personality is built” (p. 322). Elements that contribute to this secure personality include a balance of parental support with appropriately timed encouragement toward autonomy, and healthy communication patterns. Children identify with their parents and adopt their parenting pattern, “Thus the inheritance of mental health and of mental ill health through the medium of family microculture is certainly no less important, and may well be far more important, than is their inheritance through the medium of genes” (p. 323). He believed that early life models of attachment figures persist throughout life, secure or insecure.

Henry Krystal
Krystal’s epigenetic model of affect development (1988) offers insight into the developmental arrest or regression that contributes to the traits that constitute alexithymia.  Krystal assumes infantile affective precursors comprised of states of contentment or distress, out of which develop the adult affects. Contentment is the precursor to the pleasurable affects, and distress the precursor to painful affects. “In the course of maturation, affects evolve from two affect precursors into specific emotions” (p. 41). Affect development occurs along two lines: affect differentiation, and affect verbalization with concomitant desomatization. This affect development process is directly affected by the infant’s subjective experiences. As “described by Ramzy and Wallerstein (1958): The neonate has internal tensions (homeostatic disequilibria) which are ‘perceived’ (referring to a precursor of perception) not as mental status, but as diffuse tensions on an undifferentiated biopsychological level. It is only later in development that such a tension state, as for example, hunger, can be separately viewed both on the psychological level (the feeling of hunger) and on the physiological level (concomitant discernible somatic processes) [p. 175]” (p. 43).

Out of direct infant observation, Greenspan (1981) concluded that homeostasis, the regulation of excitation by the calming response, is the most crucial achievement in the first three months of life. This develops by the comforting response of the mother and gradually evolves into the child’s self-comforting ability. A considerable degree of affect differentiation occurs before affects are specific enough to be verbalized. By the age of two the child can identify affects differentiated out of contentment: delight, joy, affection; and out of distress: fear and anger. The presence of massive psychic trauma in infancy or toddler age may hinder the development of affects, and the extent to which affects remain undifferentiated and somatized is the extent to which they will lack meaning as a signal.

In adolescence the child develops the capacity to grieve loss, especially of infantile attachment to parents as well as infantile omnipotence. Affect tolerance develops parallel with the development of the affects, reflecting developing self-awareness as well as comfort with and ease in handling affects. The adolescent also confronts the need to develop tolerance for boredom, discomfort and pain; necessitating their giving themselves credit for accomplishments as would an appreciative parent. Temporary devices that assist the adolescent until they develop adult level affect tolerance include a sense of counter phobic narcissistic invulnerability and exhibitionistic tendencies. Prominent among adolescent challenges is the process of separation-individuation and establishment of ego identity. (Erikson, 1956).

Krystal on alexithymia
Consistent with his affect theory, Krystal (1988) conceives alexithymia as “a regression or an arrest in affective and cognitive development, with severe distortions resulting from infantile or catastrophic adult trauma” (p. 328) and that this is an arrest or regression from “verbalized, desomatized, and differentiated affects toward the resomatized and undifferentiated form [which] represents a predisposition to psychosomatic diseases” (p. 264). He states individuals with alexithymia have difficulty utilizing emotions as signals due to the form of their emotional responses. That is, their affects are basically somatic with minimal verbalization. Though they may function very well in their work and appear to be of superior intellectual function, he notes “a sterility and monotony of ideas and a severe impoverishment of the imagination” (p. 247) and describes them as “dull, colorless, and boring” (Krystal, 1979, p. 19) with expressionless faces and rigid posture.

Due to inadequate capacity for fantasy-making and symbolization they are left with the physiological aspects of their affects and therefore are prone to psychosomatic disorders. They may attempt to medicate these responses with drugs or alcohol which leaves them susceptible to drug addiction or alcoholism. Krystal also suggests that these individuals can only signal their distress, leaving all the functions of verbalizing such as soothing and care giving to the mother. They cannot adequately organize their thoughts into logical, coherent communication and feel they must not take over the maternal role of making sense of their distress or engage in any independent thinking, so they merely signal their distress with a lot of trivial material in chronological order. Their affective responses therefore tend to be physical and may involve various bodily systems.

Lane & Schwartz
In 1987 Lane and Schwartz proffered their epigenetic model of the cognitive development of affects, a “seminal contribution” as per Krystal (1988, p. 311). They emphasize that emotional disturbance may well be inherent in every mental health diagnostic category from mood disorders, schizophrenia and psychosomatic disorders, to the personality disorders. Emotion is a complex phenomenon of three domains: physiological (or biological); psychological (or experiential); and expressive (or social). Most mental health patients seek treatment for distress in the psychological/ experiential domain as they are in a state of emotional pain (Lane & Schwartz, 1987). They concur with Lazarus (1984) that emotion is preceded by cognition; that cognitive appraisal of the environment activates the affective response. Their theory proposes that the cognitive process undergoes transformation throughout the individual’s development, and thereby affects subsequent emotional experiences. “Our primary thesis is that emotional awareness is a type of cognitive processing which undergoes five levels of structural transformation along a cognitive-developmental sequence derived from an integration of the theories of Piaget and Werner” (p. 134). Their theory is influenced by Werner’s (1957) principle that wherever development occurs, it proceeds from a global state devoid of differentiation to a state of increasing differentiation, verbalization and hierarchic integration. The development of symbolization and language facilitates not only articulation of affect but a cognitive structure for affect. Lane and Schwartz build their developmental model upon Piaget’s hierarchical model of four stages of cognitive development – sensorimotor, preoperational, concrete operational, and formal operational periods. Their model encompasses the following five levels of structural transformation of awareness:

  1. Sensorimotor reflexive – At this first level of emotional awareness the involuntary motor responses that accompany emotional arousal become activated. These include autonomic changes and facial expression, with global arousal consisting of bodily sensation only. The individual cannot report anything though an observer can identify the quality of the emotion being experienced. There is minimal to nonexistent awareness of the separate existence of others.
  2. Sensorimotor enactive – In the second level emotion is experienced as a bodily sensation and an action tendency though conscious awareness of emotion has not yet developed. Affective states are global states of pleasure or displeasure, and action is aimed to maximize pleasure or minimize distress. The awareness of another as a separate individual is minimal.
  3. Preoperational – The representation and awareness of emotion is possible for the first time at the third level of development. Emotions are becoming a psychological as well as somatic experience and are singular – happy or sad without nuances. They are also limited in range along with their verbal descriptions. Others are recognized as separate from the self but without awareness of the other’s emotional experiences.
  4. Concrete operational – Awareness of a blend of feelings including concurrent opposing feelings and a wider range of emotions becomes manifest at this level. Emotions have become differentiated from one another both qualitatively and quantitatively. There is a capacity to modulate emotional extremes, to maintain hope in the face of challenges, and maintain different feelings for another despite current circumstances. There is a capacity to articulate complex and differentiated affects and an ability to anticipate the responses of others. Although others are recognized as separate and different based on internal as well as external attributes, the perception of another’s emotional experience is relatively undifferentiated from the perception of one’s own experience.
  5. Formal operational – The major advance at this level is greater differentiation and integration in ones appreciation of the emotional experience of others compared to ones own. There is a greater capacity for a blend of feelings of varying qualities or intensities with greater differentiation among them. The ability to describe emotions and their nuances in novel ways or with metaphors is evident. It is possible to perceive the differentiated experience of others unbiased by ones own feelings and see ones self and another’s situation through the eyes of another. The capacity to anticipate emotional consequences of personal or professional decisions increases the likelihood of making wiser decisions. By being able to anticipate the needs and reactions of ones self and others, one is able to make more satisfying and appropriate decisions or take more effective courses of action. The self has reached peak differentiation from others with recognition of the uniqueness as well as similarities between both.

Lane & Schwartz on alexithymia
The authors identify (1987) the clinical entity perhaps most directly addressed by their theory is alexithymia. They state that alexithymic individuals may self-perpetuate their undifferentiated emotions by avoiding “reflecting on and generating symbolic representations of experience. Our model suggests that an important reason for this avoidance is that unpleasant emotional arousal is experienced as overwhelming somatic distress when it is attended to” (p. 140). They suggest that individuals with high levels of cognitive complexity experience less emotional distress than those individuals with lower levels of cognitive complexity.

Theory of Affect Regulation

The most comprehensive theory of affect development is the theory of affect regulation by Taylor (1992) and Taylor, Bagby and Parker (1997). They recognize a biological basis of emotions as established by Darwin who believed emotions are an adaptation that organizes survival behavior. Emotions are a biologically based readiness toward action, but the actions taken to control how these impulses are experienced and expressed constitutes the concept of affect regulation.  They also concur with Lazarus (1991) and Zajonc (1984) who promote a cognitive theory of emotion, noting that both phylogenetically and ontogenetically, affect precedes the development of cognition but “cognition becomes linked inseparably with affects early in development . . . . cognitive processes play a vital role in the regulation of affects” (Taylor et al., p. 13). They also “regard affect regulation as a process involving reciprocal interactions between the neurophysiological, motor-expressive, and cognitive-experiential domains of emotion response systems” (p. 14). All three domains are involved in the regulation of affects, with activation in one influencing the others. Additionally, ones interaction within social relationships provides interpersonal regulation that may be either soothing or distressing. “Affects themselves may also regulate other affects . . . the positive affect of interest may attenuate fear and sadness” (p. 14).

Affect development
Taylor et al. (1997) state that undifferentiated precursor states of contentment and distress, consistent with Krystal’s theory (1988), are evident in newborn children and appear to be similar in all cultures, manifesting both physiologically and behaviorally. As maturation proceeds, these affect states undergo progressive differentiation and desomatization, while language development facilitates the articulation of symbolic representations of emotions. The authors concur with the cognitive development of affects model posited by Lane and Schwartz (1987) of structural transformation in a hierarchical developmental sequence. They credit this model for its conceptualization of affect development and the explanation it provides for the range of differences between individuals in the capacity to experience and express affects, as well as capacity for empathy.

The development of affect and the corresponding cognitive skills for regulating affect are clearly influenced by the child’s relationship with primary caregivers. Affect development in the infant will be negatively impacted by failure on the parent’s part to read the infant’s emotional cues and respond properly to externally regulate the infant’s emotional state (Bion, 1962, 1965). The development of symbolization and language facilitates growth of the child’s subjective emotional awareness during their second year. This is facilitated by the parents providing words and meaning for the child’s somatic emotional experience (Edgcumbe, 1984; Emde, 1984; Furman, 1992). But it is the child’s acquisition of language that has considerable effect upon the development of affect regulation by facilitating their verbalization of affects, receiving of feedback from others, and hearing and thinking about their feelings and how to manage them (Kopp, 1989). “The verbalization of affects leads to new experiences and a growing awareness of more complex and differentiated emotional states (Stern, 1985). This will occur only if the child is reared in a family environment where feelings are verbally labeled and validated” (Taylor et al., 1997, p. 18). Teaching children to label emotions and express them in words not only transforms behavioral expression into verbal expression but encourages cognitive appraisal of emotions (Katan, 1961). Affect tolerance, the capacity to “contain and tolerate the tensions generated by feelings and needs without always having to rely on parents” (Taylor et al., p. 19) increases with the verbalization and cognitive awareness of affects. The child also develops the capacity to use feelings of sadness, anxiety, and other emotions as signals that can be evaluated to provide guidance for reducing or managing stressors (Krystal, 1975). The ability to represent awareness of affect as well as awareness of another’s affect is generally achieved in the third year of life (Fonagy, 1991; Hobson, 1994).

Attachment development
A significant determinant of a child’s ability to regulate distressing affects and relate to others is the degree of sensitivity and responsiveness to their emotional states by their caregivers in infancy and childhood (Bretherton, 1985; Goldberg, MacKay-Soroka, & Rochester, 1994). Children can achieve secure attachment when they have experienced consistent and supportive responsiveness from caregivers. They seek comfort from their parents when they are emotionally distressed and display more positive affect, have higher levels of symbolic play than insecurely attached children and demonstrate more competence and adaptability in their relationships (Malatesta, 1990; Slade & Aber, 1992).

Insecure attachment impedes the development of affect tolerance and regulation. Insecurely attached children experienced a caregiver who was insensitive to the child’s emotional cues or rejected the child’s proximity-seeking behavior. Insecure attachment patterns can manifest in avoidance of frustrated comfort-seeking behavior and suppression of affect display; or escalating, disorganized affect displays that attempt to elicit a comforting parental response. Numerous variables on behalf of the infant or mother can contribute to disruption in the formation of secure attachment.

Early attachment experiences influence the development of an internal model of the self and others that impacts both affect regulation as well as future relationships. The attachment style developed in childhood appears to endure for a lifetime and may influence future generations. Securely attached adults seem to experience more positive affect and develop relationships that provide support in times of emotional distress. Adults with insecure attachment styles experience higher levels of negative affects such as anxiety and depression and develop relationships that provide limited emotional support.

Affect regulation development
While infant precursor emotional states are largely regulated by parents, independent affect regulating behaviors such as thumb or finger sucking quickly evolve. These precede the later emergence of a transitional object, such as a blanket or soft toy, for self-comforting in periods of distress such as separations from the mother. While these behaviors manifest even in children who later develop secure attachment, the use of a transitional object is even more critical for infants who are weaned early or have experienced early separation from the mother (Taylor, 1987). Discouragement by the mother of the child’s creation of transitional objects inhibits the development of fantasy, imaginal activities and play. These children are left with auto sensual behaviors such as hair twirling, thumb sucking or masturbation, which are primitive modes of affect regulation. Adolescent and adult equivalents may be sensation seeking behaviors such as cigarette smoking, alcohol use or overeating.

The development of pleasant self-comforting dreams, fantasies and play are largely influenced by the extent of relatedness between the infant and a nurturing mother. Imagination and creative ability play an important role in the development of personality and affect regulation. Early childhood play is an important developmental process that allows a child to manage negative affects and increase positive affects of interest and joy. Imaginal activity in play also manifests through the formation of symbolization evident in wish-fulfilling dreams and creative stories. The ability to symbolize facilitates the organization and modulation of affective experiences. As transitional objects are discarded, the acquisition of new interests and activities help provide self-soothing functions. Interests provide positive affects and enhance social relationships which in turn facilitate affect regulation.

Affect development and regulation are complex developmental processes influenced by temperamental, neurobiological and social factors. For example, an infant with a neurological defect may not be capable of adequately responding to an emotionally responsive mother. Or a depressed and emotionally unavailable or insensitive mother may be impaired in her ability to properly respond to her infant’s emotional states. Cultural factors may impact the development of affect and its regulation as they influence the degree of maternal responsiveness to an infant’s emotional cues based on sociocultural norms regarding affect significance and expression. A comprehensive theory of affect development and regulation considers the complex relationship between affect, cognition and attachment behavior as a major influence on the organization of the personality. “Crittenden (1994), for example, relates successful affect development and regulation to the experience of secure attachment; it provides predictable positive outcomes to affective communications, and thereby facilitates a satisfactory integration of affective information with cognition information. Consequently, the child is able to ‘use cognition to moderate affect and affect to inform cognition’ ” (Taylor et al., 1997, p. 24-25).

Taylor, Bagby and Parker on alexithymia
Early life failures in developing the capacity to cognitively experience and regulate affects can result in affect dysregulation and psychopathology or physical ailments. The authors consider these failures the result of arrest at or regression to lower levels of affect development as proposed by the Lane and Schwartz (1987) model. They consider functioning at these lower developmental levels to be the personality trait of alexithymia. Furthermore, though alexithymia has been associated with primitive defenses, it is not a defense mechanism itself. (Taylor et al., 1997).

Recognition of Panic Disorder as a Distinct Disorder

Klein significantly influenced the recognition of PD as a distinct diagnostic category in the 1980 American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III), while Barlow influenced the 1994 DSM-IV definition of PD. The DSM-IV conceptualization of PD considers the significant role played by the persistent anxiety over physical sensations, termed fear of fear (Goldstein & Chambless, 1978) or anxiety sensitivity (Reiss & McNally, 1985). The DSM-IV also distinguishes between unexpected (or uncued) attacks; situationally bound (or cued) attacks; and situationally predisposed attacks. See Appendix C for complete DSM-IV definition.

The DSM-IV definition of PD was also influenced by Klein’s model (1993) which restricts panic triggering cues to external stimuli and does not include internal cues. Barlow (1988) believes triggers can be external events as well as internal,  interoceptive stimuli, and Craske (1991) believes that panic attacks in PD result from internal cues including stressful life events, whereas panic by phobics is triggered by external cues. The Clark (1986) and Barsky (1992) models are predicated upon a presumption of panic that stems from interoceptive cues.

Prominent Etiological Theories of PD

Barlow – false alarm theory
Barlow (1988) proffers a biopsychosocial model that postulates panic as the misfiring of a biological system designed to serve as an alarm of pending danger. It is well established that fear is a natural reaction to environmental threats, serving to trigger a survival response to the threat. Ancestral threats such as an animal attack, to more modern threats such as armed robbery, trigger a true alarm. This alarm mobilizes a physical and cognitive fight or flight response of innate and significant evolutionary value, to either fight off the attack or take flight. But this same set of biological and psychological reactions, triggered in the absence of a life-threatening stimulus, learned or unlearned, constitutes a false alarm or a panic attack. Barlow states the causes of false alarms include biological predisposition, a history of childhood separation anxiety, and stress generated by life events.

While Barlow says “It would certainly seem logical that a biological dysregulation underlies false alarms” he goes on to explain that “at the present time there is no evidence for any specific biological marker; nor, for that matter, is there evidence for any important neurobiological differences between patients with panic disorder and nonpanickers” (1988, p. 211). He does suggest that a biological marker for panic may be chronic hyperarousal. Though it is an attribute of all with an anxiety disorder, it may interact with other variables to precipitate a false alarm.

Barlow cites a Russian study (Razran, 1961) that demonstrates fear can become conditioned to internal physiological stimuli. An association developed between a negative external event and internal sensations can be extremely resistant to extinction, even persisting indefinitely. If an association between internal cues and false alarms develops, the internal cues can signal the possibility of another false alarm. “The association of false alarms with internal or external cues results in the phenomenon of learned alarms” (Barlow, 1988, p. 229). Fear can be learned to numerous somatic cues – cardiovascular or respiratory. If a sense of threat by driving over a bridge triggers a false alarm and a racing heart which then generates fear of a heart attack, future false alarms could be triggered by the learned alarm of heart palpitations. “The association of false alarms with interoceptive or somatic cues is particularly crucial in the development of panic disorder” (p. 366). Considerable evidence indicates that people with PD have learned to fear interoceptive cues, and that therapeutic desensitization to those cues is very effective in the treatment of PD.

Barlow’s model begins with biological vulnerability: the tendency toward neurobiologically based over reactivity to stressful events that forms a platform for an initial “hard-wired alarm reaction” (p. 366). He believes the tendency to over react to stressors with exaggerated neurobiological arousal is genetically based and that “panic disorder patients are biologically predisposed to react to negative life events with neurobiological lability” (p. 368). The second component in his model is the stress trigger, which precipitates the first false alarm/panic attack. By association with interoceptive cues, the first false alarm becomes a learned alarm. That association leads to psychological vulnerability, the anxious apprehension that focuses on future alarms. This then leads to autonomic and cognitive symptoms of anxiety and further somatic cues that trigger more learned alarms. Finally, this process may lead to development of agoraphobic avoidance.

Barlow believes that panic attacks, especially in conjunction with agoraphobia, may be an adult analogue of separation anxiety experienced by children confronted with separation from their mother. However, he does note the conflicting evidence regarding a positive correlation between adult PD and childhood separation anxiety. The role of stress as a precipitant of panic attacks is well documented, and various studies have demonstrated that up to 96% of individuals with PD can identify a significant life stressor as a precursor of their first panic attack. Certain individuals may be more susceptible to the impact of stressful events based on the extent of social support, constitutional factors and a combination of cognitive and psychological traits. They would then be disposed to react to a negative life event as if it were a life threatening danger analogous to an animal attack. Barlow’s theoretical contributions to the body of knowledge regarding PD influenced the classification and description of PD in the DSM-IV.

Clark – catastrophic misinterpretation theory
Individuals with a history of panic attacks can develop panic in response to physiological agents such as inhalation of carbon dioxide, infusions of lactate or yohimbine, excessive caffeine or voluntary hyperventilation. Though these substances produce the same physiological sensations associated with panic in all individuals, those without a history of panic attacks will not experience panic in response. This fact led some theorists to suggest the difference to be the result of a biochemical disorder present only in those who react with panic. Clark (1986) suggested the difference to be the result of different interpretations as to the significance of the physiological sensations:
It is proposed that panic attacks result from the catastrophic misinterpretation of certain bodily sensations. The sensations which are misinterpreted are mainly those which are involved in normal anxiety responses (e.g. palpitations, breathlessness, dizziness etc.) but also include some other bodily sensations. The catastrophic misinterpretation involves perceiving these sensations as much more dangerous than they really are. Examples of catastrophic misinterpretations would be a healthy individual perceiving palpitations as evidence of impending heart attack; perceiving a slight feeling of breathlessness as evidence of impending cessation of breathing and consequent death; or perceiving a shaky feeling as evidence of impending loss of control and insanity (p. 462).

Clark further described the vicious cycle that develops in those who panic. An external stimulus (such as driving on a freeway) or an internal stimulus (a physical sensation, thought or image) may be perceived as a threat which generates apprehension. Physiological sensations develop or intensify from the apprehension of this inaccurately perceived threat. If these sensations are interpreted as signals of pending catastrophe, further intensification of sensations result and intensify the sense of threat. This vicious cycle can escalate and trigger a panic attack. This model can explain panic attacks that appear to ‘come out of the blue’ and those that result from heightened anxiety. Panic attacks that manifest from heightened anxiety are one of two types: a panic attack resulting from heightened anxiety about having an attack, or a panic attack triggered by the physical sensations of anxious arousal. Panic attacks ‘out of the blue’ often result from catastrophic misinterpretation of the sensations of benign physiological arousal. Normal emotions such as anger or excitement can cause increased heart rate, increased respiration or flushing. Exercise can cause increased heart rate, palpitations, or breathlessness. And actions such as bending over or quickly standing up can cause dizziness. These physical sensations can be misinterpreted as a catastrophic cue and result in panic. Individuals who misinterpret these benign sensations are often unaware of doing so and therefore the panic attack appears to come ‘out of the blue’.

The individual’s general beliefs also affect attacks. If an individual believes there is something wrong with his heart, or has ever had heart problems, there will be greater likelihood that palpitations may be interpreted as an indicator of a heart attack. Sensations related to mental functioning can also be misinterpreted such as sudden forgetfulness, confusion or depersonalization. The fear of losing one’s mind may then be triggered which may culminate in a panic attack. Clark further notes that perceiving a sensation as unpleasant is as significant as interpreting it in a catastrophic manner. Some individuals experience voluntary hyperventilation as positive and therefore do not become frightened. Those who perceive it as unpleasant will be likely to interpret it catastrophically.

Clark accepts a biological component to panic attacks in addition to his cognitive model. He believes biology can influence panic attacks in one of three ways. First, some individuals may be more physiologically sensitive and reactive to certain sensations such as breathlessness.  Second, biology may influence the degree of physiological responses to a sense of threat. Third, biology may affect the psychological processes that are involved in the catastrophic misinterpretation process.

Klein – suffocation alarm theory
Klein (1993) proposes a biological model of panic attacks in his suffocation false-alarm theory, claiming many spontaneous panic attacks are related to carbon dioxide (CO2) hypersensitivity. The central nervous system is extremely sensitive to any cues of partial suffocation. The brain, comprising only 2% of body weight but consuming 24% of used oxygen, quickly senses any rise in CO2 as a signal that suffocation is eminent. Elevated levels of CO2 can result either from extreme enclosure and resultant increased levels of inhaled CO2, or from hypoventilation. Either circumstance signals a lack of useful air and triggers a suffocation alarm system. The sense of suffocation is intense and universal, and appropriate in circumstances of actual threat of suffocation, triggering panic and the urge to flee. But in individuals with PD, “a physiologic misinterpretation by a suffocation monitor misfires an evolved suffocation alarm system. . . . followed swiftly by a brief hyperventilation, panic, and the urge to flee” (Klein, 1993, p. 306).

Though an actual rise in CO2 can be a panic stimulus, numerous psychosocial indicators can simulate suffocation and falsely trigger this suffocation alarm, resulting in distressing breathlessness, panic and the urgent need to escape to an open space to breathe. Dyspnea – difficulty in breathing or breathlessness – is practically a universal feature of PD and a frequent trigger of panic. Psychosocial stimuli such as: a sense of entrapment, immobilizing crowds, uncomfortable heat, humidity or stale air can also create a sense of limited exit and restricted capacity to breathe. Klein suggests respiratory disorders may result in PD and that PD occurs frequently with pulmonary disease. His theory holds that cues for panic attacks are external stimuli only, those situational cues arising from the environment. Sensory cues arising from within the body are not considered. Therefore the sight of an imminent uncomfortable situation is considered a cue, while the interoceptive cue of the resultant skip in heartbeat is not.

The role of suffocation anxiety in PD is illustrated by the differences between non-clinical individuals who have experienced an isolated panic attack, and those with PD. In the Norton, Harrison, Hauch & Rhodes study (1985) of young adults who had experienced one or more panic attacks within the preceding year, the most severe symptoms were trembling, sweating and pounding heart, but not difficulty with breathing. Dyspnea is a common symptom in most individuals with PD, but is absent in most with generalized anxiety disorder (Klein, 1993). In his considerable research on the biological etiology of panic, he found that bicarbonate, 5% CO2, lactate and isoproterenol can be panicogenics for those with PD but not in individuals without PD. Following up on Klein’s work with CO2, one study demonstrated that inhalation of 35% CO2/ 65%O2 precipitates symptoms of panic in individuals with PD as well as in those without PD (van den Hout, 1988). Another study reported a challenge of 35% CO2 precipitated panic in 72% of panic patients but only 4% of control subjects. McNally (1994) concludes that “the data are mixed concerning the validity of the hypersensitive carbon dioxide chemoreceptor hypothesis” (p. 57) as an explanation of PD. Although Klein’s model is biological, stating “genetic derangement may be necessary for many spontaneous panics” (1993, p. 314), it does not rule out the possibility that psychosocial factors such as separation anxiety, loss or grief may lower the suffocation alarm threshold.

Klein believes there is a qualitative difference between generalized anxiety or fear, and panic. “Spontaneous panic is not fear, although it is often confused with it” (p. 308). He considers agoraphobia the result of frequent traumatic suffocation panic attacks. His research with the antidepressant imipramine demonstrated efficacy with spontaneous panics but not with chronic anxiety, contributing to his contention that panic does not exist on the same continuum as anxiety but is an independent phenomenon. “Acute anxiety derives from fear as a recently evolved anticipation of danger. Chronic anxiety may be a primitive response to repeated traumatic sensitization. Panic seems an even more primitive response to an endogenously detected danger, specifically suffocation” (p. 308).

Taylor, Bagby and Parker – integrative theory
In addition to their comprehensive theory of affect regulation, the authors offer an integrated biological and psychological model of the etiology of PD. They consider panic to be a disorder resulting from “deficits in the cognitive processing and regulation of emotion” (Taylor, Bagby & Parker, 1997, p. 142), consistent with the alexithymia construct.  They consider panic a flood of dysregulated affect resulting from catastrophic misinterpretations of either internal physical or mental cues, or from external cues. A panic attack is considered a failure in the signal function of anxiety – instead of being perceived as a warning sign it is experienced as a flood of undifferentiated emotion – what Krystal considers “a perfect example of regression in affects to their infantile, somatic, undifferentiated form” (1992, p. 407).

They credit Bowlby’s (1969, 1973) theories of attachment and separation for explaining the process of separation and individuation from the mother and the internalization of her anxiety-regulating functions by the formation of transitional objects. These transitional objects are sensation objects in that they provide olfactory or tactile stimulation and comfort. They later acquire symbolic meaning and encourage development of imaginative ability and interests that increase the child’s capacity for self-regulation of emotions. Alexithymic individuals experience incomplete transitional object formation due to limitations in their capacity for symbolizing, and are therefore prone to continued reliance on sensation objects for soothing. Taylor et al. suggest overeating and substance abuse are adult analogues of this dependence on early sensation objects.

Consistent with Bowlby’s concept of insecure attachment, many adults with PD are very dependent on others as safe people, someone who can help regulate their anxiety and provide a sense of safety. The authors consider this reliance to be compensation for failure in the development of adequate self-regulation:

In summary, our integrated model of panic disorder includes a role for disturbances in early object relationships, which might influence certain biological mechanisms involved in affect regulation, and also influence the internal representations (Bowlby’s working model) of early experience that function as modulators of affect at a cognitive level. In addition, the model places importance on a predisposed person’s current relationships, which might trigger dysregulation of brain stem loci through separations, or function as external regulators that help stabilize hypersensitive loci (Taylor et al., p. 148).

Barsky – Theory of Amplification

The involvement of somatic cues, symptoms, and preoccupation is readily evident in the classification and description of PD as well as its etiological models. Barlow (1988) states “patients with panic disorder seem sensitized to somatic symptoms and overreport and exaggerate these events” (p. 110). Clark’s model is based on the faulty interpretation and consequential dread of somatic cues, and Klein’s theory is predicated upon over reactivity to physiological stimuli that are consistent with suffocation cues. Barsky’s (1992) amplification theory offers insight into the somatic process inherent to PD as well as other somatoform disorders. Somatosensory amplification, or what he simply terms amplification, is the “tendency to experience bodily sensation as intense, noxious, and disturbing. It includes an individual’s disposition to focus on unpleasant sensations and to consider them as pathological rather than normal” (p. 28). He proposes that amplification may play a pathogenic role in PD, as well as in hypochondriasis and major depression. He further states, “Amplification may play a role in the process of somatizing in general” (p. 31). Individuals who are more sensitive and have a low threshold to pain are amplifiers. Those less sensitive and with a higher pain tolerance are termed minimizers.

The amplification process involves three elements. First is hypervigilance to physical sensations and the tendency to focus on those that are unpleasant. Second is over focus on unusual or infrequent sensations. And third is interpretation of sensations as pathological or dangerous as opposed to benign and normal. The full range of physiological sensations can be amplified, from normal bodily functions, to minor and transient abnormalities such as an eye twitch. Or physiological symptoms of affect and symptoms of serious diseases may be amplified.

Amplification shows trait and state characteristics. Some individuals have a lifelong history of sensitivity to bodily sensations which may be learned in childhood or may be a constitutionally inherited trait. Evidence of state-like qualities manifests in the variability of one individual to amplify differently across various circumstances. Four factors can influence this variability. First is cognition, the thoughts one has about a symptom. It is normal to amplify a symptom more when it is attributed to a serious disease rather than a benign cause. Second is the context of the symptoms. If an individual recently had close contact with somebody with tuberculosis, they might react more negatively to their own sudden cough than if they had recently been exposed to someone with a cold. Third is attention to a symptom. “Attention to a symptom amplifies it, whereas distractions diminish it” (p. 29). And fourth is mood. Negative affects tend to increase an individual’s tendency to negatively assess health and symptoms. Depression tends to turn one’s focus inward and cause an individual to think they are defective and damaged. And anxiety causes individuals to catastrophize sensations as well as lower tolerance for unpleasant symptoms. Barsky reports that individuals, who report high levels of negative affects on self-report scales, also tend to report high levels of somatic symptoms (1992).

Construct Validation

Empirical validation of a new psychological construct is critical in confirming the validity of the construct and affirming that it is not a previously identified concept with a new name. “Indeed, Cronback and Meehl (1955) have argued that the validity of a construct cannot be evaluated independently of the tests that purport to measure that construct” (Taylor et al., 1997, p. 46). Toward that goal, several psychometric instruments have been devised to attempt the measure and validation of the alexithymia construct.

An interviewer-completed questionnaire, the Beth Israel Hospital Psychosomatic Questionnaire (BIQ) was developed by Sifneos in 1973 as a tool for the assessment of alexithymia. It has good face validity but several studies suggest inadequate inter-rater and test-retest reliability. Bagby, Taylor and Parker (1994) developed a modified version of the BIQ which enhanced its psychometric properties and correlation with scores on the self-report TAS-20. In the attempt to develop a standardized interview to measure for alexithymia, Krystal, Giller and Cicchetti in 1986 devised the Alexithymia Provoked Response Questionnaire (APRQ). It has high concurrent validity with the BIQ, from which it was developed, but low correlation with the TAS-20. One of the most commonly used self-report instruments for assessing alexithymia has been the MMPI Alexithymia scale (MMPI-A). The empirical criterion method of its scale construction has several flaws and several studies have demonstrated low levels of internal reliability. Both the APRQ and MMPI-A scales provide “no support for the validity of the alexithymia construct and cannot be recommended for clinical or research purposes” (Taylor et al., 1997, p. 51). Other instruments to assess the alexithymia construct, including projective techniques such as the Rorschach alexithymia indices, all lack sufficient validity and internal reliability.

Development of the TAS-20

In the absence of valid and reliable instruments to measure the alexithymia construct, Taylor, Ryan and Bagby (1985) devised the original self-report Toronto Alexithymia Scale (TAS). They used both empirical and rational methods in scale construction, initially defining five domains of alexithymia: (a) difficulty describing feelings, (b) difficulty distinguishing between feelings and accompanying bodily sensations, (c) lack of introspection, (d) social conformity, and (e) impoverished fantasy life and poor dream recall (Taylor et al., 1997). Responses were rated with a 5-point Likert scale ranging from ‘strongly disagree’ to ‘strongly agree’. After factor and item analysis, the 41 item questionnaire was pared to 26 items and four domains that were more theoretically consistent with the alexithymia construct. These four domains, or factors, were: Factor 1 (F1) difficulty identifying and distinguishing between feelings and bodily sensations, Factor 2 (F2) difficulty describing feelings, Factor 3 (F3) reduced daydreaming, and Factor 4 (F4) externally oriented thinking.

While studies demonstrated support for the discriminant and convergent validity of the TAS, and the psychometric properties of the TAS were a considerable improvement over other available instruments, the construction of the TAS prompted refinement of the alexithymia construct and its essential facets. Daydreaming, for example, was determined to negatively correlate with the first factor (Taylor et al., 1997). In 1992 the attempt at scale reconstruction led to the development of a revised, 23 item self-report scale, the TAS-R. Continuing shortcomings with the scale prompted further examination of its compositional structure. There was high correlation between factors 1 and 2, and several items cross-loaded on each factor. Further revision resulted in the TAS-20 (Bagby, Parker &Taylor, 1994; Bagby, Taylor & Parker, 1994), a 20-item self-report scale consisting of 3 factors: Factor 1 (F1) difficulty identifying feelings, Factor 2 (F2) difficulty describing feelings, and Factor 3 (F3) externally-oriented thinking. These three factors represent essential intercorrelated traits that are theoretically congruent with the alexithymia construct. The TAS-20 eliminates the theoretical overlap of the three factors and the cross-loading of items that was a liability in earlier versions, and demonstrates good internal consistency and test-retest reliability. “The psychometric properties of the TAS provide considerable empirical support for the validity of the alexithymia construct” (Taylor et al., p. 49).

Research Studies Using the Original TAS

Zeitlin and McNally
These researchers conducted a study (1993) to assess for alexithymia in patients with PD and patients with obsessive-compulsive disorder (OCD). Referencing Nemiah’s postulation (1984) that alexithymia may be prevalent in those with PD, they noted an absence of any studies to test that hypothesis. Referring to the high incidence of anxiety sensitivity among those with PD, or the fear of anxiety symptoms resulting from a belief they could have harmful consequences, the authors noted Barlow’s proposal (1988) that those with PD may try to avoid experiencing frightening physical sensations by a generalized constriction of emotional experience. They surmised that if a function of alexithymia is a constriction of emotion, individuals with PD should be more alexithymic than those with an anxiety disorder of less anxiety sensitivity such as OCD.

The PD group consisted of 27 new referrals to an anxiety disorders clinic that had been referred by their MD, meeting DSM-III-R criteria for PD. The group comprised 6 men and 21 women with a mean age of 34.4 years (SD=8.6) and age range of 23 to 57. The control group consisted of 31 new referrals to the same clinic that met DSM-III-R criteria for OCD, having been referred by an MD. This group comprised 20 men and 11 women with a mean age of 32.9 years (SD=11.8) and age range of 18 to 64.

All participants completed the original TAS and Anxiety Sensitivity Index (ASI), a 16 item self-report scale that measures fear of anxiety symptoms. Using a 4 point Likert scale, scores can range from 0 (no anxiety sensitivity) to 64 (high anxiety sensitivity), with a normative mean score of 17.8 (SD=8.8). The authors describe reliability and validity data for the ASI as favorable (1993).

The mean TAS score for the PD group was 76.48 (SD=12.36), above the cutoff score of 74 indicative of alexithymia. The mean TAS score for the OCD group was 61.16 (SD=9.82), below the cutoff score of 62 indicative of no alexithymia. The PD group scored higher on the ASI with a mean score of 37.00 (SD=10.20) compared to the OCD group mean score of 27.94 (SD=11.67). The PD group was more alexithymic than the OCD group even when the anxiety sensitivity effect was controlled through ANCOVA (F=16.21, df=1, 55, p<0.001). To measure for group gender effect, TAS and ASI scores were compared by gender for both diagnostic groups by t tests. There was no significant group gender effect demonstrated. Eighteen (67%) of the 27 in the PD group scored at or above the TAS cutoff score of 74, four (13%) of the 31 in the OCD group scored at or above the cutoff of 74.

The study concludes that PD has more association with alexithymia and anxiety sensitivity than does OCD. The authors speculate that higher levels of alexithymia stem from constriction of emotion to minimize related physiological sensation due to high anxiety sensitivity. It remains unclear as to whether alexithymia is primary and precedes PD or is a secondary consequence of PD, and whether the somatic preoccupation which is characteristic of alexithymia may predispose the individual for anxiety sensitivity and subsequent panic attacks.

Parker, Taylor, Bagby and Acklin
In 1993 these researchers published a study that evaluated the hypothesized association between alexithymia and PD, compared to alexithymia in simple phobia. Their rationale for comparing these two disorders is the association of alexithymia with PD but not with simple phobia. They expected a high incidence of alexithymia in individuals with PD due to their somatic preoccupation, difficulty identifying psychological precipitants to panic, and propensity to negatively interpret somatic symptoms of autonomic arousal consistent with the alexithymia construct. They theorized a low incidence of alexithymia in simple phobia as it is a neurotic disorder with greater capacity for regulating anxiety through complex psychological defenses.

The PD group consisted of 30 consecutive referrals to an outpatient hospital based psychiatric clinic in Honolulu, Hawaii who met DSM-III-R criteria for PD. The group comprised eight men and 22 women with a mean age of 35.6 years (SD=5.9) and age range of 25 to 53 years. Twenty (66.7%) of the patients with PD also had agoraphobia. The control group consisted of 32 consecutive referrals to the same clinic who met DSM-III-R criteria for simple phobia. This group comprised 13 men and 19 women with a mean age of 43.6 years (SD=10.4) and age range of 26 to 66 years. The PD group was significantly younger than the control group (t=3.71, df=60, p<0.001). The mean scores on Hollingshead’s educational scale were not significantly different and were within the range that corresponds to middle-class populations with some university education.

All participants completed the TAS and MMPI. The groups were compared on the TAS and selected MMPI scales by two-tailed t tests with a significance level of p<0.007. The PD group scored significantly higher on the TAS and on MMPI scales assessing dependency, anxiety, and somatic complaints. The two groups did not differ significantly on the repression or hysteria MMPI scales. A score of 74 or higher on the original TAS indicates a significant presence of alexithymic traits. Of the PD group, fourteen (46.7%) scored 74 or higher on the TAS, compared to four (12.5%) of the simple phobia group. A score of 62 or lower on the TAS indicates an absence of alexithymic traits. Of the PD group, six (20.0%) scored at or below 62, compared to 21 (65.6%) with simple phobia. There was no measure for group effect by age or gender.

The authors state that the results confirm the hypothesized prediction of higher rates of alexithymia in those with PD compared to those with simple phobia. They state that the 46.7% rate of alexithymia in the PD sample   is higher than the rate of 37% in a general psychiatric outpatient sample that was found in one unpublished study. The 12.5% rate found in the simple phobia sample compares to the rate found in another study of normal college students and adults (Taylor, Bagby & Parker, 1993). The authors were unable to determine whether alexithymia was an antecedent of panic or secondary to the effect of anxiety on these individuals. Given the rate of alexithymia found in the PD group, the authors concur with Nemiah (1984) that every individual seeking treatment for panics should be assessed for the presence of alexithymia as it has significant implications for treatment.

Joukamaa and Lepola
This Finnish pair of researchers conducted a study (1994) to test for alexithymia in individuals with PD, compared to alexithymia in a normal control group. Their purpose was to “ascertain how alexithymia and panic disorder are interrelated” (p. 33). Noting the strong association between alexithymia and somatic disorders,   psychosomatic disorders and chronic pain, they comment on the scarcity of studies of the association between alexithymia and psychiatric disorders like PD.

The PD group consisted of 50 consecutive referrals to a private outpatient clinic in Finland that were diagnosed with PD by one of the researchers. The group comprised 12 men and 38 women with a mean age of 41.7 years (SD=7.3). The control group consisted of 50 healthy students at a nursing college without noteworthy somatic or psychiatric disturbances. The group comprised 5 men and 45 women with a mean age of 28.7 years (SD=6.8). No age range data was provided on either group. The difference in mean age of each group was statistically highly significant. And the difference in social status and education levels between the groups was also considered statistically significant.

All participants completed the TAS and the General Health Questionnaire (GHQ), a 36 item self-report scale used as a measure of neurotic type psychopathology. In the PD sample, 12 (24%) were considered “true alexithymics” due to a score at or above the cutoff of 74, 20 (40%) were considered not alexithymic due to scores at or below the bottom cutoff of 62, and 18 (36%) were “probable alexithymics” due to a score between the cutoffs. The control group had no true alexithymia, and 6 (12%) were “probable”, a highly significant difference. Of the PD group, 34 (68%, p<0.001) were assessed as “psychiatric cases” on the GHQ, compared to 11 (22%) of the control group. Both alexithymia and psychiatric disorder had highly significant association with age and socioeconomic stratum, specifically oldest age and lowest socioeconomic level.

The authors conclude that alexithymia and psychic disturbance had a highly significant association in the PD sample. They consider 60% to be alexithymic (true and probable), 68% to be psychiatric, and 48% to be both, with no alexithymia within the control sample. Using both multidimensional ANOVA and logistic regression analysis, “panic disorder was highly significantly associated with alexithymia but not with psychic disturbance” (p. 35). They identify a limitation of the study in the small sample size.

Research Studies Using the TAS-20

Cox, Swinson, Shulman and Bourdeau

Citing the studies of Zeitlin and McNally (1993) and Parker, Taylor, Bagby & Acklin (1993) that found elevated levels of alexithymia in PD, these researchers conducted a study (1995) to test the hypothesis that only one dimension of the alexithymia construct may be responsible for this elevation. Cox et al. cited the need to investigate a “conceptual and measurement issue” (p. 195) and posited an alternate hypothesis to the reputed link between alexithymia and PD. They identified overlap between Clark’s (1986) catastrophic misinterpretation model, the anxiety sensitivity construct (Reiss, Peterson, Gursky & McNally, 1986), and the alexithymic dimension of difficulty identifying and differentiating emotions from physical sensations. These researchers hypothesized that a PD sample would score high only on the TAS factor (F-1) related to difficulty identifying emotions and differentiating them from physical sensations. They did not believe there would be significant correlation with the other TAS factors: difficulty describing emotions (F2) and externally oriented thinking (F3). They chose to use another anxiety disorder as a control group, social phobia. Their rationale was two-fold. Individuals with social phobia and PD experience anxiety in various circumstances as compared to simple phobia. And while both groups experience very intense anxiety, PD involves “endogenous” anxiety attacks, whereas social phobia involves “exogenous” attacks precipitated by external circumstances.

The PD group consisted of 100 outpatients at an anxiety disorders clinic that were diagnosed by a psychiatrist or clinical psychology doctoral student and met DSM-III-R criteria for PD. Levels of assessed agoraphobia among this group were: 14% without agoraphobia, 34% mild agoraphobia, 41% moderate agoraphobia, and 11% severe agoraphobia. The group comprised 33 men and 67 women with a mean age of 34.28 years (SD=9.82). The control group consisted of 46 outpatients at the same clinic that met DSM-III-R criteria for a diagnosis of social phobia. The group comprised 23 men and 23 women with a mean age of 32.46 years (SD=8.01). No age range data was provided on either group.

All participants completed the TAS-20, ASI, Panic Attack Questionnaire, Beck Anxiety Inventory, and Symptom Checklist 90 – Somatization Subscale. The mean TAS-20 results revealed no significant between group differences on any of the three TAS-20 factors. The TAS-20 cutoff scores are: 61 or greater – alexithymia; 51 or less – not alexithymia; 52 to 60 – probable alexithymia. Based on this scoring, 34.0% of the PD group was alexithymic, 28.3% of the social phobia group was alexithymic, a non- significant difference. However in both groups the identifying emotions factor, F1, was significantly higher than F2 or F3. And in both groups difficulty describing emotions, F2, was higher than externally oriented thinking, F3. The researchers also correlated the PD group’s TAS-20 results with the other anxiety assessment scales and found highly significant correlation between them and the TAS-20’s F1 and weak correlation with F2 and F3.

Cox et al. (1995) state that alexithymia levels of 34.0% in the PD group and 28.3% in the social phobia group are not only a non-significant difference, but approximate the prevalence figures of alexithymia in the general outpatient psychiatric population. They suggest that the tendency for those with PD to score high on the F1 trait can lead to many false positives on the TAS. They further speculate that the social phobia level of alexithymia may be relatively high and close to the PD level as both groups experience panic attacks that may be reflective of the F1 trait of difficulty identifying emotions.

Fukunishi, Kikuchi, Wogan and Takubo
Noting the preceding three cross-sectional studies of alexithymia prevalence in PD samples, these researchers published a longitudinal follow-up study (1997) of alexithymia in PD and social phobia samples. In a 6 month follow-up study they examined three factors: the prevalence of alexithymia in these two groups, the effect of psychiatric treatment on alexithymic traits, and what psychiatric factors are associated with a decrease in alexithymia.

The PD group consisted of 26 patients admitted to the psychiatric inpatient unit or outpatient unit at a Tokyo hospital, who met DSM-IV criteria for PD. The group comprised 12 men and 14 women with a mean age of 31.7 years, and age range of 22 to 49 years. The mean education was 13.5 years with a range of 10 to 16 years. The social phobia group consisted of 24 patients at the same inpatient and outpatient facility, who met DSM-IV criteria for social phobia without panic attacks. The group comprised 9 men and 15 women with a mean age of 33.7 years, and age range of 18 to 47 years. The mean education was 12.9 years with a range of 8 to 16 years. For a healthy control group, they recruited 25 individuals living in the Tokyo metropolitan area. The group comprised 10 men and 15 women with a mean age of 33.0 years, and age range of 20 to 48 years. The mean education was 13.4 years with a range of 11 to 16 years. Using ANOVA, chi-square and student two-tailed t tests, the demographic variables of age, educational level and gender were demonstrated to have non-significant variances.

All participants completed the TAS-20, the Hamilton Depression Scale (HDS) and Hamilton Anxiety Scale (HAS). Six months after initiation of treatment, participants again completed the TAS-20. Before initiation of treatment, 14 (53.8%) of the 26 individuals in the PD group, and 14 (58.3%) of the 24 in the social phobia group were alexithymic based on meeting or exceeding the TAS-20 cutoff score for alexithymia. The difference between these two groups was considered a non-significant difference. After treatment, the alexithymia prevalence rates were: 8 (30.8%) in the PD sample and 8 (33.3%) in the social phobia sample, again a non-significant difference. Treatment reduced the TAS-20 scores in 23 (88.5%) of the 26 in the PD group, and in 20 (83.3%) of the 24 in the social phobia group. The alexithymia prevalence rate in the healthy control group of 25 was 4 (16%) with a TAS-20 score of 52 or greater, a significant difference from the two patient groups. After psychiatric treatment there were no significant differences between the three groups.

The total TAS-20 scores were positively and significantly correlated with the HDS and HAS scores for all three groups. After treatment, there were no significant differences on the TAS-20, HDS, or HAS scores for either patient group. Also after treatment, patients who were still alexithymic (n=16) showed significantly higher HAS scores than those who were not alexithymic (n=44). The results were considered not significantly influenced by gender, age or educational level.

Noting that 23 of the 26 (88.5%) patients with PD, and that 20 of the 24 (83.3%) with social phobia showed a decrease in TAS-20 scores, the authors suggest that alexithymic traits can be reduced by psychiatric treatment for anxiety. They also found that scores on both F1 and F2, though not for F3, were reduced with treatment. They suggest that these alexithymia components are related to anxiety, whereas F3 (external thinking) may not be. The significant level of alexithymia in both patient groups (before treatment) compared to the control group, suggests alexithymia exists in many with PD and social phobia.

Fukunishi et al. (1997) reference the state or trait question about the nature of alexithymia. Freyberger (1977) suggests two types of alexithymia: primary which is a personality trait, and secondary which is a state reaction. They state that in contrast to Salminen, Saarijarvi, Aarela & Tamminen (1994) who found alexithymia to be a trait instead of a state, their study found an overall decrease in alexithymia by psychiatric treatment suggests alexithymia to be a reversible state reaction.

Bankier, Aigner and Bach
In 2001 these researchers published a comparative assessment of alexithymia levels in several psychiatric disorders: somatoform, panic, obsessive-compulsive, and in depression. They note inconsistent evidence of a correlation between alexithymia, PD, and psychosomatic symptoms, and reference a study (Bach, Bach & deZwaan, 1996) that suggests alexithymia and somatization to be separate but simultaneously occurring constructs. Similarly conflicting evidence is cited regarding a correlation between alexithymia and depression, including some evidence suggesting each exists independent of the other.

The sample groups consisted of 234 patients admitted over a four year period to the Behavior Therapy Ward at the Department of Psychiatry, University of Vienna, Austria. All diagnoses were made by use of the Structured Clinical Interview for DSM-IV. The PD group of 123 individuals consisted of 48 men and 75 women with a mean age of 36 years (SD=11.3). The somatoform disorder (SOM) group of 24 individuals consisted of 3 men and 21 women with a mean age of 38 years (SD=8.6). The obsessive-compulsive disorder (OCD) group of 59 individuals consisted of 34 men and 25 women with a mean age of 32 years (SD=9). The depression group of 28 individuals consisted of 10 men and 18 women with a mean age of 38 years (SD=11). The mean age for all participants was 35.3 years (SD=10.7) with an age range of 18 to 71 years. Of all 234 participants, 139 (59%) were women, and 95 (41%) were men.

The mean TAS-20 total score of 49 (SD=11.4) for the PD group was lower than the SOM group score of 54.9 (SD=12.4), the OCD group score of 51.9 (SD=8.7), and the depression group score of 56.6 (SD=13.6). Of the TAS-20 subfactors, F1 was significantly associated with SOM and depression; F2 with depression; F3 with OCD; while PD had a significant negative correlation with F3. Age was a non significant variable with total TAS-20 scores or any of the three subfactors. Female gender was significantly associated with F1 (difficulty identifying feelings and distinguishing them from bodily sensations) and male gender was significantly associated with F3 (externally-oriented thinking). There was significant correlation between lower educational level and TAS-20 total score and F3.

The authors note the contrast between this study and that of Zeitlin and McNally  (1993) who found significantly higher prevalence of alexithymia in a PD sample compared to an OCD sample. This study found the reverse, possibly the result of a larger sample size. The earlier study also did not investigate the multidimensionality of the alexithymia construct with a multi-factorial analysis. These researchers reference the debate as to the diagnostic value of total TAS-20 alexithymia scores, and note greater diagnostic specificity in a multifactorial application of the TAS-20. They state that the three TAS-20 subfactors can provide valuable diagnostic information about cognitive and affective aspects of various psychiatric disorders.

Conclusions and Intent of this Research Study

The alexithymia construct involves complexities of cognitive and affective processing deficits and possesses heuristic value for conceptualizing those deficits. Bowlby’s model (1969, 1973) of insecure attachment provides clarification of the etiology of limited internalized affect regulating capacity evident in the alexithymia construct. The alexithymia construct is analogous to Krystal’s (1988) model of affective regression to a somatized and undifferentiated state and is consistent with the Lane and Schwartz (1987) model of affective delay at or regression to an earlier epigenetic level of development. These developmental arrests appear to coincide with the somatization involved in the F1, and the affective verbalization deficits involved in F2, of the TAS-20.

The alexithymia construct also appears to be consistent with the salient features of PD. The somatic preoccupation, intense autonomic arousal, and affective dysregulation involved in Barlow’s false alarm (1988) and Klein’s suffocation alarm (1993) models are conceptually congruent with the somatic preoccupation of alexithymia. Clark’s (1986) catastrophic misinterpretation theory and Barsky’s (1992) amplification theory are consistent with the cognitive and affective deficits contained in the alexithymic model and appear to correlate with the TAS-20 F1 deficits in differentiating emotions from corresponding physiological sensations. By that misinterpretation and confusion between the psyche and soma, Barlow’s (1988) ‘deranged’ false alarm is triggered.

While alexithymia has been associated with several psychiatric disorders including somatoform, PTSD, OCD, depression, eating disorders and substance abuse, it would seem that alexithymia might have an even higher correlation with PD. The degree and velocity of the regression that is a panic attack is unique and extreme, while the extent of personality disorganization and affect intensity is acute. Alexithymia has been described as having a “cognitive analogue in the thought disorders of schizophrenics” (Taylor et al., 1997, p. xiii), a comparison that could also be drawn with panic attacks and the inherent, though transient, disconnection from reality that they comprise.

The data from the six research studies on the correlation between alexithymia and PD has been inconsistent. It was suggested by Bankier, Aigner and Bach (2001) that some of the inconsistency results from comparisons between studies which rely on a TAS-20 total alexithymia score only, and studies that use the total score in addition to scores from the three TAS-20 subfactors. As alexithymia is a multidimensional construct, a multifactorial analysis provides greater specificity as to the gradational involvement of each factor in a disorder.

It is therefore the purpose of this study to measure and compare TAS-20 total alexithymia scores and subfactor scores of a PD sample compared to a sample comprised of a cross-section of outpatient psychiatric diagnoses. The multiple diagnoses group was selected for a cross-section of disorders that have been associated with the alexithymia construct. The rationale is to investigate the prevalence of alexithymia in PD compared to a wide spectrum of disorders that may have variable degrees of association with the construct. Analysis of the TAS-20 subfactor scores is considered crucial in evaluating the contribution of each factor to the overall alexithymia score. It is also considered essential to measure for significant group effect by age or gender.

The purpose of this research was to test the hypothesis that alexithymia is more significantly associated with PD than with other mental health diagnoses. Alexithymia has been associated with disorders that manifest panic, depression, anxiety, substance abuse, poor impulse control, PTSD and phobia; so this study used an experimental group of individuals with PD, and a control group of individuals with a cross-section of mental health diagnoses characterized by other symptoms associated with alexithymia. Given the extent of somatization and confusion between physiological and psychological sensations in PD, this study also tested for significant differences between the groups on F1 of the TAS-20 in addition to testing for differences on F2 and F3. Also tested for was a significant variance in alexithymia prevalence between panic disordered individuals with, and those without, agoraphobia. It was also considered important to test for significant alexithymia differences by age and gender.


This total study consisted of 60 participants from the ages of 17 to 63. All the participants in this retrospective research project were drawn from my own clinical caseload at a freestanding but hospital affiliated outpatient mental health clinic. The clinic is located in a predominately Caucasian, middle class Detroit suburban community. All participants in this study voluntarily sought mental health treatment, and were diagnosed according to DSM-IV criteria.

The experimental PD group and control group each comprise 30 individuals. The experimental group consisted of 16 individuals with PD without agoraphobia, and 14 with PD and agoraphobia. A diagnostic breakdown of the control group is as follows: Adjustment Disorder with Depression and Anxiety (4); Alcohol Dependence (2); Anxiety Disorder (1); Dysthymic Disorder (10); Generalized Anxiety Disorder (GAD) (7); Posttraumatic Stress Disorder (PTSD) (3); Social Phobia (2); Specific Phobia (1).

The experimental group consisted of 18 men and 12 women, with an age range of 17 to 63 and a mean age of 37.0. The control group consisted of 15 men and 15 women, with an age range of 23 to 63 with a mean age of 38.2. Composition by race of the experimental group was: 1 African American and 29 Caucasian. Composition by race of the control group was: 1 African American and 29 Caucasian.

Participant Selection

In January 2002, I began administering the TAS-20 to all new clients who, in their presentation at intake, cited symptoms associated with the alexithymia construct. I continued to administer the scale until I obtained valid scales on 30 individuals who met DSM-IV criteria for PD, and 30 individuals who met DSM-IV criteria for disorders characterized by clinical features that have been conceptually associated with the alexithymia construct. Those 30 with a diagnosis of PD (with or without agoraphobia) constituted the experimental group, and those 30 with other diagnoses that have been associated with the alexithymia construct constituted the control group.


The test used for this research study is the psychometrically improved revision of the original Toronto Alexithymia Scale (TAS). The revised version, the TAS-20, is a 20 item self-report instrument that assesses alexithymia (Bagby, Parker & Taylor, 1994). A copyright fee was paid to the test originators for a copy of the test, scoring, and information packet concerning construction, reliability and validity. It has a three-factor structure that is theoretically consistent with the alexithymia construct. The three factors comprise three subscales: Factor 1 (F1) – difficulty identifying feelings; Factor 2 (F2) – difficulty describing feelings; and Factor 3 (F3) – externally oriented thinking.

The TAS-20 was developed using a combined rational and empirical method of scale construction. The reliability and validity of the TAS-20 has been supported by factor analysis, good internal consistency (Cronbach’s alpha = 0.81), and high test-retest correlations (r = 0.77; p < 0.01) over a 3-week period, consistent with the trait perspective of alexithymia. The TAS-20 also demonstrates high correlations with the Minnesota Multiphasic Personality Inventory-2 and Beth Israel Hospital Questionnaire. Consensual validity was also demonstrated by a positive correlation (r = 0.53; p = < 0.01) between observer ratings of alexithymia and TAS-20 scores in a psychiatric population (Parker, Bagby, Taylor, Endler & Schmitz, 1993). The TAS-20 authors indicate studies have supported the convergent and discriminant validity of the instrument, which they state also provides validity of the alexithymia construct (Taylor, Bagby & Parker, 1997).

A 1993 study by Parker et al. found that men tend to have significantly higher TAS-20 scores than do women. This contrasts with three earlier studies that used the original TAS, which demonstrated no significant gender difference (Bagby, Taylor & Atkinson, 1988; Parker, Taylor & Bagby, 1989; Taylor, Parker, Bagby & Acklin, 1992). Men may demonstrate a higher degree of alexithymia than women since the processing of emotion involves lateralized cerebral functions and women appear to have a lesser degree of hemispheric specialization than men. This would also be consistent with studies that have demonstrated that women communicate information about health problems and emotions more effectively than men (Parker et al., 1993). Age has been demonstrated to have a weak correlation with total TAS-20 scores.

The TAS-20 utilizes a five-point Likert scale with five of the items inversely scored. It is hand scored with a maximum score of 100. It uses cutoff scoring: equal to or less than 51 = non-alexithymia, equal to or greater than 61 = alexithymia. Scores of 52 to 60 = possible alexithymia. The maximum scores for each of the subscales are: Factor 1 (7 items): 35; Factor 2 (5 items): 25; Factor 3 (8 items): 40. There are no cutoff scores established for each of the three factor subscales. The following are the 20 TAS items, numbered in sequence as they appear on the test, but grouped by factor:

F1 – Difficulty Identifying Feelings

1.   I am often confused about what emotion I am feeling.
3.   I have physical sensations that even doctors don’t understand.
6.   When I am upset, I don’t know if I am sad, frightened, or angry.
7.   I am often puzzled by sensations in my body.
9.  I have feelings that I can’t quite identify.
13. I don’t know what’s going on inside me.
14.  I often don’t know why I am angry.

F-2 – Difficulty Describing Feelings

2.   It is difficult for me to find the right words for my feelings.
4.   I am able to describe my feelings easily.
11. I find it hard to describe how I feel about people
12. People tell me to describe my feelings more.
17. It is difficult for me to reveal my innermost feelings, even to close friends.

F-3 – Externally-Oriented Thinking

5.   I prefer to analyze problems rather than just describe them.
8.   I prefer to just let things happen rather than to understand why they turned out that way.
10. Being in touch with emotions is essential.
15. I prefer talking to people about their daily activities rather than their feelings.
16. I prefer to watch “light” entertainment shows rather than psychological dramas.
18. I can feel close to someone, even in moments of silence.
19. I find examination of my feelings useful in solving personal problems.
20. Looking for hidden meanings in movies or plays distracts from their enjoyment.

Note: Items 4, 5, 10, 18, and 19 are inversely keyed.


Participants signed informed consent forms both for treatment as well as participation in this study. (See Appendix A for complete Consent Form). Informed consent for a minor was obtained from both the minor and parent. Participants were asked to complete the TAS-20 after an explanation was provided regarding the construct being measured as well as the purpose of this study. As the participants were entering psychotherapy with me, I explained the value of evaluating for possible involvement of alexithymia, both for their treatment as well as for this research project. The TAS-20 was administered during the first history-taking, initial evaluation session, before treatment was initiated. I was responsible for administering and scoring all tests. The total alexithymia score is the sum of responses to all 20 items, while the score for each subscale factor is the sum of the responses to that subscale.

Data analysis was conducted using the Statistical Package for Social Sciences (SPSS) for Microsoft Windows 11.0 software. Tests conducted included Analysis of Variance (ANOVA), Analysis of Covariance (ANCOVA), and the Students t-test for Independent Samples. The total alexithymia score and the three individual factors (F1, F2, F3) that make up the score were analyzed using a two-way ANCOVA. The between factors are group (Control versus Panic Disorder) and gender (Men versus Women) and the covariate variable is age. Age is used as a covariate in order to control for possible confounding of age with the main effects of group and gender on the outcome measures.

There is a significant main effect of the group variable of Control vs. Panic Disorder, (F=13.8, df=1,55, p =0.001) with means of 50.5 and 61.6 respectively. The main effect of gender is not significant (p=0.21), the means for males and females are 58.4 and 53.1 respectively. However, we do obtain a very significant group by gender interaction effect (F=5.8, df=1,55, p=0.02). The covariate age is also significant (F=4.9, df=1,55, p=0.03) The R-square for the age variable is 0.31. This indicates that over 30% of the variability in the TAS-20 score can be explained by age differences.

In order to examine the observed significant group by gender interaction we conducted separate ANCOVAs for each gender using group as the between factor and age as the covariate. Among the male subjects we do not obtain a significant difference for the group variable (p=0.29). The means for the Controls and the Panic Disorder Groups are 55.5 and 60.9 respectively. Among the female subjects, we do observe a very significant effect of the group variable (F=28.3, df =1,24, p=0.0001). The means for the Control and Panic Disorder groups are 45.5 and 62.7 respectively.

Examining the individual TAS-20 factors revealed a significant group effect for Factor 1 (F1) – (F=19.7, df=1,55, p=0.001) with means of 17.3 and 24.3 for Controls and Panic Disorder groups respectively.  The R-square is 0.29 for the age covariate on F1. The group effect on Factor 1 (F1) is observed to be significant for both Males and Females (p=0.014 and p=0.001 respectively).

For Factor 2 (F2) we do not see a group effect (p=0.10). However, we do observe a significant gender difference on F2 (F=4.6, df=1,55, p=0.037). The means for the males and females on F2 are 15.8 and 12.7 respectively. There is also a gender by group interaction effect on F2 (F=5.1, df=1,55, p=.03). As seen in the T.A.S. and F1, we also observe a significant group effect on F2 in the female subjects (F=13.4, df=1,24, p=0.001), but not in the male subjects. The covariate age is also significant (F=9.3, df=1,55, p=.004). The R-square is 0.27 for the age covariate on F2. The means on Factor 2 (F2) are 10.9 and 15.1 for the Control and Panic Disorder groups respectively.

On Factor 3 (F3) we do not observe either group or gender effects. However a significant group by gender interaction is observed. (F=5.1, df=1,55, p=0.03). The mean score for the females are 17.5 and 22.1 for the controls and panic disorder groups respectively, resulting in a significant increase of 4.6 points (F=7.5, df=1,24, p=0.01). While in the males the difference is not significant (p=0.74) with scores of 22.1 and 21.6, a difference of only 0.5 points. The age covariate is not significant on F3.

The data derived from this study provide evidence of a significant relationship between alexithymia and Panic Disorder, and support for the theory that the component of alexithymia most significantly associated with alexithymia is the difficulty in identifying feelings and differentiating between emotions and concomitant physiological sensations (F1 on the TAS-20). The purpose of this discussion is to provide interpretation of the study results, evaluate their implications and theoretical consequences, as well as address limitations of the study.

The Hypotheses

Hypothesis 1
There is a significant relationship between alexithymia and PD.

The data support the first hypothesis that there is a significant group difference in alexithymia levels between the experimental (PD) group and the control group. While there is no significant main effect of the group variable of men vs. women in total alexithymia scores with respective mean scores of 58.4 and 53.1, there is a very significant group by gender interaction effect. This is demonstrated by the significant elevation of alexithymia in the PD group women compared to control group women, but non-significant increase in alexithymia in the PD group men compared to control group men. This indicates that the significant PD group vs. control group difference in alexithymia is strictly the result of the very significant difference in alexithymia scores between the PD group women and control group women.

The TAS-20 alexithymia cut-off score of 61 is reached by the PD Group with a mean score of 61.6. Of the 30 individuals in the PD group, 16 (53.3%) score in the alexithymic range, 8 (26.7%) are in the possibly alexithymic range of 52 to 60, and 6 (20%) are in the non-alexithymic range. Of the 16 participants in the PD group who reach the alexithymia cutoff score, 8 (50%) are women and 8 (50%) are men. But the 18 men in the PD group comprise 60% of the PD sample while the 12 women comprise 40%.

Over half of the PD group evidences alexithymia and another quarter of the group is in the possibly alexithymic range. Only one-fifth of the PD group is non-alexithymic.

Hypothesis 2
There is no significant relationship between alexithymia and other DSM-IV diagnostic groups that have been conceptually associated with the alexithymia construct.

The data support the null hypothesis that there is no relationship between alexithymia and other diagnostic categories that are associated with the alexithymia construct. The control group fails to achieve the TAS-20 cutoff score of 61 with a mean score of 50.47, below the non-alexithymia cutoff score of 51. Of the 30 control group participants, 8 (26.7%) score in the alexithymic range, 3 (10%) are in the possibly alexithymic range, and 19 (63.3%) are in the non-alexithymic range. Of the 8 in the alexithymic range, 6 (75%) are men, and 2 (25%) are women. The control group is comprised of 15 (50%) men and 15 (50%) women.

The control group as a whole evidences a lack of alexithymia with two-thirds of its individual participants scoring in the non-alexithymic range. Only one-quarter of the control group evidences alexithymia.

Hypothesis 3
There is no significant difference in alexithymia between PD with Agoraphobia and PD without Agoraphobia.

The data support the hypothesis that there is no significant group difference in alexithymia levels between PD with Agoraphobia, and PD without Agoraphobia. Participants with agoraphobia (N=14) have a mean total alexithymia score of 63.5 and participants without agoraphobia (N=16) have a mean total alexithymia score of 59.9, a non-significant difference. Of the 14 with agoraphobia, 7 (50%) reached the TAS-20 cutoff score of 61. Of the 16 without agoraphobia, 9 (56%) reached the cutoff score of 61. Theoretical consideration for these results will be discussed later in this section under the topic of avoidant type “Safety Behaviors”.

Hypothesis 4
There is a significant relationship between the TAS-20 Factor 1 and PD.

The data also support the hypothesis of a significant group effect for F1 with means of 24.3 for the PD group and 17.3 for the control group, a significant difference, and it is the only TAS-20 subscale with significant between group difference. It is also the only subscale without significant gender variance, with mean scores of 25.5 and 23.4 respectively for the PD group women and PD group men, and 17.0 and 17.6 respectively for the control group women and control group men.


Gender as a Covariable

The significant effect by gender is consistent with the literature on the design and validation of the TAS-20 (Bagby, Parker & Taylor, 1994). It is noted that the derivation sample demonstrated a “small but statistically significant difference between the mean TAS-20 score for men . . . and the mean score for women” (p. 27). Another study by members of this group stated “a significant main effect was found for gender … with men scoring higher than women” (Parker, Bagby, Taylor, Endler & Schmitz, 1993, p. 227). That study also found no gender effect for the TAS-20 subscale F1, but found significant gender effect for F2 and F3, with men scoring higher than women.

This current research study found gender to have a considerable effect on total TAS-20 alexithymia scores, and on the F2 and F3 subscale scores. The only component of the TAS-20 that was not influenced by gender was subscale F1.

Implications of Gender Differences

While this study finds a significant group difference in alexithymia prevalence between a PD sample and a control sample, that overall difference is attributable to the women in the study. The women in the control group are non-alexithymic with a TAS-20 mean score of 45.5 (51 or less = non- alexithymia), while the men in the control group fall into the possibly alexithymic range with an overall mean score of 55.5 (52 to 60 = possible alexithymia). Both the men and women in the PD group are alexithymic with mean scores of 60.9 and 62.7 respectively (61 or greater = alexithymia). Only the difference between the control group women and PD group women is significant, with mean scores of 45.5 and 62.7 respectively. The non-significant difference between the men is 55.5 and 60.9 for the control and PD groups respectively.

The gender differences demonstrated by this study are consistent with the common perception that women tend to be more emotionally attuned than men, and are consistent with the findings of earlier studies. While the control women were non-alexithymic, the control men were possibly alexithymic. The gender effect in this study may facilitate identification of the most prominent alexithymic trait in the etiology of PD. It may also serve to differentiate the relative involvement of each alexithymic trait in PD.

The only dimension in which there were no significant gender differences – no gender interaction effect – is on the TAS-20 F1 subscale. This is the only portion of the TAS-20 that reflected a common effect of PD on both men and women.

These results would seem to suggest that the alexithymic trait most significant in the etiology of PD involves difficulty with affect identification and differentiation of affects from their attendant physiological sensations. The TAS-20 F1 subscale of difficulty in identifying emotions and differentiating between an emotion and its physiological sequelae is clearly associated with somatization, resonating with Nemiah’s original formulation of PD as “the prototypical psychosomatic disorder” (1984, p. 134).

Gender Effect in Earlier Studies

Of the six earlier cited studies that measured alexithymia in PD samples compared to control samples, three found no effect by gender on TAS-20 overall scores while the other three did not test for gender effect. One study (Zeitlin & McNally, 1993) noted that while “men are often more alexithymic than women” (p. 659) their results did not show a gender effect on TAS-20 scores. That study did not conduct a subfactorial analysis to separate results on the three TAS-20 subscales.

TAS-20 Subscale Results in Earlier Studies

Conducting subfactorial analysis facilitates differentiation of the relative involvement of each factor in the etiology or manifestation of PD. Of the six previous studies of alexithymia in PD compared to a sample, three measured the multidimensionality of the alexithymia construct by separate evaluation of the three TAS-20 subscales. The other three only considered overall TAS-20 alexithymia scores. One study (Bankier, Aigner & Bach, 2001) reported no gender effect on overall TAS-20 scores but found that female gender was significantly associated with Factor 1 and male gender was significantly associated with Factor 3 of externally oriented thinking. A 1995 study (Cox, Swinson, Shulman & Bourdeau) found a highly significant association between Factor 1 and a PD group (n=100), but also found a similar association between F1 and a social phobia control group (n=46). This same study did not measure for a gender effect.

Age Effect

A two-way ANOVA on age found a significant negative correlation between age and alexithymia scores. As age increases, alexithymia scores tend to decrease. Older individuals demonstrate lower alexithymia levels, possibly reflecting a positive correlation between increased self-awareness and advanced age.

Age Effect in Earlier Studies

Of the six previously cited studies, one found an age effect while two did not and the other three did not test for age effect. The one study that found an age effect described a highly significant positive correlation between advanced age and alexithymia scores in individuals with both alexithymia and a “psychiatric disturbance” as determined by the General Health Questionnaire (GHQ-36) (Joukamaa & Lepola, 1994). The additional covariate of another psychiatric diagnosis might confound the evidence for an independent age effect.

Evidence of Alexithymic Traits in the Early Treatment of PD

The significant presence of alexithymia in individuals with PD is supported by my clinical experience in treating these individuals. From the inception of treatment to the point of termination, a primary focus of treatment is reducing alexithymic features by increasing affect awareness and affect regulation. The cognitive and affective deficits that constitute alexithymia are prominent in the process that facilitates entry into treatment and become a focal point in the early phase of treatment.

Many of these individuals do not self-identify an emotional difficulty and seek relief by contacting a psychotherapist. Instead, a sizeable percent have a history of identifying somatic concerns and seeking out medical solutions. Often this includes extensive use of medical facilities including their own primary care physician (PCP) and emergency medical services such as hospital emergency rooms. According to a 1984 study (Katon), 6% to 10% of patients in primary medical care actually have PD, with somatic complaints that tend to be neurological, cardiological and/or gastrointestinal. Other diagnostic categories such as substance abuse are also strongly associated with a circuitous route to treatment, often through legal channels or through exasperated family members.

Individuals with PD often come to psychotherapy after extensive use of medical services in the vain attempt to find symptom relief. They seek psychotherapeutic interventions out of their own frustration with the limited relief obtained from medical services, or by referral from the medical provider. Even before the first appointment in a therapist’s office, there is often a lengthy history of frustrated efforts to find medical solutions that is usually not present with individuals seeking therapy for other diagnoses. This type of faulty self-appraisal and related behavior would seem to be closely associated with the TAS-20 F1 subscale: difficulty identifying feelings and differentiating between emotional sensations and physical sensations. The tendency for somatization predisposes the panic disordered individual to overuse of medical services. The data from this study is consistent with this observation in the elevation on F1 for both panic disordered males and females. This finding is also consistent with Clark’s hypothesis (1986) that panic attacks are the consequence of the catastrophic misinterpretation of strange or uncomfortable physical sensations.

When panic disordered individuals present for treatment, the initial symptoms they present focus on their intense anxiety as well as on the somatic components of their symptoms, and there may be a degree of disorganization in the presentation. Alexithymic traits are evident in their confusion between emotional and physiological symptoms, and frequent uncertainty as to whether they are experiencing a medical or psychological crisis. While they are generally certain that they are extremely anxious, they are often uncertain as to whether the anxiety is the central problem or is a symptom of another more insidious physical disorder. Their affective awareness is generally limited to intense anxiety over another panic attack and/or the strong fear of the possible catastrophic implications of another panic attack. They may or may not be aware of the considerable depression that is often comorbid to PD. There is often limited awareness of the precipitants of their initial attacks; with an over focus on the external circumstances that surround initial and subsequent attacks, similar to the TAS-20 F3 subscale of externally oriented thinking.

While their recollection of their first panic attack is usually quite vivid, it is not uncommon that recollection of even earlier panic attacks emerges later in treatment.  Difficulty identifying emotional issues and affect laden events that preceded the onset of their first panic attacks appears to be directly related to the traits that define the alexithymia construct. Quite often the onset of panic attacks is related to separation fears or significant life events: death, illness, geographical move, relationship breakup, pregnancy/childbirth, injury, accident, or threat of any of these. It is an early marker for the presence of alexithymia when a client, during the initial evaluation period, struggles to identify any such issue or related affect as a precipitant to the onset of panic, only to have it surface later in therapy – a frequent occurrence in my work with these individuals.

Often several sessions are necessary before the individual begins to develop insight into the underlying psychosocial dynamics that contributed to their initial attack(s). This insight only develops with consistent psychotherapeutic probing which underscores the patient’s limited capacity for achieving these insights without assistance, consistent with the cognitive deficits of alexithymia. Insight into the underlying issues responsible for the individual’s first panic attack is vital to the treatment of PD as this insight challenges their previously unproductive formulation that their anxiety is appropriate dread of a physiological catastrophe. This insight can facilitate identification of the alexithymic traits that precipitate and perpetuate panic attacks especially the tendency toward external thinking, somatic preoccupation, and difficulty identifying and verbalizing feelings and their causes. After a period of insight-oriented therapy, the individual often seems to develop a better capacity to elucidate the precipitating events of panic attacks, as well as their emotional sequelae. This could be an argument for the mutability of alexithymia – that it can be mitigated with insight-oriented, psychodynamic or cognitive psychotherapy.

Insight seems to develop in tandem with decreasing anxiety, reflecting a circular relationship. As anxiety decreases, the capacity for higher level cognitive functioning is enhanced; and the development of cognitive mastery contributes to reduced anxiety. Individuals with PD who seek treatment tend to minimize the extent and significance of anxiety disorders or symptoms in family of origin members. Often at least one parent has or had panic attacks or is a chronic anxious worrier, typically involving health or safety issues. The PD individual may have limited awareness of the impact of this role-modeling, or possible genetic loading, on their own anxious tendencies. It is common for PD individuals to evidence limited separation and emancipation from a parent, often the mother, and often the similarly anxious parent.  This tends to manifest in frequency of contact, by telephone or in person, not uncommonly daily or even multiple daily contacts. This symbiosis represents the lack of internalized affect regulating capacity consistent with the cognitive and affective deficits of alexithymia. The lack of an internalized comforting maternal object contributes to perpetual reliance on the original object as a comforting source. The characteristic reliance on a safe person by individuals with PD has its roots in this lack of internalized comforting and affect regulation, reflecting Bowlby’s (1973) formulation of proximity-maintaining behavior in insecure or anxious attachment patterns.

The Role of Alexithymia in the Perpetuation of PD

The cognitive and affective deficits of alexithymia involved in the etiology of PD are also involved in the maintenance of the disorder. A vital dilemma exists as to the chronicity of PD: how is it that even hundreds of trials of an unreinforced catastrophic fear can resist extinction? Seligman (1988) wondered why individuals would continue to misinterpret panic attacks as heart attacks after multiple experiences to the contrary. In response, Salkovskis (1988) theorized that these individuals continue to catastrophically misinterpret sensations because they believe they avert catastrophe with protective maneuvers. He suggests these individuals engage in subtle avoidance or safety behaviors that undermine the test of the catastrophic hypothesis. Clark (1988) recognized the impact these preventative measures have on sustaining irrational fears: if someone has the erroneous thought that they may faint and therefore quickly sit down to avoid falling, they miss an opportunity to disprove their false belief.

Another explanation is that individuals refuse to reject a catastrophic hypothesis for reasons described by the philosopher Hume (1748, 1977) who suggested that a generalization cannot be proven by a finite number of positive occurrences. Because someone has not driven off a bridge during a dizzy spell one hundred times, does not mean they could not drive off it during the next dizzy spell. While panic disordered individuals may begin to accept the lack of probability that their catastrophic fear will materialize, it is ultimately an act of faith when they do so (McNally, 1994).

It is my proposition that the deficits of alexithymia are central to the perpetuation of PD as they support reliance on the use of safety behaviors to avert catastrophic outcomes. The safety behaviors in turn support these cognitive and affective deficits and inhibit the testing of catastrophic hypotheses. In other words, I believe there exists a circular relationship between alexithymia and safety behaviors, each sustaining the other and thereby PD. This relationship appears to be closely related to the fear of fear cycle which Goldstein and Chambless termed “the most central phobic element” (1978, p. 51).

Limited awareness of faulty, irrational cognitions and resultant catastrophic fears perpetuates reliance on safety behaviors to avoid catastrophe. Many of these maneuvers are conscious efforts to avoid catastrophe, but many seem to have developed without conscious participation. A part of the uncovering process of insight and cognitive therapy is the revealing of these behaviors, their intent, and their positive and negative consequences. It is of significance that these behaviors often have a magical thinking or superstitious quality to them, suggesting their primitive nature. It appears to me that alexithymic deficits are prominent in interfering with the conscious awareness and evaluation of these behaviors and thereby perpetuate them. The catastrophic hypothesis these behaviors seemingly prevent remains untested and the fear of fear cycle continues.

The Role of Safety Behavior

Psychotherapeutic success with individuals with PD involves identification of irrational cognitions, fears, and safety behaviors, so as to facilitate conscious testing of underlying catastrophic hypotheses. This becomes possible when alexithymic traits are adequately mitigated to facilitate this conscious process. In the event that underlying cognitive and affective processes do not become conscious, hundreds of further unreinforced trials may continue without significant increase in the likelihood of spontaneous extinction. In my clinical experience with panic disordered individuals, it appears that remission of panic attacks proceeds from amelioration of the cognitive and affective deficits of alexithymia.

Examples of Safety Behaviors

The degree to which safety behaviors contribute to the perpetuation of panic level anxiety is considerable and widely recognized. They are similar to the avoidance behaviors of simple phobia or agoraphobia and can be explained by operant or classical conditioning models. An operant model suggests that safety behaviors can best be understood as freely emitted behaviors that increase or decrease relative to their consequence or value. If any behavior reduces anxiety and seems to prevent a panic attack, heart attack or other catastrophe, it is reinforced and likely to be repeated. A classical conditioning model suggests that anxiety has become associated with certain stimuli and the avoidance of these stimuli is reinforced by the termination of the anxiety state.

In simple phobias or agoraphobia, avoidance behavior reinforces fear of the avoided object or situation by precluding desensitization via exposure to the feared stimuli. Similarly, safety behaviors reinforce fear of external or interoceptive cues by avoiding a test of the catastrophic hypothesis associated with those cues. Safety behaviors are reinforced by their capacity to terminate aversive anxiety states but contribute to the continuation of an anxiety disorder. This study found no significant difference in alexithymia prevalence between the PD with agoraphobia individuals, and the PD without agoraphobia individuals. I suggest this is the result of their difference without a difference. That is, both these groups employ avoidance behaviors with identical function and consequence but with apparently insignificant surface distinctions.

One type of safety behavior is avoidance. This includes the set of behaviors that constitute agoraphobia, but also include more subtle avoidance behaviors manifested by individuals with PD either with or without agoraphobia. Often the difference in the avoidant behaviors demonstrated by individuals within these two different diagnostic categories is the degree of subtlety, or symbolization, of the behavior. It is not accurate to consider only the individual with PD and agoraphobia as a phobic avoider, and one with PD without agoraphobia as a non-avoider. Each group generally manifests some degree or variant of avoidant safety behaviors, blurring the clear distinction between the two diagnostic categories.

Panic disordered individuals with agoraphobia manifest classic symptoms and behaviors that constitute agoraphobia. Anxiety over situations and places of limited escape, or unavailability of assistance in the event of panic, cause them to avoid going far from home or leaving home altogether. Their anxiety can prompt them to avoid crowds, certain modes of transportation, or standing in line. These and similar situations are avoided or are endured with marked anxiety (American Psychiatric Association, 1994).

But there are other avoidance behaviors that do not qualify as classic agoraphobic manifestations as they constitute subtler forms of avoidance. They are more subtle as they are selective, discriminating forms of avoidance as compared to the more generalized agoraphobic forms of avoidance. Specific aspects of anxiogenic situations may be avoided such as avoiding one particular stretch of road as opposed to avoiding all driving situations. Specific bridges, tunnels, roads, or driving conditions may be avoided, while others are not. Anxiogenic physiological sensations such as dizziness, breathlessness, or accelerated heart rate may be avoided by avoiding substances or activities that produce them. These may include avoidance of: caffeine; sugar; amusement park or playground rides; physiological exertion such as sports activities or exercise; or even the witnessing of others participating in these behaviors.

A significant variant of avoidant-type safety behavior employed in the presence or absence of agoraphobia, is the avoidance of being alone via the use of a safe person. The safe person is an external comforting source, typically a spouse or first degree relative such as a parent. The overt reliance on a comforting individual suggests delay in the development of internalized, comforting parental introjects and the continuation of proximity-maintaining behavior (Bowlby, 1973). The dependence on a safe person also represents significant lack of “differentiation of the self” – what Bowen (1978) considered the highest level of maturational development.

The agoraphobic individual may be obvious about their resistance or refusal to leave home without a safe person. The non-agoraphobic individual may manifest this behavior in symbolized form involving the use of an object representation or transitional object. Instead of requiring the presence of a safe person, there may be reliance on a cell phone that represents immediate connection with a comforting source. A transitional object can be any object imbued with comforting value – a picture, a ‘good luck charm’, or any object with representational or talisman value. This transitional object can be similar to a child’s but tends to have more symbolic meaning as opposed to comforting sensory properties like a child’s security blanket or stuffed animal.

Closely related to the safe person is the concept of a safe place. The agoraphobic is unwilling to leave their only true safe place – their home, but may do so with a safe person or alone but with considerable anxiety. For non-agoraphobics, a safe place is a place with representational value – places away from home wherein the person is in proximity to a source of perceived safety or assistance. Examples are a hospital, a restroom or freeway rest area, or a friend or relative’s home.

Another form of safety behavior is distraction. An individual experiencing an internal or external anxiety cue while driving a car may turn up the radio, open a window, sing aloud, or force conversation with anyone else in the car in an attempt to distract their attention from the cue and incipient catastrophic cognitions. Physiological maneuvers such as shifting in one’s seat, muscle clenching, or moving about can be attempts to distract one’s attention from incipient cues for panic.

Another type of safety behavior is the use of escape, or taking flight from an anxiogenic situation. These behaviors constitute fleeing from a store or mall, leaving the line at a cash register, or pulling off a road or freeway. Often the flight is from a situation that represents suffocation or restriction of breathing, to one that represents a capacity for unrestricted breathing. Examples are: leaving an enclosed or crowded room, escaping a traffic jam or backup, going to an open door or window or going outside for fresh air.

Another set of safety behaviors involve the use of relaxation techniques to avoid a panic attack. These responses include deep breathing, progressive relaxation techniques, and meditation. While there is a very legitimate use of these methods during the early phase of treatment of PD, they become potentially counterproductive measures when they become the solution to panic.

Most of these safety behaviors have preliminary value for the panic disordered individual in their attempts to find ways to abort the fear of fear cycle. Behaviors that constitute mastery of relaxation techniques have an inherent generalized value that can contribute to a legitimate sense of self-control and affect regulation. But relaxation techniques, like other safety behaviors, become counterproductive and self-limiting if they interfere with identification and resolution of the underlying process of panic. These behaviors can interfere with the identification of internal and external anxiogenic cues, and the amplification and catastrophic misinterpretation that follow. These safety behaviors not only support the alexithymic deficits that impede the awareness of and challenge of false catastrophic hypotheses, they also promote the alexithymic tendency to consider a problem as physiological that can be ameliorated with palliative physiological interventions.

Case Illustration of Safety Behavior and Alexithymia

A case illustration that revealed to me the powerful role of safety behavior and its relationship to alexithymia, involved a woman in her mid-thirties diagnosed with PD without agoraphobia. Panic attacks that she had experienced on a frequent basis for several years were in remission for several months after approximately one year of insight and cognitive therapy, with me as her therapist. Despite the absence of panic attacks, she revealed in a session that she still experienced significant anticipatory anxiety each morning when first leaving her home. She complained of not understanding why she still remained so anxious on a daily basis when her panic attacks had been absent for many months.

In my attempt to help reveal the chain of negative cognitions that precede the anxiety she experiences each morning when she leaves her home, she could only identify that her head “feels foggy.”  Not only could she not identify specific cognitions, she could not identify specific affects and was only able to identify a vague somatic sensation. In pursuing identification of more specific cognitions or affects, she said “I feel afraid of how I feel.” With further exploration, she identified a fear that her “anxiety could escalate . . . could overwhelm me . . . be out of control . . . go crazy”. She was surprised to realize that everyday she had the identical sequence of thoughts and feelings and yet never recognized them on her own. It was evident that despite these fears, she had developed adequate coping strategies to preclude catastrophizing these thoughts to the point of panic. She asked me an important question that essentially asked why she continues to experience the same thoughts and feelings each morning when nothing bad ever happens.

Her question prompted me to explore for the existence of safety behaviors that may have been preventing further panic attacks, but were interfering with identification and modification of negative cognitions. She was able to reveal that each morning she employs the use of deep breathing and muscle clenching while driving her car, and her strong feeling that these techniques were preventing further panic attacks. After we established more rational cognitions that challenged her irrational ones, I wrote down her chain of negative cognitions on a 3” x 5” index card along with the positive, rational alternate cognitions. She was given the homework assignment: to review this chain of irrational thoughts and the rational alternate cognitions, each morning before she left her home. And she was to refrain from using her safety behaviors. At her next appointment two weeks later, she stated her index card was dog-eared as she takes it with her wherever she goes, though only read it a few times as she remembers vividly what it says. She reported a dramatic decrease in her anxiety to the point it had been negligible for that entire two week period. She said, “just being aware of it [the card and what it said] stops it in its tracks [her anxiety] . . . keeps it from building . . . keeps it logical . . . stops it from mounting.” She also said that she felt comforted by carrying around that index card, “I know its there.” She stated the main benefit of recounting her chain of negative thoughts, and the positive alternate thoughts, each morning came from “identifying it [her anxiogenic thoughts] versus pushing it away which doesn’t help”.

What occurred?
This woman evidences the alexithymic trait congruent with the TAS-20 F1 subscale when she identifies a physiological feeling in lieu of an emotional feeling in her description of her emotions as a “foggy” head sensation. She had difficulty congruent with the TAS-20 F2 subscale of difficulty verbalizing emotions, evident in her limited vocabulary for affects. Her paucity of affects was also evidenced during these sessions by her wooden affect and frequent blank look and stare while discussing these powerful emotional experiences. She also demonstrated considerable lack of affective and cognitive awareness as she had previously been unaware of the underlying, catastrophic cognition of her fear of going crazy. While her safety behaviors facilitated a decrease in panic attacks, they were also counter productive to the extent they kept her irrational thoughts out of her consciousness, and therefore unavailable to evaluation, challenge and potential extinction.

What helped?
It appeared that the identification of her core catastrophic cognition was most instrumental in the remission of her anxiety. Identifying this facilitated the cognitive evaluation and challenge of her catastrophic hypothesis, and conscious daily test of it without reliance on safety behaviors that could render these thoughts unconscious.

What are other possible explanations?
While the presence of alexithymic traits and safety behavior is evident in this case illustration, there are other possible explanations for the decrease in anxiety that resulted from my interventions. She may have found comfort in the transitional object value of the index card, having been hand-written by me and representing my omnipresence during her trips away from home. This possibility would imply employment of another safety behavior, albeit one with more symbolic value and suggestive of hierarchical advancement in affect regulation (Taylor, 1987). Also, given her alexithymic traits, she may have been sustaining herself without panics by reliance on other, as of yet unidentified, safety behaviors. Lastly, a placebo effect may be responsible for progress, derived from a sense of safety or nurturance resulting from my involvement, or our expectation that a homework assignment was to be effective.

My Formulation of the Role of Alexithymia in the Etiology of Panic Disorder
My formulation of panic disorder and the role of alexithymia in its etiology constitute a biopsychosocial model

Biological components
Biological deficits may be significant to the alexithymic difficulties experienced by those with psychosomatic disorders, which I believe includes PD, in identifying and verbalizing affects. These deficits contribute to the tendency to discharge affects somatically instead of cognitively (MacLean, 1949; Nemiah, 1984). There is also considerable support for the theory of disruption in communication between the emotional right hemisphere and verbal left hemisphere (Hoppe, 1981). The biological alarm triggered by a sense of suffocation seems to be frequently involved in PD. Many panic attacks involve the fight or flight response, manifested by the attempt to escape from a sense of enclosure and restricted breathing, such as crowds, stores or traffic jams (Barlow, 1988; Klein 1993). I give much consideration to Barlow’s conceptualization of a biological vulnerability for panic attacks in the face of the multi-generational presence of PD in families.

Psychosocial components

Alexithymia seems to result from biological deficits and deficits in personality development, rather than stemming from unconscious conflicts – consistent with a deficit model rather than a defense model. Bowlby’s attachment theory (1969, 1973) accurately describes the impact of early life attachment patterns in the clinging behavior manifested by adults with PD, exemplified by the manifestation of a safe person. I see evidence of his proposition that anxiety is activated attachment behavior in the many safety behaviors of PD that seek to avoid separation and maintain proximity. Bowlby’s model provides explanation for the high rate of occurrence of PD in multiple generations. An anxiously attached parent seems to be high risk for replicating a faulty attachment pattern with their child, and/or model anxiety for their child.

My formulation of affect development is predicated upon Krystal’s (1988) infantile precursor states of contentment or distress, and the development of adult affective states via differentiation, verbalization and desomatization. I believe Krystal is precise in his conceptualization of alexithymia as arrest at or regression to an earlier undifferentiated and somatized affective level. This formulation has considerable heuristic value for conceptualizing psychosomatic disorders such as PD.

Also of considerable theoretical value is the Lane and Schwartz (1987) epigenetic model of cognitive and affective development that explains development of affects proceeding from increasing verbalization and differentiation. I find value in their hierarchical model especially for its capacity to explain the somatic traits of PD as they represent arrest at, or regression to, an early somatized affective level.

It appears to me that somatosensory amplification is generally involved in the panic process (Barlow, 1988; Barsky, 1992). The amplification theory provides considerable explanation for the dread of normal physiological sensations by individuals with PD, and why normal sensations are often experienced as noxious sensations by these individuals. The amplification concept suggests reasons for the overwhelming power of sensations and why individuals with PD feel they may be overwhelmed by their physiological experiences and lose control of themselves. The amplification concept provides insight into the somatization process as it explains how somatic sensations can become the primary or even exclusive focus of attention. The significant elevation on the TAS-20 F1 subscale (difficulty identifying emotions and differentiating them from physiological sensations) by the PD men and women in this study appear to be evidence of the involvement of somatosensory amplification in individuals with PD.

In my judgment, the dynamic of greatest prominence in the etiology of PD is that of the catastrophic misinterpretation (Clark, 1986). It seems to me that the catastrophic hypothesis is the spark that ignites the ingredients of biological factors, cognitive and affective deficits of alexithymia, life events and somatic sensations into the explosion of the panic reaction or panic attack. But I believe the term attack is a misnomer, suggesting as it does a spontaneous event, for a panic attack is more accurately a reaction to multiple factors and events with the catastrophic misinterpretation or hypothesis being the flashpoint. The cognitive and affective deficits of alexithymia and the principles of classical and operant learning theories become prominent in the chronicity of panic attacks as they support the “fear of fear cycle”.

Limitations of this Study

Limitations to this present research seem to fall into three categories: (a) sample characteristics, (b) instrumentation, and (c) theoretical limitations. Limitations of the sample selection include small sample size. Inferences drawn from small samples make it questionable to generalize these results to a larger population. The significant effect of gender might make it beneficial to conduct similar research with more homogeneous populations to reduce gender by group interaction and develop better understanding of the reasons for such significant gender differences in alexithymia, in both PD groups as well as control samples. The age range of this study’s participants is another limitation. As this study found a decrease in alexithymia with advanced age, future studies with more homogeneous by age samples could provide clearer data as to alexithymia prevalence between groups without the confounding effect of age.

This current research may have limitations due to instrumentation. The TAS-20 is a self-report measure designed for this type of research and is currently the most accepted instrument for these purposes. It was therefore deemed the most appropriate psychometric tool for this research project. But it is well-known that self-report measures may not provide an accurate portrayal of reality and may unintentionally reflect bias on the part of the participant. Another limitation is the absence of data from a second instrument or other measure of alexithymia. While at this time there are no other self-report instruments with adequate construct validity, standardized observer-ratings of alexithymia could provide cross comparison of data.

Other limitations of this research are a function of the topic of research itself. There are inherent limitations to research on abstract concepts such as alexithymia. A tester must rely on accuracy and honesty in the participant’s self-report of traits such as affective awareness and verbalization, and tendencies for somatization. It is uncertain if a participant with limitations in the area of objective self-awareness can therefore be expected to produce an accurate self-portrayal.

Treatment Implications

This current research yielded results that lend themselves to important treatment implications. The positive correlation between PD and somatic traits inferred by the TAS-20 F1 subscale, in both men and women, has implications of value. That these are the only consistent alexithymic traits distinguishing both genders with PD from controls, it can be implied that these traits require specific treatment interventions. The alexithymic deficits of PD suggest a primitive affective level of functioning wherein the individual operates from an early non-verbal, somatized level of affective development. Treatment interventions need to facilitate functioning at higher developmental levels via affective identification and verbalization to effect the desomatization of affects.

This current research yields data that contribute to the growing body of knowledge concerning the role of alexithymia in the etiology of psychosomatic and other disorders of affect regulation. The data support the heuristic value of the alexithymia construct in conceptualizing and explaining some of the deficits involved in disorders characterized by affect dysregulation. The results of this current research are consistent with findings of earlier studies while at odds with the results of others.

Zeitlin and McNally (1993) found greater association between alexithymia and PD than with OCD. Parker, Taylor, Bagby and Acklin (1993) found higher rates of alexithymia in those with PD compared to those with simple phobia. And Joukamaa and Lepola (1994) found more alexithymia in those with PD than in a healthy control group. The results of this current research are consistent with these studies in finding greater correlation between alexithymia and PD than with other disorders.

On the other hand, Cox, Swinson, Shulman and Bourdeau (1995) found a non-significant difference in alexithymia between a PD sample and a social phobia sample. Fukunishi, Kikuchi, Wogan and Takubo (1997) found a non-significant difference between PD and social phobia samples. And Bankier, Aigner and Bach (2001) found lower alexithymia scores for a PD sample compared to somatoform, obsessive-compulsive and depression samples.

While research results may be inconsistent, or even contradictory, the current research results are consistent with the author’s experience in treating those with PD compared to other mental health disorders. The degree of affective unawareness in psychosomatic disorders such as PD is commensurate with the degree of somatization and tendency for catastrophic explanations of somatic experiences. The F1 subfactor on the TAS-20 isolates this trait and identifies the difficulty in distinguishing between emotions and related physiological sensations. It would be expected that there would be significant elevation on F1 by those with PD compared to controls, and this current research confirms this. Some of the inconsistent findings by earlier studies are related to a lack of subfactorial analysis.

The most prominent implications of this current research derive from two sources. First is the significant involvement of the TAS-20 F1 deficits in PD for both genders. And second is the considerable gender difference in the prevalence of alexithymia. Given the extent of difference between genders in the manifestation of alexithymic traits, it is all the more noteworthy that the only area of congruence between the genders is in the elevation of F1 scores in PD. The implication is that the F1 deficits – difficulty identifying emotions and differentiating them from their related physiological sequelae – are the alexithymic component most involved in PD. The somatization process inherent in this component underscores Nemiah’s (1984) conceptualization of PD as a psychosomatic disorder. The Krystal model (1988) as well as the Lane and Schwartz model (1987) suggest the origin of the somatization in PD to be the arrest at or regression to an earlier, non-verbal and somatized level of affect development. The implications for treatment then would appear obvious – that affects need to become more well identified and verbalized to become more cognitive and less somatic.

Directions for Future Research

The results of this current research provide additional empirical support for the significance of certain alexithymic traits in the etiology of PD. These current research hypotheses seem to be appropriate directions for future research given the large body of literature supportive of the theory outlined in the literature review. It would be of value for future research to address some of the limitations of this current research that are noted earlier. It would be beneficial for future research to be conducted with larger sample groups, possibly controlling for age and gender using a narrower age range or single gender studies.

Given the multi-faceted nature of the alexithymia construct, future studies might include additional self-report measures or observer-rating instruments designed to isolate each of the alexithymic traits to provide more information as to which traits are most prominent to PD. Observer-ratings could also address the inherent problem of the accuracy of self-reported data. Valuable future research might involve the use of pre and post testing of specific alexithymic traits to obtain information as to the interconnectedness of traits and whether they respond in tandem or independently to treatment interventions. Additionally, pre and post testing might demonstrate mutability of alexithymic traits, and thereby influence the state or trait controversy regarding the nature of alexithymia.

Closing Statement

“Each is liable to panic, which is exactly, the terror of ignorance surrendered to the imagination. Knowledge is the encourager, knowledge that takes fear out of the heart, knowledge and use, which is knowledge in practice. They can conquer who believe they can. It is he who has done the deed once who does not shrink from attempting again.”
– Ralph Waldo Emerson (1803-1882)

This current research offers validation of the above quotations premise, that ignorance provides opportunity for the imagination to induce fear while knowledge can quell it. The cognitive and affective deficits of alexithymia constitute ignorance – or rather the lack of self-knowledge – that allows the imagination to generate panic attacks. This research provides support for the hypothesis that the aspect of alexithymia most responsible for chronic panic attacks is the limited capacity for identifying emotions and the tendency to confuse them with physical sensations. The cognitive therapy modality strives to increase self knowledge – thereby reduce alexithymic traits – and has demonstrated the greatest success of any modality in reducing the frequency and intensity of the panic attacks of PD. Cognitive therapy aims to facilitate progression in the epigenetic model of affect development, upward from early levels of alexithymic traits to higher levels of desomatized, verbalized affects. Ultimately, the goal of treatment is the facilitation of the successful completion of normal developmental epigenesis – from alexithymia to lexithymia.

  • American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders, 4th ed. Washington DC: American Psychiatric Association.
  • Bach, M., Bach, D. & deZwaan, M. (1996). Independency of alexithymia and somatization: A factor analytic study. Psychosomatics, 37, 451-58.
  • Bagby, R.M., Parker, J.D.A. & Taylor, G.J. (1994). The twenty-item Toronto Alexithymia Scale – I: Item selection and cross-validation of the factor structure. Journal of Psychosomatic Research, 38, 23-32.
  • Bagby, R.M., Taylor, G.J. & Atkinson, L. (1988). Alexithymia: A comparative study of three self-report measures. Journal of Psychosomatic Research, 32, 107-16.
  • Bagby, R.M., Taylor, G.J. & Parker, J.D.A. (1994). The twenty-item Toronto Alexithymia Scale – II: Convergent, discriminant, and concurrent validity. Journal of Psychosomatic Research, 38, 33-40.
  • Bankier, B., Aigner, M. & Bach, M. (2001). Alexithymia in DSM-IV disorder: Comparative evaluation of somatoform disorder, panic disorder, obsessive-compulsive disorder, and depression. Psychosomatics, 42, 235-40.
  • Barlow, D.H. (1988). Anxiety and its disorders. New York: Guilford Press.
  • Barlow, D.H., Cohen, A.S., Waddell, M.T., Vermilyea, B.B., Klosko, J.S.,
  • Blanchard, E.B. & DiNardo, P.A. (1984). Panic and generalized anxiety disorders: Nature and treatment. Behavior Therapy, 15, 431-49.
  • Barsky, A.J. (1992). Amplification, somatization, and the somatoform disorders. Psychosomatics, 33, 28-34.
  • Bion, W.R. (1962). Learning from experience. London: Heinemann.
  • Bion, W.R. (1965). Transformations. London: Heinemann.
  • Bowen, M. (1978). Family therapy in clinical practice. New York: Jason Aronson.
  • Bowlby, J. (1969). Attachment and loss: Attachment. New York: Basic Books.
  • Bowlby, J. (1973). Attachment and loss: Separation. New York: Basic Books.
  • Bretherton, I. (1985). Attachment theory: Retrospect and prospect. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research: Monographs of the society for research in child development (vol. 50, pp. 3-35). New York: Basic Books.
  • Cairns, R.B. (1966). Attachment behavior of mammals. Psychological Review, 73, 409-26.
  • Cairns, R.B. & Johnson, D.L. (1965). The development of interspecies social preferences. Psychonomic Science, 2, 337-8.
  • Clark, D.M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24, 461-70.
  • Clark, D.M. (1988). A cognitive model of panic attacks. In S. Rachman & J.D. Maser (Eds.), Panic: Psychological perspectives (pp. 71-89). Hillsdale, NJ: Erlbaum.
  • Cox, B.J., Swinson, R.P., Shulman, I.D. & Bourdeau, D. (1995). Alexithymia in panic disorder and social phobia. Comprehensive Psychiatry, 36, 195-98.
  • Craske, M.G. (1991). Phobic fear and panic attacks: The same emotional states triggered by different cues? Clinical Psychology Review, 11, 599-620.
  • Crittenton, P.M. (1994). Peering into the black box: An exploratory treatise on the development of self in young children. In D. Cicchetti & S.L. Toth (Eds.), Rochester symposium on developmental psychopathology: Disorders and dysfunctions of the self  (vol. 5, pp. 79-148). Rochester, NY: University of Rochester Press.
  • Edgcumbe, R.M. (1984). Modes of communication: The differentiation of somatic and verbal expression. Psychoanalytic Study of the Child, 39, 137-54.
  • Emde, R.N. (1984). The affective self: Continuities and transformations from infancy. In J.D. Call, E. Galenson & R.L. Tyson (Eds.), Frontiers of infant psychiatry (vol. 2, pp. 38-54). New York: Basic Books.
  • Erickson, E.H. (1956). The problem of ego identity. Journal of the American Psychoanalytic Association, 4, 56-72.
  • Fonagy, P. (1991). Thinking about thinking: Some clinical and theoretical considerations in the treatment of a borderline patient. International Journal of Psychoanalysis, 72, 639-56.
  • Freyberger, H. (1977). Supportive psychotherapeutic techniques in primary and secondary alexithymia. Psychotherapy and Psychosomatics, 28, 337-42.
  • Fukunishi, I., Kikuchi, M., Wogan, J. & Takubo, M. (1997). Secondary alexithymia as a state reaction in panic disorder and social phobia. Comprehensive Psychiatry, 38, 166-70.
  • Furman, E. (1992). On feeling and being felt with. Psychoanalytic Study of the Child, 47, 67-84.
  • Goldberg, S., MacKay-Soroka, S. & Rochester, M. (1994). Affect, attachment and maternal responsiveness. Infant Behavior and Development, 17, 335-9.
  • Goldstein, A.J. & Chambless, K.L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 47-59.
  • Greenspan, S.I. (1981). Psychopathology and Adaptation in Infancy and Early Childhood. New York: International Universities Press.
  • Harlow, H.F. (1961). The development of affectional patterns in infant monkeys. In B.M. Foss (Ed.), Determinants of infant behaviour, Vol. 1. New York: Wiley.
  • Heiberg, A. & Heiberg, A. (1977). Alexithymia – an inherited trait? A study of twins. Psychotherapy and Psychosomatics, 28, 221-5.
  • Hobson, R.P. (1994). On developing a mind. British Journal of Psychiatry, 165, 577-81.
  • Hoppe, K.D.  & Bogen, J.E. (1977). Alexithymia in twelve commisurotomized patients. Psychotherapy and Psychosomatics, 28, 148-55.
  • Horney, K. (1952). The paucity of inner experiences. American Journal of Psychoanalysis, 12, 3-9.
  • Horton, P.C., Gewirtz, H. & Kreutter, K.J. (1994a, April). Alexithymia and panic disorder. American Journal of Psychiatry, 151, 627-8.
  • Horton, P.C., Gewirtz, H. & Kreutter, K.J. (1994b, July). Panic disorder and alexithymic: No overlap. American Journal of Psychiatry, 151, 1099.
  • Hume, D. (1748, 1977). An enquiry concerning human understanding. Indianapolis: Hackett.
  • Joukamaa, M. & Lepola, U. (1994). Alexithymic features in patients with panic disorder. Nordic Journal of Psychiatry, 48, 33-36.
  • Katan, A. (1961). Some thoughts about the role of verbalization in early childhood. Psychoanalytic Study of the Child, 16, 184-8.
  • Katon, W. (1984). Panic disorder and somatization: Review of 55 cases. American Journal of Medicine, 77, 101-6.
  • Kelman, N. (1952). Clinical aspects of externalized living. American Journal of Psychoanalysis, 12, 15-23.
  • Klein, D.F. (1993). False suffocation alarms, spontaneous panics, and related conditions: An integrative hypothesis. Archives of General Psychiatry, 50, 306-17.
  • Kopp, C.B. (1989). Regulation of distress and negative emotions: A developmental view. Developmental Psychology, 25, 343-54.
  • Krystal, H. (1962). The opiate withdrawal syndrome as a state of stress. Psychiatric Quarterly, 36, 53-65.
  • Krystal, H. (1968). Massive psychic trauma. New York: International Universities Press.
  • Krystal, H. (1975). Affect tolerance. Annual of Psychoanalysis, 3, 179-219.
  • Krystal, H. (1979). Alexithymia and psychotherapy. American Journal of Psychotherapy, 33, 17-28.
  • Krystal, H. (1988). Integration and self-healing: Affect, trauma, alexithymia. Hillsdale, NJ: Analytic Press.
  • Krystal, H. (1992). Psychoanalysis as a ‘normal’ science. Journal of the American Academy of Psychoanalysis, 20, 395-412.
  • Krystal, H., Giller, E.L. & Cicchetti, D.V. (1986). Assessment of alexithymia in posttraumatic stress disorder and somatic illness: Introduction of a reliable measure. Psychosomatic Medicine, 48, 84-94.
  • Krystal, H. & Raskin, H. (1970). Drug Dependence. Detroit: Wayne State University Press.
  • Lane, R.D. & Schwartz, G.E. (1987). Levels of emotional awareness: A cognitive-developmental theory and its application to psychopathology. American Journal of Psychiatry, 144, 133-43.
  • Lazarus, R. (1984). On the primacy of cognition. American Psychology, 39, 124-126.
  • Lazarus, R. (1991). Emotion and Adaptation. New York: Oxford University Press.
  • Lorenz, K.Z. (1935/1957). ‘Der kumpan in der unvelt des vogels.’ J. Orn. Berl., 83. English translation in C.H. Schiller (Ed.), Instinctive Behavior. New York: International Universities Press.
  • MacLean, P.D. (1949). Psychosomatic disease and the ‘visceral brain’: Recent developments bearing on the Papez theory of emotion. Psychosomatic Medicine, 11, 338-53.
  • Malatesta, C.Z. (1990). The role of emotions in the development and organization of personality. In R.A. Thompson (Ed.), Nebraska symposium on motivation (vol. 36): Socioemotional development (pp. 1-56). Lincoln: University of Nebraska Press.
  • Marty, P. & de M’Uzan, M. (1963). La ‘pensee operatoire’. Revue Francaise de Psychanalyse, 27, 1345-56.
  • McDougall, J. (1974). The psychosoma and the psychoanalytic process. International Review of Psychoanalysis, 1, 437-59.
  • McDougall, J. (1984). The “dis-affected” patient: Reflections on affect pathology. Psychoanalytic Quarterly, 53, 386-409.
  • McNally R.J. (1994). Panic Disorder: A critical analysis. New York: Guilford Press.
  • Nemiah, J.C. (1975). Denial Revisited: Reflections on psychosomatic theory. Psychotherapy and Psychosomatics, 26, 140-7.
  • Nemiah, J.C. (1984). The psychodynamic view of anxiety. In R.O. Pasnace (Ed.), Diagnosis and treatment of anxiety disorders (pp. 117-137). Washington DC: American Psychiatric Press.
  • Nemiah, J.C. (1997). Alexithymia: Theoretical considerations. Psychotherapy and Psychosomatics, 28, 199-206.
  • Nemiah, J.C., Freyberger, H. & Sifneos, P.E. (1976). Alexithymia: A view of the psychosomatic process. In O.W. Hill (Ed.), Modern trends in psychosomatic medicine (vol. 3, pp. 430-9). New York: Appleton-Century-Crofts.
  • Nemiah, J.C. & Sifneos, P. E. (1970). Affect and fantasy in patients with psychosomatic disorders. In O.W. Hill (Ed.), Modern trends in psychosomatic medicine (vol.2, pp. 26-34). New York: Appleton-Century-Crofts.
  • Norton, G.R., Harrison, B., Hauch, J. & Rhodes, L. (1985). Characteristics of people with infrequent panic attacks. Journal of Abnormal Psychology, 94, 216-21.
  • Parker, J.D.A., Bagby, R.M., Taylor, G.J., Endler, N.S. & Schmitz, P. (1993). Factorial validity of the 20-item Toronto Alexithymia Scale. European Journal of Personality, 7, 221-32.
  • Parker, J.D.A., Taylor, G.J. & Bagby, R.M. (1989). The alexithymia construct: Relationship with sociodemographic variables and intelligence. Comprehensive Psychiatry, 30, 434-41.
  • Parker, J.D.A., Taylor, G.J., Bagby, R.M. & Acklin, M.W. (1993). Alexithymia in panic disorder and simple phobia: A comparative study. American Journal of Psychiatry, 150, 1105-07.
  • Ramzy, I. & Wallerstein, R.S. (1958). Pain, fear, and anxiety. The Psychoanalytic Study of the Child, 13, 147-90.
  • Razran, G. (1961). The observable unconscious and the inferable conscious in current Soviet psychophysiology: Interoceptive conditioning, semantic conditioning, and the orienting reflex. Psychological Review, 68, 81-147.
  • Reiss, S. & McNally, R.J. (1985). Expectancy model of fear. In S. Reiss & R.R. Bootzin (Eds.), Theoretical issues in behavior therapy (pp. 107-21). San Diego: Academic Press.
  • Reiss, S., Peterson, R.A., Gursky, M. & McNally, R.J. (1986). Anxiety sensitivity, anxiety frequency, and the prediction of fearfulness. Behaviour Research and Therapy, 24, 1-8.
  • Ruesch, J. (1948). The infantile personality. Psychosomatic Medicine, 10, 134-44.
  • Salkovskis, P.M. (1988). Phenomenology, assessment, and the cognitive model of panic. In S. Rachman & J.D. Maser (Eds.), Panic: Psychological perspectives (pp. 111-36). Hillsdale, NJ: Erlbaum.
  • Salminen, J.K., Saarijarvi, S., Aarela, E. & Tamminen, T. (1994). Alexithymia – State or trait? One year follow-up study of general hospital psychiatric consultation out-patients. Journal of Psychosomatic Research, 38, 681-5.
  • Schaffer, C.E. (1993). The role of adult attachment in the experience and regulation of affect. Doctoral dissertation, Yale University.
  • Scott, J.P. (1963). The process of primary socialization in canine and human infants. Monographs of the  Society for  Research in  Child Development, 28, 1-47.
  • Seligman, M.E.P. (1988). Competing theories of panic. In S. Rachman & J.D. Maser (Eds.), Panic: Psychological perspectives (pp. 321-29). Hillsdale, NJ: Erlbaum.
  • Shipley, W.U. (1963). The demonstration in the domestic guinea-pig of a process resembling classical imprinting. Animal Behavior, 11, 470-4.
  • Sifneos, P.E. (1967). Clinical observations on some patients suffering from a variety of psychosomatic diseases. Acta Medicina Psychosomatica, 7, 1-10.
  • Sifneos, P.E. (1972/1973). Is dynamic psychotherapy contraindicated for a large number of patients with psychosomatic disease? Psychotherapy and Psychosomatics, 21, 133-6.
  • Sifneos, P.E. (1973). The prevalence of ‘alexithymic’ characteristics in psychosomatic patients. Psychotherapy and Psychosomatics, 22, 255-62.
  • Sifneos, P.E. (1975). Problems of psychotherapy of patients with alexithymic characteristics and physical disease. Psychotherapy and Psychosomatics, 26, 65-70.
  • Sifneos, P.E. (1987). Anhedonia and alexithymia: A potential correlation. In D.C. Clark & J. Fawcett (Eds.), Anhedonia and Affect Deficit States (pp. 119-27). New York: PMA Publishing.
  • Sifneos, P.E. (1988). Alexithymia and its relationship to hemispheric specialization, affect and creativity. Psychiatric Clinics of North America, 11, 287-92.
  • Slade, A. & Aber, J.L. (1992). Attachments, drives, and development: Conflicts and convergences in theory. In J.W. Barron, M.N. Eagle & D.L. Wolitzky (Eds.), Interface of psychoanalysis and psychology (pp. 154-85). Washington DC: American Psychological Association.
  • Stern, D.N. (1985). The interpersonal world of the infant. New York: Basic Books.
  • Taylor, G.J. (1984a). Alexithymia: Concept, measurement, and implications for treatment. American Journal of Psychiatry, 141, 725-32.
  • Taylor, G.J. (1984b). Psychotherapy with the boring patient. Canadian Journal of Psychiatry, 29, 217-22.
  • Taylor, G.J. (1987). Psychosomatic medicine and contemporary psychoanalysis. Madison, CT: International Universities Press.
  • Taylor, G.J. (1992). Psychosomatics and self-regulation. In J.W. Barron, M.N. Eagle & D.L. Wolitzky (Eds.), Interface of psychoanalysis and psychology (pp. 464-88). Washington, DC: American Psychological Association.
  • Taylor, G.J. (1994). The alexithymia construct: Conceptualization, validation, and relationship with basic dimensions of personality. New Trends in Experimental and Clinical Psychiatry, 10, 61-74.
  • Taylor, G.J. (2000). Recent developments in alexithymia theory and research. Canadian Journal of Psychiatry, 45, 134-42.
  • Taylor G.J., Bagby, R.M. & Parker, J.D.A. (1991). The alexithymia construct: A potential paradigm for psychosomatic medicine. Psychosomatics, 32, 153-64.
  • Taylor G.J., Bagby, R.M. & Parker, J.D.A. (1992). The revised Toronto Alexithymia Scale: Some reliability, validity, and normative data. Psychotherapy and Psychosomatics, 57, 34-41.
  • Taylor G.J., Bagby, R.M. & Parker, J.D.A. (1997). Disorders of affect regulation. Cambridge, UK: Cambridge University Press.
  • Taylor, G.J., Parker, J.D.A., Bagby, R.M. & Acklin, M.W. (1992). Alexithymia and somatic complaints in psychiatric out-patients. Journal of Psychosomatic Research, 36, 417-24.
  • Taylor G.J., Ryan D. & Bagby, R.M. (1985).  Toward the development of a new self-report alexithymia scale. Psychotherapy and Psychosomatics, 44, 191-9.
  • Telch, M.J., Lucas, J.A. & Nelson, P. (1989). Nonclinical panic in college students: An investigation of prevalence and symptomatology. Journal of Abnormal Psychology, 98, 300-6.
  • Van den Hout, M.A. (1988). The explanation of experimental panic. In S. Rachman & J.D. Maser (Eds.), Panic: Psychological perspectives (pp. 237-57). Hillsdale, NJ: Erlbaum.
  • Werner, H. (1957). The concept of development from a comparative and organismic point of view. In D.B. Harris (Ed.), The concept of development: An issue in the study of human behavior. Minneapolis: University of Minnesota Press.
  • Zajonc, R.B. (1984). On the primacy of affect. American Psychologist, 39, 117-23.
  • Zeitlin, S.B. & McNally, R.J. (1993). Alexithymia and anxiety sensitivity in panic disorder and obsessive-compulsive disorder. American Journal of Psychiatry, 150, 658-60.

(Dissertation Copyright 2004  –  All Rights Reserved)