Reply (Martin M. Katz)

This is a reply to this comment by Donald F. Klein

Don Klein questions the interpretation of the three dimensions that are identified in the book as constituting the structure of the depressive disorder. He asks whether these dimensions, the “anxiety-agitation-somatic”,"depressed mood-retardation”, and “hostility-interpersonal sensitivity”, derived from a mathematical factor analytic approach, reflect independent components of the disorder or are representing the presence or interaction of more than one syndrome. I believe he is asking whether the analysis has identified the presence of both a depressive and an anxiety disorder in the same patient, an interpretive position by many that a large segment of the severe depressive disorders combine both. That position has received a great deal of attention in the literature. He is asking whether the depressive syndrome or the mixture of syndromes is responsible for the variety of symptoms that are manifested. Further, he questions whether the neurotransmitter systems with their different associations with the components of psychological functioning are possibly analogous to Galen’s “humors”, I presume reflecting the various temperaments with which the humors were associated, or whether the dimensions should be viewed from the standpoint of the historical Sydenham “nosology’, reflective of different diagnostic classes having different sets of phenomena and course characteristics.

So it is important to clarify the meaning of the dimensions derived from factor analysis of the 11 mood, behavior, cognitive and somatic components measured across the six hospital Collaborative Study (CDS) sample of severe depressive disorder (Katz et al, 1984; Maas et al 1980). The primary aim in conducting that factor analysis was to develop measures that would describe in reliable and comprehensive detail the range of behavioral and somatic phenomena manifested across this group at that point in time, when the disorder was near or at the peak of the patient’s psychopathology. The component measures would be applied in the study so as to determine their relationships with the functioning of specified neuroendocrine and central neurotransmitter systems, previously implicated in the genesis or nature of the disorder, in order to (1) test then extant hypotheses about the underlying neurobehavioral mechanisms of the disorder and (2) through measuring subsequent changes in that baseline, uncover the sequence, timing and nature of the clinical actions of the relatively new antidepressant tricyclic agents being applied to their treatment.  The study was initiated in 1970 when the “catechol amine hypothesis” derived from a relatively small patient sample had not been tested definitively, and the drugs were still relatively new in their application. Aware that the analyses of behavioral measures were aimed at a point in time, and not at the long term course of the disorder, it was not an aim to develop a new classification scheme for the depressive disorders. Care had also been taken by adhering to the operationally derived criteria represented by the Research Diagnostic Criteria (RDC) (Spitzer et al 1979) that the patients met all the distinctive criteria for unipolar or bipolar depression.  The behavioral measures were selected to ensure comprehensive description of the “phenomena” of the disorder at that point in time, presumably around the peak of the illness, so that the state of the patient could also be measured at subsequent points during the brief, four week treatment course.

The measures served very well in these respects and the important results on neurobehavioral associations and specific drug effects reported in earlier publications (Katz et al, 1987, 1991, 1994), are summarized in detail in the book. I note that although the specific behavioral and mood tests were chosen to measure the major components of the disorder, the dimensions that are derived from the analysis of their associations are generated by a strictly mathematical procedure, Hotelling’s principal components analysis. The analysis tells us that the three identified dimensions account for the large majority of variance (75%) among the correlations of the 11 components. In the book I try to stay within the limits of this cross-sectional analysis of the psychopathology of these patients, and explain what the principal components tells us about the structure of the disorder at that time. I will try to briefly interpret that in answer to Klein’s question.

Although adhering to the requirements of the RDC for the selection of patients with a “major depressive disorder” protected against including patients with a mixed disorder for depressive and anxiety disorder in the study, it cannot completely rule out the inclusion of some patients of that type.  In view of the rigor with which the method of selection was conducted, however, the small segment of patients with combined disorders that may have slipped that screen would represent a minor influence on the results in this study.

I, therefore, view the dimensions as applying to the structure of the major depressive disorder at that point in time, and not as the result of the mixed disorder, defined as the combined depressive and anxiety disorders, and not as a structure that necessarily persists through all phases or throughout the course of the disorder.  

In answer to Klein’s question about the manner in which the dimensions relate to the neurobehavioral mechanisms underlying the varying manifestations of symptoms that identify the disorder, I offer the interpretation provided in the book.  I view the anxiety-agitation and the depressed mood-retardation dimensions as reflecting activity of two central nervous system states, the first, an excited state of “arousal”, that has a distinctly negative tone, the second, a slowed down or sedated tone, reflecting a down mood and retarded motor activity.  The states occur concurrently in time so the intensity of each and the fact that they are opposite in emotional tone make them appear as “opposed” in their interaction, and thus responsible in large part, for the basic turmoil and distress suffered by the patient at that point in time.  The emotional and somatic state produced would in turn explain the range of symptoms which also appear opposed in quality, sadness, motor retardation, loss of interest at one end, anxiety, anger, self-punitiveness, physical agitation, at the other.  Thus, I view the “conflictual interplay” of independent neurobehavioral components, as generating the array of symptoms manifested, and reflective of the dynamics that define the disorder at the height of its severity.

I welcome Don Klein’s attention to this issue and any further consideration by colleagues of these results and theory toward increasing understanding of the nature and neurobehavioral bases of the major depressive disorder.


 Katz MM, Koslow SH, Berman N, Secunda S, Maas JW, Casper R, Kocsis J, Stokes P A multivantaged approach to the measurement of behavioral and affect states for clinical and psychobiological research. Psychological Reports Monographs, 1984; 55:619-671.

 Katz MM Katz MM, Koslow SH, Maas JW, Frazer A, Bowden CL, Casper R, Croughan J, Kocsis J, Redmond E . The timing, specificity, and clinical prediction of tricyclic drug effects in depression. Psychol Med 1987;17:297-309.

Katz MM, Koslow SH, Maas JW, Frazer A, Kocsis J, Secunda S, Bowden CL, Casper RC. Identifying the specific actions of amitriptyline: Interrelationships of behavior, affect and plasma levels in depression. Psychological Medicine, 1991; 21, 599-611. 

Katz MM, Maas JW, Frazer A, Koslow SH, Bowden CL, Berman, N, Swann AC, Stokes PE . Drug-induced action on brain neurotransmitter systems and change in the behavior and emotions of depressed patients. Neuropsychopharmacology 1994; 11, 89-100.

Katz MM.  Depression and Drugs: The Neurobehavioral Structure of a Psychological Storm  New York, Springer; 2013.

Maas JW, Koslow S, Davis J, Katz MM, Mendels J, Robins E, Stokes PE, Bowden CL. Biologica1 component of the NIMH-Clinical Research Branch Collaborative Program on the Psychobiology of Depression: I. Background and theoretical considerations. Psychological Medicine, 1980; 10:759-776.

Spitzer RL, Endicott J, Robins E. Research diagnostic criteria: rationale and reliability. Archives of General Psychiatry. 1979; 35: 773-782.


Martin M. Katz

May 8, 2014