Clinical and biochemical aspects of depressive disorders: III. Treatment and controversies

SYNAPSE 8:185-211 (1991)

Clinical and Biochemical Aspects of Depressive Disorders: I. Introduction.

Classification, and Research Techniques SALLY CALDECOTT-HAZARD, BARRY H. GUZE, MITCHEL A. KLING, ARTHUR KLING, AND

LEWIS R. BAXTER Laboratory of Biomedical and Environmental Science (S.C.-H.) and Department of Psychiatry (B.H.G.,

L.R.B.), Uniuersity of California at Los Angeles, Los Angeles 90024; and Department of Psychiatry, Veterans Administration Medical Center, Sepulveda 91343 (A.K. i, California; Clinical Neuroendocrinology Branch,

Division of Intramural Research Programs, National Institute of Mental Health, Bethesda, Maryland 20892 (M.A.K.)

KEY WORDS

ABSTRACT The present review focuses on recent data from clinical and animal research concerning the biochemical bases of depressive disorders, diagnosis, and treatment. In addition to integrating these data, problems and future directions in this research are discussed. The review is presented in three parts. This study, Part I, describes diagnostic classification schemes for depressive disorders, some epidemiologcal and biological correlates of the classifications, and research techniques for investigating depressive disorders. Research techniques include animal models, human biochemical techniques, and Positron Emission Tomography. In a future issue, Part I1 will discuss various transmitterheceptor theories of depressive disorders, e.g., noradrenergq serotonergic, cholinergic, and dopaminergic, GABAergic, and peptidergic theories. Also in a future issue, Part I11 will discuss treatments for depression and some of the controversies in the field.

Depressive disorders, Humans, Animal models

INTRODUCTION During recent decades, a wealth of information has

accumulated on the biological, particularly biochemical, basis of depressive disorders and their treatments. Studies in human patients have often focused on the identification of biologcal markers for depression or on the evaluation of antidepressant drug efficacy. Animal studies have frequently investigated the physiological and biochemical mechanisms of antidepressants, or the mechanisms underlying depressed behavior, assessed using animal models. Although these data are comple- mentary and of importance to both groups of researchers, the quantity of data limits one’s ability to amass and integrate the facts. Therefore this review integrates both clinical and animal research, with the intent of producing a more comprehensive understanding, by neurobiologsts and physicians, of the current issues in depressive disorders. Areas that are discussed include the classification and treatment of depressive disorders, techniques for studying depressive disorders, various transmitterh-eceptor theories of depression, as well as problems, controversies, and future directions for research.

Historically, three terms-“mood disorder,” “affective disorder,” and “depressive disorder’’-have all referred 0 1991 wII,E;Y-I,ISS, I N C

to a group of clinical syndromes whose primary feature is a disturbance of mood, either depressed or elated, “accompanied by related cognitive, psychomotor, psychophysiological, and interpersonal difficulties” (Kler- man, 1983). However, technically, the three terms are not identical. Mood disorder, the most general phrase, refers to a prolonged disturbance of emotion that influ- ences one’s entire psychic life (APA, 1987). Affective disorder is concerned with emotional states attached to individual thoughts (Stedman, 19781, and depressive disorder is a less precise phrase that relates to depression as well as to mania in bipolar disease. In spite of the technical differences, these three terms are used inter- changeably in this review to facilitate integration of information from different sources.

The review is presented in three parts: Part I, introduction, classification, and techniques for studying depressive disorder; Part 11, transmitter/receptor theories of depression; and Part 111, treatments, controversies, and conclusions regarding depressive disorders. (Parts I1 and I11 are to be published separately.)

~ .-

Received October 26,1990 accepted in revised form December 18.1990 Address reprint requests t o Dr. Sally Caldecott. 3385 South U.S. Hlghwiy

17-92. Suite 21OM). Casselberry. FL32707.

186 S. CALDECOTT-HAZARD ET AL.

CLASSIFICATION OF CLINICAL SYMPTOMS Historical subtypes of depressive disorders

(S. Caldecott-Hazard) An understanding of the etiologies and biological

mechanisms of various depressive disorders will ultimately provide the best basis for classification. How- ever, for the past and present, subtype categorization has been based largely on symptoms that predominate in different patients. This method of classification has often led to overlap and confusion between subtypes, resulting in periodic changes in the diagnostic categories. A major goal of classification is to obtain subtypes that can be reliably applied, that meaningfully predict prognosis and course of illness, and that respond to specific treatments (Liebowitz et al., 1984). Although these goals have not yet been reached, the refinement of subtype criteria and ongoing research to validate these criteria signify progress in the right direction,

E. Kraepelin (1921) first attempted to link classification with prognosis when he distinguished psychoses that did not worsen over time (his “manic-depressive” category) from those that did (his “dementia praecox” category). Later, other categories of depressive disorders were developed. Two widely used classification dichotomies were “psychotic-neurotic” and “endogenous-reactive” subtypes of depression. Unfortunately, the criteria for these classifications were not clearly defined. “Psychotic” referred either to the presence of hallucinations and delusions, to severe symptoms, or to functional incapacity. “Neurotic” referred to the opposite extremes (mild symptoms). “Endogenous” initially indicated the absence of external precipitating events, but also referred to a group of symptoms including early morning waking, loss of appetite, weight loss, and depressed mood that was not reactive to external stimuli. “Reactive” depression referred to that state precipitated by stressful events or depressed mood that could be temporarily reversed by pleasurable events (Liebowitz et al., 1984).

Leonhard et al. (1962) proposed a bipolar-unipolar dichotomy for depressive disorders that was based on the existence of manic episodes in bipolar but not in unipolar depression. Although these classification terms have been modified, the bipolar-unipolar concept remains one of the most widely accepted subclassifications for depressive disorders (Schwartz et al., 1987). However, other classifications have been proposed that are more inclusive than the bipolar-unipolar subdivi- sions. For example, a primary-secondary classification defines depression in terms of the respective absence or presence of an additional, associated psychiatric illness (e.g., schizophrenia).

Starting in 1952 and continuing to the present, the American Psychiatric Association sponsored and published classifications of “depressive” (“affective” or “mood”) disorders in their Diagnostic and Statistical Manual of Mental Disorders, or DSM (APA, 1987). Classifications in each edition of the manual reflected

the current state of knowledge regarding mental disorders, and hence many of the same problems of clarity and reliability existed in the DSM as in other classifications. Specific criteria for classification categories did not even exist until the publication of DSM-I11 (1974). However, specific clinical, pharmacological, and biological data about the various classification categories have been accumulating; these data are promoting the process of subtype validation, especially within the most current DSM editions, DSM-IIIR (1987) and DSM IV, scheduled for publication in 1993. Some of the validating data are described under Epidemiological and Bio- logical Correlates of DSM Classifications. However, first, a more detailed discussion of the subclassifications and criteria of DSM-IIIR mood disorders is presented.

DSM-IIIR classification of mood disorders (B.H. Guze and L.R. Baxter)

Mood disorders are, by definition, characterized by a prolonged disturbance of emotion that colors the whole of one’s psychic life (APA, 1987). In DSM-IIIR, they are divided into the general categories of bipolar and depressive disorders. The hallmark feature of bipolar disorders is the occurrence of either a manic or hypomanic episode, which is usually also associated with a history of depressive episodes. A manic episode involves a distinct period of an abnormally and persistently elated mood, accompanied by other cognitive and phys- iolopcal disturbances (see Table I). A hypomanic episode also involves an elevated, irritable mood and associated symptoms, but the severity is milder than mania. Also, there is no marked impairment of social or occupational functioning, no hospitalization, and no delusions in hypomania. In contrast, the essential feature of a depressive disorder is one or more episodes of depression without manic or hypomanic episodes.

Bipolar disorder is further subclassified into three categories. The first and most widely recognized is bipolar disorder proper. This is characterized by one or more manic episodes that usually alternate with one or more “major depressive” episodes (Table 11). The second category of bipolar disorder is cyclothymia (Table III), in which there are numerous hypomanic episodes that alternate with periods of mild depressive symptoms, not considered major depressive episodes. In cyclothymia, similar to bipolar disorder, the rate of alternation from hypomanic to depressive states varies with the patient. Bipolar disorder “NOS” (not otherwise specified) is the third category of bipolar disorder. This is a residual category that includes the subtype “Bipolar 11,” characterized by hypomania alternating with major depressive episodes.

There are, likewise, three categories of depressive disorders. Major depressive disorder, the most severe form, consists of one or more major depressive episodes (Table 11). Dysthymia (Table IV), in contrast, occurs in individuals with a history of depressed mood more days

DEPRESSION: CLINICAL AND BIOCHEMICAL ASPECTS 187 TABLE I . Criteria for a manic episode

1. A distinct period of abnormally and persistently elevated, expansive, or irritable mood.

2. During the period of mood disturbance, a t least three of the following symptoms have persisted (four if the mood is only irritable) and have been present to a significant degree: a. inflated self-esteem or grandiosity. b. decreased need for sleep, e.g., feels rested after only 3 hours of

c. more talkative than usual or pressure to keep talking. d. flight of ideas or subjective experience that thoughts are racing. e. distractibility, i.e., attention too easily drawn to unimportant

or irrelevant external stimuli f. increase in goal-directed activity (either socially, at work or

school, or sexually) or psychomotor agitation. g. excessive involvement in pleasurable activities that have a

high potential for painful consequences, e.g., the person engages in unrestrained buying sprees, sexual indiscretions, or foolish business investments

3. Mood disturbance sufficiently severe to cause marked impairment

sleep.

in occupational functioning or in usual social activities or relationship with others, or to necessitiate hospitalization to prevent harm to self or others.

4. At no time during the disturbance have there been delusions or hallucinations for as long as 2 weeks in the absence of prominent mood symptoms (i.e., before the mood symptoms developed or after they have remitted).

5. Not superimposed on schizophrenia, schizophreniform disorder, delusional disorder, psychotic disorder NOS.

6. It cannot be established that an organic factor initiated and maintained the disturbance. Note: Somatic antidepressant treatment (e.g., drugs, ECT) that apparently precipitates a mood disturbance should not be considered an etiologic organic factor.

than not for a t least 2 years when the condition is not severe enough to meet criteria for a major depressive episode. The third category is depressive disorder NOS, which, like bipolar disorder NOS, is a residual category. In previous DSM-I11 classifications, this category was called atypical depression. Unfortunately, as in other classification schemes, the criteria were not clearly defined (Liebowitz et al., 1984). Often, this category included patients who did not have the “melancholi~~~ symptoms (Table V) associated with what was characterized as a “typical” depressive episode. For example, these patients showed hypersomnia andor weight gain as compared t o hyposomnia and weight loss. Some of these patients also expressed prominent anxiety, pho- bias, or hysterical features. In DSM-IIIR, these symptoms are now included under the classification of major depressive disorder, but these subgroupings of symptoms may still have pharmacotherapeutic importance.

All depressive episodes in bipolar or depressive disorders may also include psychotic features, melancholic features, or seasonal patterns of onset (APA, 1987). Psychotic and melancholic features derive from the previous classification dichotomies psychotic-neurotic and endogenous (or melancholic)-reactive depressions. However, in DSM-IIIR usage, their criteria are strictly defined and clearly limited. Psychotic features refer to the presence of delusions or hallucinations, which may be mood congruent or incongruent (appropriate or not with the particular mood). The content of mood congruent hallucinations or delusions is consistent with the

TABLE II. Criteria for a major depressive episode

1. At least five of the following symptoms have been present during the same 2-week period and represent a change from previous functioning: a t least one of the symptoms is either depressed mood, or loss of interest or pleasure. (Do not include symptoms that are clearly due to a physical condition, mood-incongruent delusions or hallucinations, incoherence, or marked loosening of associations.) a. depressed mood (or can be irritable mood in children and

adolescents) most of the day, nearly every day, as indicated either by subjective account or observation by others.

b. markedly diminished interest or pleasure a t all, or almost all, activities most of the day, nearly every day (as indicated either by subjective account or observation by others of apathy most of the time).

c. significant weight loss or weight gain when not dieting (e.g., more than 5% of body weight in a month), or decrease or increase in appetite nearly every day (in children, consider failure to make expected weight gains).

d. insomnia or hypersomnia nearly every day. e. psychomotor agitation or retardation nearly every day

(observable by others, not merely subjective feelings of restlessness or being slowed down).

f. fatigue or loss of energy nearly every day. g. feelings of worthlessness or excessive or inappropriate guilt

(which may be delusional) nearly every day (not merely self- reproach or guilt about being sick).

h. diminished ability to think or concentrate, or indecisiveness, nearly every day (either by subjective account or as observed by o thers) .

suicidal ideation without a specific plan or a suicide attempt or a specific plan for committing suicide.

i. recurrent thoughts of death (not just fear of dying), recurrent

2. It cannot be established that an organic factor initiated and maintained the disturbance, and the disturbance is not a normal reaction to the death of a loved one (uncomplicated bereavement). Note: Morbid preoccupation with worthlessness, suicical ideation, marked functional impairment or psychomotor retardation, or prolonged duration suggest bereavement complicated by major depression.

3. At no time during the disturbance have there been delusions or hallucinations for as long as 2 weeks in the absence of prominent mood symptoms (i.e., before the mood symptoms developed or after they have remitted).

4. Not superimposed on schizophrenia, schizophreniform disorder, delusional disorder, or psychotic disorder NOS.

TABLE III. Criteria for cvclothvmia

1. A chronic mood disorder lasting a t least 2 years. 2. During the 2 years there have been numerous episodes of

hypomania, mild depressed mood, or loss of interest or pleasure from usual activities.

3. There have been no periods longer than 2 months without mood symptoms.

4. No marked impairment in social or occupational functioning. 5. No clear evidence of a major depressive episode or manic episode

during the first 2 years of the disturbance (or 1 year in children and adolescents). Note: After this minimum period of cyclothymia, tbere may be superimposed manic or major depressive episodes, in which case the additional diagnosis of bipolar disorder or bipolar disorder NOS should be given.

schizophrenia or delusional disorder.

maintained the disturbance.

6. Not superimposed on a chronic psychotic disorder, such as

7. It cannot be established that an organic factor intitiated and

typical depressive themes such as inadequacy, guilt, disease, death, or deserved punishment. Hallucinations or delusions that are mood incongruent do not contain the above themes; rather, they may contain persecutory ideas, thought insertion (others are putting alien

188 S. CALDECOTT-HAZAFLD ET AL.

TABLE IV. Criteria for dvsthvmia TABLE V. Criteria for melancholic t vDe of deuressiue eDisode

1. A chronic disturbance of mood lasting most of the day, more days than not, for a t least 2 years.

2. During these 2 years there are no periods without depressive symptoms longer than 2 months.

3. Associated symptoms: the same a s in major depression but milder.

4. Depressed mood (or can be irritable mood in children and adolescents) for most of the day, more days than not, a s indicated either by subjective account or observation by others, for a t least 2 years (1 year for children and adolescents).

5. Presence, while depressed, of a t least two of the following: a. poor appetite or overeating. b. insomnia or hypersomnia. c. low energy or fatigue. d. low self-esteem.

The presence of a t least five of the following must occur: 1. Loss of interest or pleasure in all or most activities. 2. Lack of reactivity to usually pleasurable stimuli. 3. Depressed mood worse in the morning improving a s the day

4. Early morning awakening. 5. Psychomotor retardation or agitation visible to the clinician. 6. Significant anorexia or weight loss, defined a s greater than 5% of

7. No significant personality disturbance antedates the onset of the

8. One or more prior major depressive episodes followed by complete

9. Previous good response to antidepressant treatment such a s

progresses.

body weight in 1 month.

major depressive episode.

or near complete recovery.

lithium. monoamine oxidase inhibitors. tricvclic antidemessants.

6.

7.

8.

9.

10.

~~

e. poor concentration or difficulty making decisions. f. feelings of hopelessness. During a %-year period (1 year for children and adolescents) of the disturbance, never without the symptoms in 4 or more than 2 months a t a time. No evidence of a n unequivocal major depressive episode during the first 2 years (1 year for children and adolescents) of the disturbance. Note: There may have been a previous major depressive episode, provided there was a full remission (no significant signs or symptoms for 6 months) before development of the dysthymia. In addition, after these 2 years (1 year in children or adolescents) of dysthymia, there may be superimposed episodes of major depression, in which case both diagnoses are given. Has never had a manic episode or a n unequivocal hypomanic episode . Not superimposed on a chronic psychotic disorder such as schizophrenia or delusional disorder. It cannot be established that an organic factor initiated and maintained the disturbance.

thoughts into the patient’s mind; Maguire, 1988), or delusions of control (being controlled by outside forces).

Melancholic features are basically a subset of criteria that were previously referenced as “typical” depressive symptoms. For example, they include weight loss, anorexia, and insomnia, but not weight gain and hypersomnia. The definition of insomnia is also limited in this subtype to early morning awakening and worsened mood in the morning (see Table V). A seasonal pattern of depression refers to a consistent pattern between the onset of symptoms and a particular 60-day period of the year. The onset of the depression should be independent of seasonally occurring psychosocial stressors. Also, at least three episodes of mood disturbance that follow a seasonal pattern must occur in three separate years, and seasonal episodes of mood disturbance outnumber nonseasonal episodes by more than 3 to 1.

Table VI provides a summary and comparison of the various symptoms that may accompany manic and major depressive episodes (in bipolar and major depressive disorders), as well as dysthymic and cyclothymic disorders.

Epidemiological and biological correlates of DSM-IIIR classifications (S. Caldecott-Hazard)

Bipolar-nonbipolar dichotomy

A number of studies have provided evidence that the classification categories of bipolar and depressive disor-

ders are valid and distinct diagnostic entities. This evidence comes from studies of symptom prevalence, familial prevalence, genetics, response to treatment, and possibly also from neurochemical research. A consensus of several studies is that the incidence of depressive (nonbipolar) disorder in industrialized countries is approximately 3.2% for men and 4 5 9 . 3 % for women. In contrast, the incidence of bipolar disorder is less than 1% (Weissman and Boyd, 1984). Whereas there are approximately twice as many women as men who have nonbipolar depression, the numbers of both sexes having bipolar disorder are not significantly different (Weissman and Herman, 1977). According to retrospec- tive studies, the peak period of onset of bipolar disorder may also be slightly younger (twenties versus thirties) than for nonbipolar disorder (Loranger and Levine, 1978; Weissman and Boyd, 1984). However, longitudinal studies indicate overlap in ages of onset of the two disorders (Weissman and Boyd, 1984).

Twin and family studies suggest a heritable predisposition in both bipolar and nonbipolar depression. Monozygotic twins have a higher concordance rate (if one twin has a disorder, they both do) than dizygotic twins. However, some studies also suggest that there may be a genetic specificity for polarity, i.e., both twins are bipolar or both are nonbipolar (Nurnberger and Gershon, 1984). In addition, patients with bipolar and nonbipolar depression have a greater number of first- degree relatives who have depressive disorders than do controls. A greater percentage of the relatives of patients with nonbipoIar disorder also have nonbipolar, as compared to bipolar, disorder. Conversely, the relatives of patients with bipolar disorder have either bipolar or nonbipolar disorder in roughly equal frequencies (Ger- shon et al., 1987).

Studies of bipolar disorder in the Amish initially indicated that the disorder might be inherited via a dominant gene on chromosome 11 (Egeland et al., 1987). However, similar studies in other populations (where the manic-depressive disease also appeared to be inherited via a dominant gene) did not find any linkage to chromosome 11 (Kolata, 1987). Subsequent studies of

DEPRESSION: CLINICAL AND BIOCHEMICAL ASPECTS 189

TABLE VI. Summary of possible symptoms of mood disorders (compared to normals)

Melancholic Mania Major depression depression Dysthymia Cyclothymia

Mood Reactivity of

mood Symptoms worse

in a m Self Esteem Amount of

talking Distractibility Goal directed

activity Psychomotor

activity Impairment of

occupational or social functioning

Involvement in pleasurable activities

Eating/weight Sleep Suicidal

thoughts

t ni

ni

t i

t 1

1

yes

t

ni 1

no

1 ni

ni

1 I

1 1

1 or 1

yes

1

1 or t 1 or I yes

1 !

yes

ni ni

ni ni

1 or 1

ni

1

! 1

ni

mile 1 ni

ni

mild 1 n i

mild t mild 1

mild 1

no

mild 1

mild I or t mild 1 or i

no

mild 1 & 1 ni

ni

mild 1 & 1 ni

ni ni

ni

no

mild 1 & t

ni ni no

i = increased. 1 = decreased. ni = not included in the definition of this disorder

the original Amish group cast further doubt on this linkage data (Kelsoe et al., 1989). Regardless of whether the bipolar disease gene is located on chromosome 11 or not, the genetic basis for this disorder is supported by twin and familial studies, and ongoing molecular research should eventually solve this problem.

In addition to genetic factors, stressful life events probably also play a role in the onset of depression, particularly nonbipolar disorder. Seven out of eight international studies reported significant increases in the number of stressful events that preceeded the onset of nonbipolar depressive disorder as compared to per- sons who did not have depression. The role of stressful events in precipating bipolar disorder was not as clearly supported by the literature (Paykel, 1983).

Although controversial, some studies have reported that treatment response may differ in bipolar and nonbipolar patients. Lithium may be more effective in bipolar than in nonbipolar patients, and tricyclic antidepressants (such as imipramine) may be more important for the treatment of nonbipolar as compared to bipolar disorder (Prien et al., 1984). Antidepressant drugs (especially tricyclics) also have been reported to induce manic episodes or rapid cycling between mania and depression in bipolar patients (Pickar et al., 1984; Wehr and Goodwin, 1979). Some studies suggest that neurochemical data may differentiate bipolar from nonbipolar depression, but other conflicting reports also exist. For example, Schildkraut et al. (1978) and Beck- mann and Goodwin (1980) reported that patients with bipolar depression excrete less of the norepinephrine metabolite 3-Methoxy-4-hydroxy-phenethyleneglycol (MHPG) in their urine than patients with unipolar

(nonbipolar) depression. According to a review by Siever (19871, most studies over the past decade support these data. However, a comprehensive NIMH-CRB collabora- tive study found no differences in MHPG excretion between bipolar patients, unipolar (nonbipolar) patients, or controls. However, urinary norepinephrine and epinephrine levels were significantly elevated in unipolar as compared to bipolar patients. Norepineph- rine levels in bipolars were not different from controls, but bipolar epinephrine levels were increased as compared to controls (Koslow et al., 1983).

It was initially thought that Positron Emission To- mography (PET) studies in depressed patients might also differentiate between bipolar and nonbipolar disorders. Baxter et al. (1985) and Buchsbaum et al. (1984) both reported that prefrontal cortex glucose metabolism was decreased in bipolar depressed patients as compared to unipolar (nonbipolar) patients or normals. Later, however, Baxter et al. (1989) reported that unipolar and bipolar depressives both had decreased lateral left prefrontal lobe metabolism as compared to normals. Although other cerebral metabolic differences between bipolar and unipolar patients were also reported by either the Baxter or Buchbaum group, the findings were contradicted by data from the other group. Thus cerebral metabolic distinctions between bipolar and nonbipolar depressive disorders await further research.

Other classifications The subclassifications of melancholic (previously also

named endogenous^' depression) and atypical depression have each been linked to data that, in part, support


their validity as separate categories of depressive disorder. Although these data are not as robust as those for bipolar-nonbipolar classifications, they have importance for the treatment of symptoms associated with these subclassifications. Melancholic symptoms, such as depressed mood that is worse in the morning, loss of interest in pleasurable activities, or weight loss, have been reported to be more responsive to treatment with tricyclics than nonmelancholic depression (Kdoh and Garside, 1963; Mendels and Cochrane, 1968). Carroll et al. (1981) also reported that more patients with endogenous depression fail to suppress cortisol release caused by dexamethasone than do nonendogenous patients. However, Carroll’s criteria for endogenous depression were not identical to those for endogenous/ melancholic depression described in DSM-IIIR, and other researchers report no differences between dexamethasone’s effect in endogenous versus nonendogenous patients (Coryell et al., 1982).

Atypical depression, including symptoms of anxiety, panic, hypersomnia, and weight gain, reportedly re- sponds preferentially to MA0 inhibitor antidepressants (Liebowitz et al., 1984). Although these symptoms have been incorporated into the general category of depressive disorder in DSM-IIIR, they may still have value as a therapeutic entity.

Conclusions The revision of classification schemes over the years,

particularly the clarification and limitation of criteria for the classification categories, has increased both the reliability and the usefulness ofthese diagnostic tools. A number of studies have provided data to support the distinctiveness of various classification categories, most notably bipolar-nonbipolar distinctions. However, it is clear that many more studies need to be performed. Data on the etiologies and biological mechanisms underlying different depressive symptoms will provide the ultimate basis on which classification is determined.

TECHNIQUES FOR STUDYING DEPRESSIVE DISORDERS AND THEIR BIOCHEMICAL BASIS

Animal models of depression (M.A. Kling and A. Kling)

Although there have been many advances concerning the biological underpinnings of major depression, it remains an entity defined primarily in terms of clinical descriptive criteria. As such, its diagnosis can be arrived at only after excluding other known causes of the same syndrome. Animal models can often aid in defining pathophysiologic mechanisms that may prevail in human disease states; however, the development and validation of such models for major depression have been hampered by a number of factors. Among these are: the intrinsically experiential nature of some of the core symptoms of major depression; species differences in the contextual appropriateness of various behaviors

that are associated with human depression; difficulty in defining robust biological alterations in human depression in sufficient neurobiologic detail; and the possibil- ity that major depression may in fact consist of multiple subsyndromes with different, even divergent patho- physiologies. Nevertheless, a number of animal models of depression have been proposed and, in some cases, shown to reproduce some of the more observable and quantifiable features of the major depressive syndrome.

McKinney and Bunney (1969) have proposed criteria for assessing the validity of animal models for depression. These authors recommended that putative models resemble human depression with respect to etiology, biochemistry, symptomatology, and treatment. A subsequent animal model core group report (Reite et al., 1983) more extensively delineated specific features of human depression that could be assessed in animal models, either individually or collectively. Although the McKin- ney and Bunney criteria represented an important step forward in providing organizing principles for guiding research on animal models, Willner (1984) and others have pointed out that the etiology and biochemistry of human depressions are themselves poorly and incom- pletely understood, and as such do not provide an adequate basis for evaluating models. Hence, Willner (1984) proposed criteria for evaluating prospective models with regard not only to face validity, as embodied in the McKinney and Bunney criteria, but also with regard to predictive (the model correctly identifies treatments known to be effective or ineffective in treating human depression, and in a potency order that corresponds to that for human depression) and construct validity (the model identifies features that are not only homologous with human depression but are consistent with at least some aspects of the theoretical or empirical basis of human depression). It should be borne in mind, however, that even models that do not fully satisfy these criteria may have value to the extent that they generate further testable hypotheses that may refine the study of depression, either in humans or experimental animals. We may term this property, for example, heuristic ualid- ity.

The value of animal models is not necessarily absolute but is in part dependent on the specific purpose to be served by the model. As an example, the development of new antidepressant drugs often requires the screening of a large number of compounds and as such customarily utilizes models that can be tested in a large number of animals, usually rats, at minimum expense and prefer- ably with a minimum of complexity and duration of treatment. For such purposes, predictive validity may be sufficient. However, models designed to further information regarding pathogenetic andor pathophysiologic aspects of depression may be more complex and utilize species ethologically more similar to humans. For these goals, construct validity is crucial.

In this section we review some of the available animal models with attention not only to their validity, but also


to their heuristic value in generating hypotheses regarding the pathophysiology of human depressions. For additional information, the reader is referred to one or more of several excellent recent reviews of animal models of depression (Harris 1989; Reite et al., 1983; Will- ner, 1984,1990).

Clinical and biological characteristics of major depression as related to animal studies

One factor that complicates the development of animal models of depression is its syndromal nature a t the present levels of understanding. Hence, major depression is currently conceived as a constellation of symptoms and signs defined entirely in descriptive terms and by a consensus review of all available validating data. The lack of reliably quantifiable measures of these features lends some subjectivity to the diagnosis, as well as leading to some ambiguity as to what constitutes the core features of the illness. Most authors agree, almost by definition, that despondent or dysphoric mood is the cardinal feature of major depression; however, this symptom is ubiquitous in the population, and its occurrence, at least transiently, is an inevitable and perhaps even necessary feature of the human condition. There- fore, intensity and duration criteria have been applied in an attempt to define a “syndromal” or critical level of this symptom. Such criteria are arbitrary and have shifted somewhat as diagnostic criteria have been revised. Moreover, this symptom, although obvious to an observer in extreme cases, is often dependent on the patient’s subjective report, and great variation exists in the degree to which individuals express or can tolerate such feelings before the patient complains about them. Because of the above, and the inherent difficulties in assessing internal emotional state in experimental animals, most attempts at model development have focused on other clinical features of depression, or on biological disturbances that have only partially been validated. Some of the clinical features of major depression that are applicable to animal models are: anhedonia, or inability to experience pleasure; decreased reactivity of mood to usually pleasurable stimuli; alterations in arousal mechanisms, including either psychomotor agitation or retardation, generalized anxiety, and disruptions of sleep pattern; and disturbances of appe- titive function, leading to decreased or increased appetite (often reflecting a change in the perceived desirabil- ity of food) andfor food intake, andfor gastrointestinal distress.

Biological disturbances that have been the focus of animal models include changes in neurotransmitters, their metabolites, or hormone levels. Neurotransmitter changes have been reported in the catecholaminergic, serotonergx, dopaminergic, cholinergic, GABAergic, and peptide (corticotrophin-releasing hormone, argin- ine vasopressin, and somatostatin) systems. However,

the results from human studies, on which the animal studies are based, are often conflicting. One attempt to reconcile the conflicting data relies on a closer examination of the clinical picture of depression and its potential heterogeneity. In this regard, many of the symptoms and physiological changes characteristic of the depression subsyndrome, melancholia, are strikingly similar to those seen in acute responses to stressors across a variety of mammalian species. Hence, it has been proposed (Gold et al., 1988a,b; Kling et al., 1989) that an intense arousal, psychomotor agitation, lack of reactivity of mood, decreased sleep and decreased appetite and food intake seen in melancholia might represent a pathological activation of neurobiological systems involved in the response to stress, or alternatively a deficiency in those mechanisms that ordinarily counter- regulate and restrain the organism’s response to stress. In contrast, depression characterized by fatigue, leth- argy, psychomotor retardation, preservation or en- hancement of mood reactivity, and increased sleep and appetite might represent a pathological inactivation of neurobiologic systems involved in the response to stress (Gold et al., 1988a,b; Kling et al., 1989). Furthermore, animal models might mimic one subtype of depression better than another, and they provide a medium in which activation or inhibition of neurobiologic systems mediating stress responses can be studied.

Biochemically/pharmacologically based models Catecholamines. One of the first conceptual sche-

mata to propose a biological basis for human depression was the catecholamine hypothesis as formulated by Schildkraut (1965). This hypothesis was in fact borne out of observations that treatments that tended to deplete brain catecholamines, e.g., giving the drug reserpine, were associated with the induction of depressive syndromes. Simultaneously, preclinical investigations were showing that reserpine and related drugs produced a behavioral and physiological syndrome that could be reversed by pretreatment with clinically effica- cious antidepressants, namely tricyclics and monoamine oxidase inhibitors, MAOIs (Costa et al., 1960; Max- well and Palmer, 1961; Theobald et al., 1964). However, the generalizability of this model is limited by the fact that some clinically useful antidepressants, such as mianserin and trazodone, are inactive in the reserpine test, whereas a number of drugs that are not effective antidepressants, do reverse reserpinelike agents’ effects (Silvestrini, 1982; Van Riezen, 1979). Moreover, reserpinized animals show evidence of physical debili- tation to a degree that is undesirable in a model of depression. Although it continues to be of some historical and conceptual value, reserpine’s current usefulness in developing novel antidepressants is limited, as it mainly identifies drugs that enhance noradrenergic function acutely.


Other models based on producing alterations in nor- adrenergx function have utilized the administration of alpha-methylparatyrosine (AMPT), clonidine, and 6-hydroxydopamine (6-OHDA). AMPT is of conceptual value because it inhibits the rate-limiting step in catecholamine synthesis, namely, tyrosine hydroxylase. AMPT produces a depressivelike syndrome in monkeys (see Table VII) (Redmond et al., 1971a,b).

The syndrome seen after AMPT administration in the stump-tailed macaque (M. arctoides) included: (1) general loss of interest as indicated by decreased initiation of social interactions, decrease in exhibited threat and aggression, and a drop in social rank; (2) anhedonia and diminished sexual activity; (3) decreased locomotor activity in which the subject appeared lethargic and often sat in a huddled position with head bowed and arms crossed on the chest; this posture may be interpreted as a sedative effect of the drug; (4) a “flattened and even sad-appearing facial expression of this ordinarily highly expressive species, which is coincident with dimished vocalizations, (5) appetite disturbances, which can only be suggested by the correlation of weight loss occurring with AMPT treatment and recovery after cessation of treatment; however, this could be tested more directly with operant procedures; and (6) decreased self-care, i.e., grooming. Since grooming in primates serves a social function as well as to keep the skin and hair free of parasites, it may be analogous to personal hygiene in humans, often neglected in depressive disorders. Sleep architecture was not studied but could easily be recorded in such subjects using EEG radiotelemetry (Kling et al., 1984). Cognitive function was not tested in the studies referred to but would require motivation- independent measures. Biological rhythms were not assessed, but studies could easily be carried out. Bio- chemical changes included, as expected, a decrease in 3Methoxy-4-hydroxy-phenethyleneglycol (MHPG) and 3Methyoxy-4-hydroxy-mandelic acid (VMA) extraction during treatment, which increased following cessation of treatment.

Thus administration of AMPT may serve as a useful model of noradenergic (and dopaminergic) dysfunction in depressive disorders. In particular, this syndrome seems most representative of the lethargic subtype of depression. Because this drug inhibits the synthesis of both norepinephrine (NE) and dopamine (DA), it would seem of considerable heuristic interest to contrast this syndrome with behavioral changes resulting from the administration of inhibitors of dopamine beta-hydroxylase (DBH), which would block synthesis of NE specifically.

Clonidine is an alpha-2 adrenergc agonist that acts on presynaptic receptors located on cell bodies on the locus coeruleus to inhibit its spontaneous activity. As such, it would be expected to produce effects similar to other treatments that reduce noradrenergic activity in the brain. Clonidine does, in fact, induce a behavioral

syndrome in rats consisting of sedation, analgesia, hypothermia, and depression of conditioned avoidance behavior. Of interest in relation to this model is the observation that repeated treatments with desipramine or electroconvulsive shock (more than single treatments) are effective in reversing the behavioral depression induced by clonidine (Green et al., 1982). Moreover, mianserin, which as mentioned above is not identified by a number of classical models of antidepressant activity, did antagonize the effect of clonidine on conditioned avoidance behavior (Gower and Marriott, 1980; Robson et al., 1978). This action is consistent with the idea that mianserin, in addition to its effects on serotonin receptors, also may block alpha-2 receptors (Robson et al., 1978). Indeed, other drugs that act more specifically as alpha-2 adrenergic antagonists, such as idazoxan, appear to have antidepressant effects in humans (Osman et al., 1989). However, questions about the specificity of the clonidine effect may be raised based on recent observations that clonidine binds to histamine receptors as well as to a novel class of receptors in the brainstem; these areas in the brainstem mediate blood pressure regulation, and the novel receptors are neither noradrenergic nor histaminergic, but specifically recog- nize the imidazole moiety shared by clonidine, idazoxan, and a number of similar acting drugs (Bricca et al., 1989; Parini et al., 1989). Further evaluation of the usefulness of this model thus appears to be predicated upon comparison of the effects of clonidine to those of drugs that are more specific for alpha-2-adrenergic receptors.

Another means of reducing catecholaminergic function is the neurotoxin 6-hydroxydopamine (6-OHDA), which destroys catecholaminergic terminals. Studies in rats indicate that this drug, when given centrally, produces sustained decreases in catecholamine concentrations in catecholamine terminal-rich brain regions. The relative effects of this drug on norepinephrine versus dopamine neurotoxicity can be controlled either by pretreatment with desipramine, which protects norepinephrine-containing terminals selectively by blocking the uptake of 6-OHDA into the presynaptic terminal, or by using low doses, which somewhat selectively destroys dopamine-containing terminals because of the relatively greater affinity of their uptake system for 6-OHDA compared to noradrenergic terminals.

In one investigation in nonhuman primates, Red- mond et al. (1973) examined the effects of intraventric- ularly administered 6-OHDA on social behavior in a semifree-ranging colony of rhesus monkeys on Guaya- can Island, Puerto Rico. Four subjects were treated with increasing doses of 6-OHDA at 12-hour intervals for 3 days (a total of 31 mg), producing a near 70% depletion of brain NE; 4 others were implanted but untreated and served as field controls.

After release, the treated subjects were slow to return to their social groups, were frequently peripheral to


TABLE VII. Selected studies of pharmacological and behavioral models of human depressive disorder in non-human primates

Separation in Infants & Juveniles Serotonin Serotonin (Peer)

AMPT PCPA agonists antagonists (Protest) (Despair) (Separation) Symptoms‘

a. General loss of interest

IPlay in juveniles ISocial interactions lAuto & hetero groom iThreat & aggression Social rank Uninterested in

surroundings iResponse of female

to male attempts at copulation.

IMounting and presentations

IHuddling with head bowed and arms crossed.

IActivity, appeared lethargic. ? Sedation.

loss of facial, vocal, & postural evidence of affective responsivity. Mask- like facies.

Look “sad”; marked

3

Unchanged t Approach except when t Grooming become toxic Coalition

!Behavior

i Approach - Social IGrooming withdrawal

i Avoiding

Protest iExploration

b. Anhedonia and sexual behavior

Unchanged Unchanged Unchanged - - -

ILocomotion

Alert

Unchanged t Locomotion c. Disturbed psychomotor regulation

lhcomotion tActivity IActivity

Alert

d. Sadness Mask-like facies in subject.

e. Sleep disturbances

f. Appetite disturbances

g. Cognitive disturbances

h. Disturbances of biological rhythms

i. Biochemical changes

- ?

Weight loss t Appetite

Not tested -

? -

- iREM -

i Appetite - iFood and water.

- - -

Weight loss

Not tested

? - !Heart rate IHeart rate ITemperature ITemperature

iMHPG and VMA !after DC Rx

IMHPG, 5-HIAA. - -

!Serotonin

tcortisol. Yes - -

in Hypoth.

j. Reversibility & duration of disorder

appearance & changes in physiology

k. Physical

1. Other

Yes, long duration.

Poorly groomed

Yes Yes

Ataxia, hair - loss, paresthesias, look sick.

- - Looks into space

Mother withdraws

- tVocal lVocal lVocal iPlay

m. Mother-infant from infant.

’Categories a-h are from the Animal Models Group Report. The present authors have added i-m. f = increased. 1 = decreased.

their group or isolated, and were attacked by the normal animals on their initial contact with their group. Subse- quent observations found that, compared to the field controls, the treated subjects showed decreases in auto and hetero-grooming, total social initiated behaviors, and threat and attack behaviors. Treated animals would frequently assume a slumped posture with ex- pressionless facies, but did not appear sedated. Two treated females never returned to their group of infants. In many respects their behavior resembled that seen after AMPT treatment.

The use of kin-related subjects added an important dimension to these studies in that this study design allowed for the observation of disruptions in social bondings that would not have been observable had nonrelated subjects been used, as is often the case in laboratory studies. There are, however, some important drawbacks to field studies, which include limited observations, due to environmental conditions, and difficulty in obtaining biological measures. Considerable time is often required to locate subjects when they do not remain within a group, and experimental subjects may


die from attacks by other members of the colony or predators. These issues, as they pertain to the development and use of animal models of depression, are cov- ered in more detail below in connection with models involving manipulation of the social context.

As with the AMPT model, it is ofconsiderable interest to establish the relative roles of norepinephrine and dopamine in aspects of the symptom complex of the 6-OHDA model. This may be feasible by utilizing the techniques mentioned above for obtaining specific de- pletions of NE or DA. The potential role of DA in the decreased motor activity seen in this syndrome is further discussed below (see Cross Validation of Animal Models).

The use of amphetamine withdrawal as a model of depression is an outgrowth of the catecholamine hypothesis, but it differs from the above in that it relies on inducing a neurobiologic alteration by removing rather than introducing a pharmacoloac stimulus, in this case chronic amphetamine. As such, it may be more amena- ble to bioloqc and pharmacologic study because of the lack of acute interfering effects of the drug used to produce the syndrome. The behavioral syndrome associated with this paradigm reproduces a number of the clinical features of human depression, notably decreased motor activity (Caldecott-Hazard et al., 1988; Lynch and Leonard, 1978; Seltzer and Tonge, 1975). Additionally, in this syndrome, direct, objective evidence for an anhedonialike state is provided by studies showing decreased responding for intracranial self- stimulation (ICSS) (Leith and Barrett, 1976), which was shown to be due to an increased threshold for ICSS (Cassens et al., 1981), and thus a decreased efficacy of normally rewarding stimuli rather than secondary to motor impairment. Finally, this effect is ameliorated by tricyclic antidepressants, the MA01 pargyline, and the atypical antidepressant mianserin (Kokkinidis et al., 1980; Lynch and Leonard, 1978; Seltzer and Tonge, 1975).

Serotonin. Whereas depletion of serotonin (5-HT) may play an accessory role in several of the models mentioned above (e.g., reserpine, and possibly amphetamine withdrawal), other animal models have been proposed that produce alterations in serotonin transmission more selectively.

The acute administration of 5-HTP, the immediate precursor of serotonin, produces a behavioral depression that is reversed by pretreatment with tricyclic antidepressants and MAOIs. Behavioral depression is also prevented by “atypical” antidepressants believed to act through serotonergic mechanisms, such as mianserin, iprindole, and trazodone. However, like many of the pharmacologic potentiation models mentioned above, the 5-HTP potentiation model suffers from some serious discrepancies with the pharmacotherapy of depression. Not the least of these is the observation that

fluoxetine, now known to be one of the most potent antidepressants available (Benfield et al., 19861, actually enhances 5-HTP-induced behavioral depression. There are also important discrepancies between the acute and chronic effects of tricyclics and MAOIs on 5-HT-induced behavior. These data, along with more recent information regarding the complexity of serotonergic transmission in the brain, suggest that more sophisticated models of serotonergic dysfunction need to be developed in order to assess the role of this system in the pathophysiology of depression and to account for the known pharmacology of antidepressant drugs.

Another agent that has been used to probe the role of serotonergic mechanisms in depressive disorders is parachlorophenylalanine (PCPA), which inhibits the synthesis of serotonin by blocking tryptophan hydroxylase, the first obligatory step in this synthetic pathway. Although PCPA administration in monkeys has been shown to increase “huddling” behavior (Kraemer and McKinney, 1979) and to produce “masklike” facies, weight loss, and to increase hair-pulling behavior (Raleigh et al., 1983), this drug did not appear to significantly affect locomotor activity, level of interest, or to produce any evidence of anhedonia (see Table VII). It is of interest that at the doses used, PCPA decreased the excretion not only of 5-HIAA, as expected, but also of MHPG. This observation is consistent with the finding of Potter et al. (1985) showing that antidepressant drugs selective for serotonin and norepinephrine uptake blockade may have parallel effects on both serotonergic and noradrenergic measures in depressed patients, indicating that activity of the serotonergic and noradrenergic systems are not necessarily indepen- dently perturbable. The data cited above further suggest that the role of serotonin in human depression may be more difficult to characterize than that of norepinephrine.

Although the 5-HTP and PCPA-related models seem to be of only limited value, the administration of serotonin antagonists in vervet monkeys (C. aethoiops) did result in decreased approach behaviors, locomotion, and appetite (Table VII). This effect was reversible after cessation of treatment. Opposite effects were observed after administration of serotonin agonists. No metabolic studies were reported in these experiments, but many behaviors associated with the depressive categories are seen with serotonin antagonists and the reverse with agonists. Thus drugs acting directly at specific serotonin receptors may ultimately prove more useful than those that affect serotonergic function globally. This is not only understandable in the light of recent data indicating considerably greater diversity of serotonin receptors than previously recognized but can now be more directly addressed given the development over the past several years of a number of drugs much more specific for these subtypes than previously available (Schmidt and Peroutka, 1989).


Lesion-based models

Olfactory bulbectomy. The olfactory system has long been known to have anatomic connections with the limbic system, and in nonhuman mammalian species tends to be well developed and intimately involved in reproductive mechanisms. Moreover, the olfactory af- ferents are the only primary sensory fibers to have direct access both to the outside world (at the olfactory mucosa) and to the cerebral cortex. In addition to being highly interconnected with several limbic structures, the olfactory bulb receives inputs from the locus coeruleus and raphe nuclei. Bilateral lesions of the olfactory bulb in the rat produce, with a latency of 10-14 days, a behavioral syndrome characterized by hyperactivity, irritability, a variety of aggressive behaviors including predatory attack and infanticide, and variable alterations in intermale and territorial aggression, the intensity and direction of which are species dependent (Leonard and Tuite, 1981; Willner, 1984,1990). Bulbec- tomized rats also show deficits in passive avoidance learning, which, however, appears largely attributable to the hyperactivity. This behavioral syndrome is also associated with hypersecretion of corticosteroids. Most or all of these altered behaviors, as well as the hypercor- ticism, are reversed by antidepressant drugs represent- ing diverse classes, including tricyclics, selective serotonin uptake blockers, serotonin antagonists, and the dopamine uptake blocker bupropion (Willner, 1984). Of interest is the observation that most of these drugs require multiple administrations to be effective, at least when given systemically.

The bulbectomy model is of interest in a number of respects. First, it lends itself well to drug treatment paradigms because, once established, the syndrome is stable and not dependent on the administration of drugs that may interact with the antidepressants under study. As such, it may represent a viable alternative to pharmacologic potentiation models as a screening procedure for new antidepressants. Second, it may represent a model of a specific subtype of depression, namely, agitated depression, in which enhanced arousal and motor activity and irritability are key features. This type of depression, although relatively uncommon, is more likely to be associated with hypercortisolism than depressions associated with decreased arousal (Brown et al., 1988). Interestingly, this model provides evidence for linkage of altered serotonin function with hypercortisolism in human depression. Hence, many of the behavioral as well as biochemical features of this syndrome may be related to a deficiency of serotonergic transmission in the amygdala (Willner, 1990). At the same time, decreased global indices of serotonin metabolism may be preferentially associated with agitated depressions and with aggressive behaviors, including violent suicides (Asberg et al., 1980; Banki et al., 1981). Third, most of the ameliorative effects of antidepres-

sants on this model seem to require repeated treatments.

Like most other models, there are a few inconsisten- cies in the properties of the bulbectomy model. One is that the MA01 tranylcypromine increases both the passive avoidance deficit and the hyperactivity. A related issue is the difficulty in demonstrating that the passive avoidance deficit is not simply a function of the hyperactivity, and the therapeutic effects of antidepressants therefore attributable to decreased locomotor activity (Willner, 1990). The dissociability between effects of antidepressants on motor activity and activity in an animal model of depression are discussed further below in connection with the Porsolt forced-swim test.

Internal capsule lesions. Two reports in abstract form appeared several years ago describing depression of responding for intracranial self-stimulation in rats by lesions of the internal capsule, in the regon of the telencephalic-diencephalic border (Cornfeldt et al., 1982; Szewczak et al., 1982). Although the deficit was reported to be alleviated rather specifically in these studies by repeated treatment with a variety of antidepressants, little or no further information has appeared regarding this paradigm and it must still be regarded as preliminary at best.

Models based on alteration of the psychosocial context

It has long been known that episodes of depression in humans are frequently associated with psychosocial stressors (reviewed by Anisman, 1984). Traditional psy- choanalytic conceptualizations of depression tended to focus on separation and loss as factors associated with depressivelike syndromes (e.g., Freud, 1917); this view is supported by empirical epidemiologxal investigations in which by far the stressors most commonly associated with depressive episodes are separation from and/or loss of a loved one (Anisman et al., 1984). Accord- ingly, many investigators have attempted to use this type of stressor in animals to induce a depressivelike syndrome. However, loss may be a type of stressor that i s most applicable to species with a highly sophisticated, integral social structure and less relevant to animals such as rats, which have been used extensively in other models of depression and in behavioral paradigms in general for a variety of reasons. Studies in such species have tended to utilize stressors with stimulus properties designed to reproduce various cognitive structures associated with depression, e.g., uncontrollable or un- predictable stimuli.

Separationlloss models. Beginning in the early 1960s, several groups reported a characteristic pattern of behavioral responses in infant monkeys separated from their mothers following the formation of strong social bonds (Table VII) (Seay et al., 1962; Jensen and Tolman, 1962). These were observed to undergo a reproducible, almost stereotyped sequence beginning with an initial,

196 S. CALDECOTT-HAZARD ET AL

short-lived “protest” phase characterized by marked increases in motor activity and verbal output (specifically “distress vocalizations”), and heart rate, followed by a “despair” phase consisting in part of decreased activity and vocalization, social withdrawal and lack of play behavior, decreased food and water intake, and decreased heart rate, which may progress to death from wasting in some cases. This pattern of behavioral changes was noted to be strikingly similar to that observed in children separated from their mothers, termed “anaclitic depression” by Spitz (1946) and “separation anxiety” by Bowlby (1952, 1961). Subsequent studies revealed physiological changes resembling those seen in human depression, such as alterations in catecholamine and indoleamine metabolism, pituitary- adrenal activity, sleep EEG architecture, and circadian rhythms (Reite et al., 1981). Moreover, some or all aspects of the syndrome can be ameliorated by antidepressant treatments such as imipramine (Suomi et al., 1978), desipramine (Hrdina, 1979), and electroconvulsive therapy (Lewis and McKinney, 1976).

Although maternal separation does bear some resem- blance to the syndrome of adult depression, it is most analogous to the syndrome of maternal separation in humans. Thus as separation from or neglect by parental figures early in life has been proposed as a risk factor for later depression (Lloyd 19801, this model may be most useful for examination of its long-term consequences. For example, Kraemer and McKinney (1979) showed that monkeys who had been separated early in life were more sensitive to the “depressogenic” effects of AMPT than control subjects, suggesting that early life separation may have sustained neurobiologic effects that can be latent unless unmasked by specific interventions.

“Learned helplessness”mode1s. In the rnid-l960s, Se- ligman and colleagues found that dogs subjected to inescapable, uncontrollable electric shocks showed subsequent deficits in learning to terminate shock under conditions in which it was escapable (Overmier and Seligman, 1967). they termed this phenomenon “learned helplessness,” reasoning that their inability to terminate (control) the shock in the first paradigm led the animals to learn that their behavior was ineffective and that this cognitive structure generalized to the subsequent paradigm. The idea of such a “learning deficit” was seen as analogous to some of the cognitive disturbances observed in patients with depression, who frequently voice feelings of helplessness of powerlessness to cope with environmental stressors. Experimentally induced behavioral deficits of this type have since been observed in rats and other species, both nonprimate (Maier and Seligman, 1976) and primate (Rush et al., 19831, including humans (Breier et al., 1987).

Since this hypothesis was first proposed, it has been the subject of intense controversy focusing on the mechanism underlying the behavioral deficits observed in animals exposed to inescapable stress. Whereas some

(Maier, 1984; Maier and Seligman, 1976; Seligman and Weiss, 1980) have argued €or a change in the animal’s cognitive set as the explanation, others have proposed that other, more fundamental neurochemical (although not necessarily incompatible) mechanisms may be at work. Hence, Weiss and colleagues have advanced data suggesting that the inescapable stress produces a depletion of norepinephrine andor a decrease in tyrosine hydroxylase activity in the region of the locus coeruleus (Hughes et al., 1984; Lehnert et al., 1984; Weiss et al., 1981). These data indicate that this depletion occurs mainly in recurrent inhibitory collaterals from the LC and results in a net disinhibition of LC activity, altering the animal’s responsiveness to subsequent stressors. The idea of heightened noradrenergic function may help to explain the attentional (Minor et al., 1984,1988) and motor (Anisman et al., 1979; Glazer and Weiss, 1976; Maier and Jackson, 1979) impairments that have been observed in association with this syndrome.

Irrespective of the controversy over the mechanism of the apparent learning deficits in uncontrollably stressed animals, a number of groups have reported reversal of these deficits by several antidepressant treatments, including tricyclic drugs, MAOIs, various “atypical” agents, and ECS (Dorworth and Overmier, 1977; Leshner et al., 1979; Petty and Sherman, 1980; Sherman et al., 1982). Moreover, Zacharko et al. (1983) showed suppression of responding for ICSS in inescap- abIy shocked rats, which was reversed by chronic desipramine treatment (Zacharko et al., 1984). Further- more, although many of the behavioral deficits are relatively transitory (Maier, 1984; Weiss et al., 1982), Desan et al. (1988) have shown a more sustained effect of inescapable shock on running wheel activity in rats, lasting 2-6 weeks, which was reversed only after 7 days or more of treatment with desipramine.

Another type of inescapable stress model that has been advocated as a screening test for antidepressants is known as the forced swim test, or Porsolt test after its developer (Porsolt et al., 1978). This test relies on the tendency of rats, when forced to swim in a confined space, to eventually assume a position of relative immobility, movingjust enough to keep the head above water. Porsolt termed this state “behavioral despair” and found that a number of clinically effective antidepressants reduced the amount of immobility assumed by the animals when subjected to a second trial a t doses that actually decreased motor activity in an open field test. Among the antidepressants that were positive in this test were tricyclics; MAOIs; the serotonergic agents fenfluramine, tolaxatone, and viloxazine; and nomifen- sine (Porsolt et al., 1978). However, the latter drug, like the psychostimulants d-amphetamine and caffeine, increased motor activity or caused stereotypy in open field testing, making motor stimulation difficult to rule out as a cause of reduced immobility with forced swimming. There are a few false negatives and a larger number of


false positive drugs with this test (see Willner, 1984, 1990). Cross-validation of animal models

Due to the fact that each of the models described thus far only partially resembles features of human depression, it may be helpful to compare the properties of several models using common measures. This technique was used in a recent study where brain glucose metabolism was compared in three depression models (amphetamine withdrawal, AMPT, and chronic stress), which were previously described in the literature. Each of the three models was previously reported to show decreased motor activity (Katz et al., 1981; Lynch and Leonard, 1978; Rech et al., 1966) and reduced reward behaviors (Cassens et al., 1981; Katz, 1982; Poschel and Ninteman, 1966). Amphetamine withdrawal, AMPT, and chronic stress also each induce depression in humans (Ainsman and Zacharko, 1982; Gillin et al., 1985; Watson et al., 1972). Each of these models resulted in an increase in GMR (glucose metabolic rate) in the lateral habenula with a decrease in the dorsomedial prefrontal cortex and anterior ventral thalamic nucleus (Calde- cott-Hazard et al., 1988). There was also a decrease in total forebrain GMR. The behavioral and GMR effects of the three methods were partially antagonized by tranylcypromine. The authors suggest that because the behavioral and 2DG changes were not idiosyncratic, they can be generalized to a state of “depressed exploratory behavior in the rat and may be analogous to human studies of glucose metabolism in bipolar patients who also show a decrease in global metabolic rate (Baxter et al., 1985).

Perhaps one of the most intriguing findings in the study of Caldecott-Hazard et al. (1988) was the increase in GMR in the lateral habenula after AMPT, amphetamine withdrawal, and chronic stress. The habenula receives fibers from the lateral pre-optic region of the hypothalamus and from the medial pallidum while sending projections to the substantia nigra, central gray and dorsal raphe, and reticular formation in the habenular peduncular tract. This pallido-habenular-nigral loop forms an accessory extrapyramidal circuit. De- creased limbic dopamine tends to decrease nonspecific locomotor exploratory behavior, whereas decreased ni- grostriatal DA tends to affect motor responding to in- centive and sequencing (e.g., stereotyped behavior) in the rat. Thus these three methods would be affecting DA systems similarly and be reflected in the decreased exploratory behavior.

In a study by Danysz et al. (1988), the efficacy of the tricyclic drugs desipramine and amitriptyline, and the selective serotonin uptake blockers zimelidine and alaproclate were compared in five different putative models of depression: the forced swim test (after Porsolt), inescapable electric shock, reversal of clonidine-induced hypothermia, a social dominancekompetition para-

digm, and instrumental learning using a DRL schedule. Desipramine and amitriptyline were found to be active in all of the tests used in this study. Zimelidine, which has been shown to be an effective antidepressant but which was withdrawn from the market because of ad- verse effects, was effective only in reducing shock- induced behavioral deficits, and in the DRL and social dominance tests (Danysz et al., 1988). Alaproclate, which has shown promise as an antidepressant in preliminary studies (Asberg et al., 1986), was effective in all except the social dominance test. Although the more general effects of the tricyclic drugs perhaps reflect their broader pharmacological spectrum of action, this type of multiple-comparison strategy may be useful in exploring the efficacy of prospective antidepressants as well as providing information regarding biochemical mechanisms involved in the models themselves. Genetic/environmental risk factor-based models

One of the drawbacks common to most, if not all, of the animal models detailed above is that they are based on pharmacological or behavioral interventions made in, presumbly, initially normal animals. However, a large body of evidence indicates that the predisposition to major depression is inheritable and hence not all individuals are equally at risk. Hence, one could argue that the development of more sophisticated and valid animal models is contingent upon the identification ind/or reproduction of presumed genetically determined predisposing factors in experimental animals. This type of approach has begun to be explored, for example, by Suomi et al. (1981) by attempting to identify behavioral and/or physiological characteristics in infant monkeys that reliably predict “anxious” or “depressed” features later in life. These investigators noted a strong correlation between heart rate reactivity in a classical condi- tioning paradigm performed in the third month of life in surrogate-reared rhesus monkeys and subsequent self- clasping behavior (an index of anxiety in these animals) over the ensuing year. Analysis of the genetic back- ground of these animals indicated a heritability of greater than 0.7 for this heart rate change measure. Such indices could, in principle, then be used for selective breeding of “anxious” animals.

Another model that has been examined in some detail is of rats bred for central cholinergic supersensitivity, analogous to a condition that may underlie human affective disorders (Overstreet, 1986; see review section on Acetylcholine).

A third model, in the common marmoset (C. jacchus jacchus), a New World primate, possesses several potential advantages over other models. First, their small size allows for the establishment of more nativelike social groups within the limited space of the laboratory set- ting. Second, they have strong ethological parallels with human behaviors (Johnson, 1990, or refs therein). For example, animals of this species tend to be monoga-


mous, a t least in captivity (Evans and Poole, 1983; Kleiman, 1977). Limited observations suggest that this is also the case in the wild. These animals also form strong pair-bonds (Abbott and Hearn, 1978; Box, 1975; Woodcock, 1982) and maintain a more or less nuclear family structure that more closely resembles the human social structure than does many of the Old World primates (Epple, 1975). Moreover, the period of psychoso- cia1 maturation seems prolonged relative to the animal’s life-span, and the fathers and older siblings appear to play a much more active role in rearing of the infant than in other nonhuman primate species (Abbott and Hearn, 1978; Kleiman, 1977). These animals have been noted to develop a wasting syndrome that is characterized in part by social withdrawal, listlessness, decreased feeding, weight loss, weakness and muscular dystrophy, hair loss, and ultimately death (Morin, 1982). Although a variety of factors have been implicated in this syndrome, including parasitosis and nutri- tional factors (Brack and Rothe, 1981; Morin, 19721, different modifications of the social conditions has been observed to either induce or ameliorate this syndrome (Johnson, 1990). In addition, limited observations (Gold P.W., unpublished) suggest that this syndrome may respond to ECS. In examining some of the factors that may predispose to the development of this syndrome, Johnson et al. ( 1990) have conducted longitudinal studies examining mother-child behavior from birth under both natural conditions and during imposed changes in the social structure, such as isolation and the formation of new peer groups. In these studies, physically abusive parental behavior toward infants was associated with reduced weight and stature and decreased play behavior a t later stages of life compared to peers who received more “caring” behaviors such as carrying, grooming, and social exploration. In postpubertal animals, morning plasma cortisol levels during a relatively un- stressed, pair-bonded condition was significantly associated with ultimate social status during later peer group formation. Differences were also found in the plasma cortisol response to isolation, during which the animals tended to show signs of wasting, according to ultimate dominance status.

This model, or others like it, may help identify behavioral andor biochemical risk factors originating early in development that predispose the animal to a vulnerability to later alterations in stress responsiveness. More- over, the closer similarity in the social structure of this species compared to that of humans may yield a closer concordance of “depressive” symptomatology with that of human depression than with other species.

Conclusions/future directions A number of the models mentioned above remain to be

fully explored, both from the point of view of refining the paradigms and testing of their validity in modelling the human syndrome of depression. For example, whereas

separation from the mother produces a syndrome that in some respects mimics human depression, it may not be representative of some of the more subtle forms of deprivation that may be more common predisposing factors to human depression than overt separation from the mother. Hence, it would be of interest to observe the long-range effects on social functioning of, for example, interventions in the mother, such as treatment with benzodiazepines or neuroleptics, which may blunt the mother’s responsiveness to the infant.

Another possible approach is suggested by recent advances in molecular biology and gene cloning. It has now become possible to identify specific genes through sequencing techniques, generate multiple copies of them through powerful cloning methods, and transfect them into fertilized oocytes from naive animals, which are then reimplanted into the host, producing “transgenic” animals whose native DNA structure has then been systematically altered. This technique has re- cently been used in the study of diabetes and other autoimmune diseases; mechanisms of oncogenesis; ret- roviral infection; and mechanisms of tissue-specific gene expression (Babinet et al., 1989; Gordon, 1989; Hinrichs et al., 1988). The still more recent application of these techniques in neuroscience (Rosenfeld et al., 1988), coupled with the identification and cloning of a number of critical genes in the neurochemical pathways implicated in human depression, suggests the possibil- ity of generating mutations in specific regions of these genes, systematically altering their regulatory properties. These types of regulatory mutations have been produced for members of the family of steroid receptor genes (Evans and Arriza, 1989). Such altered genes could then be transfected into oocytes and the animal’s behavior and physiology examined either naturalisti- cally, or in some of the behavioral and pharmacological paradigms mentioned above. In particular, responses to deprivation, inescapable shock, or other stressors could be examined to see if such animals were more vulnerable or sensitive to these stimuli than untreated animals. One could also test whether clinically useful antidepressant drugs could ameliorate any behavioral abnormalities observed. Whereas this technique depends in part on prior knowledge of which genes may be implicated in depression, there are a number of possible candidates already suggested by existing data.

The development of more useful animal models for depression, however, is in large part dependent upon the identification of robust and consistent biologcal markers based on neurochemical systems that have some comparability of function across species. Among these, the systems involved in mediating responsiveness to stress appear to be among the most promising, as adaptation to stress is a common feature of virtually all species and because of the similarity of the symptom complex of melancholic depression to that of acute responses to stress (Kling et al., 1989).


Biochemical research techniques in patients with depressive disorders (S. Caldecott-Hazard)

A wide variety of biochemical testsitechniques have been used in humans to study the pathological mechanisms underlying depressive disorders. Data cojlected from these tests are discussed in other sections of this review, particularly those on various transmitter/ receptor theories of depressive disorder. However, it is also useful to compare the tests themselves without regard to the data that are obtained. An overview of some of the tests, including individual ranges of applicability, rationales for usage, and problems, is valuable for understanding the data already collected and for developing new, more powerful analytical tools.

Table VIII is a summary of many of the more frequently used biochemical tests, what they measure, the source of the test sample, usual conditions of measurement, and their rationale for measurement. Although controversial, each of these tests has been reported at least once to show significant differences between depressed and normal individuals (see subsequent sections). With the exception of the relatively new technique, Positron Emission Tomography (see following section), the methods for performing each test are not described here. Readers are referred to some excellent articles or books for determining neurotransmitter concentrations in plasma and tissue (Abramson, 1974; Enna, 1980; Henry et al., 1975; Ziegler et al., 1976), hormone concentrations in plasma and urine (Lefkowitz et ai., 1970), changes in receptor densities and affinities (Bennett and Yamamura, 19851, or changes in CAMP response (Mann et al., 1985; Wang et al., 1974).

Neurotransmitter and metabolite concentrations Measurements of neurotransmitter and metabolite

concentrations in CSF, plasma, or urine (Table VIII) are some of the most frequently used tests for studying depressive disorders. Reasons for this popularity may be the relative ease of drawing samples, the ability to take repeated samples with or without preceding physiological challenges (e.g., changes in body position or the placement of a limb in cold water), and a variety of methods for quantitating the transmitters or metabolites. The decision as to which neurotransmitters andor metabolites are measured depends on a number of factors, including the hypothesis being tested, the estimated amount of each substance present in normals, and the sensitivity and accuracy of the assay technique €or each substance. Although assays for biogenic amines have been in use since the late 1940s (Von Euler, 1947), the accuracy and sensitivity of these techniques have increased greatly over the years, culminating with radioreceptor assays and radioimmunoassays (Enna, 1985). The rationale for the transmitter and metabolite tests is that they may be convenient, indirect measures of changes in brain neural activity that underlie symptoms of depressive disorders. Metabolites such as

MHPG were previously thought to origmate primarily in the brain and pass into the CSF, plasma, and urine. However, mare recent studies suggest that the propar- tion of MHPG that derives from the brain is lower than originally estimated and that it may be further metabo- lized in the periphery to vanillylmandelic acid (VMA) (Kopin et al., 1983; Mardh et aI., 1981)). Also, transmitters and other metabolites, especially in urine, are probably derived primarily from peripheral sources rather than the brain (Maas and Landis, 1968; Maas et al., 1973). Still, several studies have indicated that changes in these peripheral substances may, under a variety of circumstances, parallel changes in the same substances in the brain (Crawley et al., 1978; Kopin et al., 1983; Maas et al., 1973).

Some problems with the measurement of neurotransmitter and metabolite concentrations are that parallel changes in peripheral and cerebral substances may not occur in all depressive conditions and that the assump- tion of metabolites paralleling neural activity by reflecting the release of neurotransniitters may not be accurate (Siever, 1987). Furthermore, although data exist that concentration changes occur in depressed individuals as compared to normals, the direction of change differs from one report to another, and reports of “no differences” also exist for most data. Perhaps some of this variability is due to the fact that concentrations of transmitters and metabolites are not only determined by neuronal activity, but also depend on the clearance of these substances via reuptake into presynaptic neuron terminals, metabolism, and excretion (all of which may vary with individuals). Age, gender, diet, time of day and year, and activity may also influence concentrations and distribution of these substances in body fluids (Siever, 1987).

Neurotransmitter Concentrations have also been measured in postmortem brain from suicide victims, and the assay techniques are similar to those for concentration measurements in body fluids (Stanley et al., 1986). Whereas the postmortem technique may be a direct measure of brain changes, it has many problems that are similar to those encountered in tests of receptor changes in postmortem tissue. These issues are discussed under the heading on receptor changes.

Hormone concentrations The measurement of hormone concentrations in the

plasma, serum, or urine (Table VIII) is also thought of as a convenient, indirect measure of changes in brain neurons and transmitters (Halbreich, 1987). Hormone systems are known to be regulated by transmitters in the brain. For example, research has indicated that the release of corticotropin (ACTH) from the pituitary is stimulated by corticotropin-releasing hormone (CRH), which in turn is inhibited by norepinephrine and GABA, and enhanced by acetylcholine (Martin and Reichlin, 1987). Thus changes in these brain transmitters in depressive disorders are thought to cause altered con-

200 S. CALDECOT"r-HAZARD ET AL.

TABLE VIII. Examples of biochemical techniques in patients with depressive disorders' Source of Usual conditions Rationale for

What was measured sample of measurement measurement

I . Concentrations of: Neurotransmitters, Metabolites (a) Norepinephrine

Epinephrine Serotonin Endorphins GABA

(b) Norepinephrine Epinephrine Serotonin Endorphins GABA

(c) Norepinephrine EDineahrine

11.

111.

IV.

V

MHPG CSF HVA

5HIAA

MHPG

MHPG MET -~ VMA, NM

Concentrations of: Norepinephrine Serotonin Hormone Concentrations: Cortisol Melatonin

Pharmacological Challenges: Clonidine stimulation

of growth hormone release. TRH stimulation of TSH

release. Dexamethasone suppression

of cortisol release. 5-HTP stimulation of

ACTH or cortisol. Cholinergic agonist

stimulation of ACTH. h-CRH stimulation of

ACTH. Changes in Receptor Densities: (a) Alpha adrenergic

(b) Beta adrenergic receptors

receptors

(c) Muscarinic cholinergic

(d) Beta adrenergic

(e) Serotonin receptors

receptors

receptors

Imipramine binding sites

(f) Imipramine binding sites

VI. Changes in Transmitter Reuptake: Serotonin

VII. Changes in cAMP Response: Stimulation of cAMP by Isoproterenol.

Plasma

Urine

Brain tissue.

Plasma and urine

Plasma or serum

Blood platelets

Blood leukocytes, connective tissue

Fibroblasts

Fibroblasts Brain tissue. Brain tissue.

Platelets

Platelets

Lymphocytes

Single measures during depressed, manic, or euthymic states.

Multiple measures during circadian cycles.

Measures that follow physiological challenges; eg, cold stress or changes in body position.

Postmortem

Single measures during depression, mania, or euthymia.

Multiple measures during circadian cycles.

Single or multiple measures following the pharmacological challenge.


culture w/o receptor change

Fibroblasts can he grown in

Postmortem Postmortem


Single state measure or multiple circadian cycle measures.

Single measures during depression, mania, or euthymia.

Stimulation done in vitro.

Indirect measures of changes in brain transmitters; ie, brain metabolites such as MHPG may pass into CSF, plasma, and urine.

Other transmitters or metabolites may show changes in the periphery that parallel, but do not derive from changes in the brain.

Direct measures of changes in brain transmitters.

Indirect measures of changes in brain transmitters; ie, hormone systems are regulated by transmitters in brain.

These drugs affect neurotransmitters in brain that regulate hormone systems.

Changes in these various peripheral receptors may parallel changes in similar receptors in the brain.

Changes in these peripheral receptors may be genetic markers of depressive disorder.

Direct measures of changes in brain receptors.

These sites modulate the uptake of serotonin into platelets. Changes in peripheral sites may parallel similar changes in the brain.

Serotonin transporter in periphery is similar to that in brain.

Studies of receptor responsiveness. Results may parallel changes in adrenergic receptors in brain.

(continued)


TABLE VZZI. Examples of biochemical techniques in patients with depressiue disorders‘ (Continued)

Usual conditions Rationale for What was measured sample of measurement measurement VIII. Positron Emission

Tomography: Glucose metabolism Brain Single or multiple measures Direct, in vivo measures of

brain metabolism, blood Amino Acid Pools Brain Circulation and/or euthymia. flow, and receptor Relative receptor Measures following changes.

Source of

in depression, mania,

densities physiological stressors.

’Abbreviations: MHPG = 3Methoxy-4-hydroxy phenethyleneglycol; NM = norrnetanephrine; VMA = 3Methoxy-4-hydroxy-mandelic acid; HVA = Homovanillic acid; MET = Metanephrine; 5HIAA = 5Hydroxy-indoleaceticacid TRH = Thyrotropinreleasing hormone; TSH =Thyroid stirnulatinghormone; 5-HTP = 5Hydroxytryptophan; ACTH = Corticotropin; h-CRH = Human corticotropin releasing hormone.

centrations of hormones in the periphery. Also, these abnormal hormone systems are thought to have altered sensitivities to drugs that mimic the regulatory neurotransmitters (i.e., clonidine, an alpha-2 adrenergic agonist; or arecholine, a muscarinic, acetylcholine receptor agonist), or that are exogenously applied, hormone- releasing factors (i.e., CRH or thyrotropin-releasing hormone) (Meltzer, 1987).

Similar to measures of neurotransmitters and their metabolites, the sampling of hormones in plasma, serum, or urine can be done repeatedly with or without prior physiological or pharmacological challenges. Also, radioreceptor assays or radioimmunoassay can quanti- tate hormone levels in various body fluids. One of the problems with the hormone measurement tests is that, again, like neurotransmitter tests, the data are variable and conflicting. In fact, the dexamethasone test, “the most extensively evaluated bioloacal test in psychiatry” (Arana and Baldessarini, 1987) is probably the most controversial in terms of its usefulness for differentiat- ing subtypes of depressive disorder from one another o r from other psychiatric diseases.

Receptor densities Changes in the density of various receptors have been

studied in platelets and leukocytes of the blood, in fibroblast cells of the connective tissue, and in postmortem brain tissue (Table VIII). The rationale for measuring blood and connective tissue receptors is that some of the receptors, a t least, are pharmacologically similar to receptors in brain tissue (Nadi et al., 1984; Siever, 1987). Thus changes in peripheral receptors due to depressive disorders might parallel changes in brain receptors. Studies in patients and their relatives with a history of depressive disorders have also suggested that the abnormalities in similar receptors throughout the body could serve as global genetic markers for depressive disorder (Nadi et al., 1984).

The rationale for measuring brain receptors in postmortem tissue is that they may be direct measures of changes in various brain transmitter systems. In studies of depressive disorders, the brains of suicide victims are compared to the brains of normal individuals. Inter-

estingly, whereas the majority of suicide victims are retrospectively diagnosed as having had depressive symptoms, only approximately 50% are diagnosed as having had depressive disorder (Asberg et al., 1987).

Although most of the receptors measured in living or postmortem cells are for biogenic amines or acetylcholine, imipramine binding sites have also received a great deal of attention. These sites are thought to modulate the uptake of serotonin into neurons and platelets, and numerous reports have suggested that they may be abnormal in depression (Meltzer and Lowy, 1987). Whereas receptor studies have tended to utilize single samples of blood cells or fibroblasts, longitudinal studies of these cells could provide interesting, additional data. Fibroblasts have an added advantage of being grown in culture without altering the properties of the receptors. Thus multiple studies can be performed from the same cell sample.

Some of the problems with the various receptor binding tests are that, like all clinical biochemical tests in depressive disorder, there is a great deal of variability in the data, both when using the same test and when comparing data across tests. There is also the question of whether peripheral and central brain receptors really do function in parallel in certain diseases. For example, some studies suggest that beta adrenergic receptor densities are decreased in the periphery of depressed patients but that the same receptors in the brain of postmortem suicide patients are increased (Siever, 1987). Postmortem tissue studies from suicide victims are particularly vulnerable to variability. Although pre- cautions are taken when analyzing this tissue, the time of day when death occurred (receptors may cycle in their densities), the interval between death and tissue analysis, the patient’s age and gender, and drug history, and the region of brain studied may all affect the data (Asberg et al., 1987; Stanley et al., 1986). Variability in imipramine binding tests may, in part, be due to the use of ligand displacing agents that are relatively nonspecific for the serotonin transporter site (Marcusson et al., 1986). Although recent tests are using more specific displacing agents, these techniques have not yet been applied to the study of receptor changes in depressive disorders.


Changes in transmitter reuptake Serotonin reuptake into platelets has been reported to

be reduced in major depressive disorder both in single state tests and in circadian cycle studies. The serotonin transporter in platelets is thought to be similar to the serotonin transporter in neurons in the brain (Stahl, 1977). However, the serotonin reuptake techniques need additional testing. The variation in serotonin reuptake in platelets of normal individuals is great enough to limit the usefulness of this test by itself as a diagnostic for depressive disorder (Meltzer et al., 1983).

Changes in CAMP responses The in vitro application of beta adrenergic agonists

(i.e., isoproterenol) to lymphocytes causes an increase in the beta receptor coupled second messenger, CAMP (Extein et al., 1979; Mann et al., 1985; Pandey et al., 1979). Furthermore, the magnitude of the cAh4P response provides a measure of the receptor's functional responsiveness. Like receptor density studies, the relevance of this technique is that receptor changes in the periphery either may parallel changes in similar brain receptors or may serve as genetic markers of depressive disorder.

Unlike other biochemical techniques, the data from this test quite consistently show a decreased responsiveness to isoproterenol in depressed patients, even in the absence of changes in receptor densities (Extein et al., 1979). However, because these data conflict with reports of increased beta receptor binding in postmortem brains from suicide victims, the applicability of peripheral changes to central ones needs further clarification.

Positron emission tomography (PET) The unique aspect of this technique is its ability to

directly image and measure brain glucose metabolism, blood flow, amino acid pools, and receptor changes in vivo. Glucose is the primary energy source for the brain, and its metabolism thus reflects the functional activity of brain cells. PET scans can be repeated for longtudi- nal studies of either metabolism, blood flow, or receptor changes across different mental states in a patient. Conversely, if a patient is in one mental state for a prolonged period or has repeated episodes of the same state, changes in all of these modalities could be assessed during the same mental state.

Relatively few PET studies of patients with depressive disorders have been performed to date. Further- more, and perhaps also because of the scarcity of studies, the data are quite variable. Differences in the experimental groups, protocols of the experiments, and scanning technology all contribute to the variability. However, even within one research center, repeat studies have produced conflicting data (Baxter et al., 1985, 1989). Although this technique is clearly promising, additional studies are needed.

Conclusions As Siever (1987) discussed in his review article, re-

search on the biochemical basis of depressive disorders is changing its emphasis from relatively static techniques (single measure of concentrations of neurotransmitters, metabolites, and hormones in body fluids) to more dynamic techniques, such as longitudinal studies of concentration changes or concentration changes following physiological or pharmacological challenges. An- other major change in biochemical techniques has occurred through the development of radioactive isotopes and new technologies associated with them. Changes in receptor densities and affinities, transmitter reuptake, and brain glucose metabolism, amino acids, and blood flow can now be readily analysed. Even concentration techniques are more sensitive and versatile because of radioreceptor and radioimmunoassays; for example, the measurement of CAMP allows one to test the responsiveness of beta adrenergic receptors.

Unfortunately the data collected using most of these biochemical tests have been quite variable. At present, the results of these tests cannot be relied upon as bases for clinical diagnosis. Future studies, employing the techniques discussed in this review, need to emphasize greater standardization of protocols, patient groups and analysis technologes, both within and across research centers.

Positron emission tomography and mood disorders (B.H. Guze and L.R. Baxter)

Fundamental principles of PET scanning Positron Emission Tomography (PET) is an analytical

imaging technique that provides in vivo measurements of the anatomical distribution and rates of specific biochemical reactions (Phelps et al., 1979, 1985). It provides tomographic images and quantitative estimates of the distribution of compounds labeled with positron emitters (Phelps et al., 1982). Elements such as carbon, nitrogen, oxygen, or fluorine (substituting for hydrogen) exist as short-lived isotopes that decay by positron emission.

Most positron emitting isotopes have a half-life ranging between 2 and 110 minutes. Because of this, PET tracers usually are prepared on site. This requires a cyclotron and facilities for radiochemistry (Pawlik et al., 1986).

The radioactive substance emits a positron. Positrons act as positively charged electrons. The emitted positron travels a short distance (e.g., a fluorine positron travels approximately 2 mm) before combining with an electron. When this occurs, there is annihilation of the emitted positron by the electron. Two 511 KeV photons are produced by this annihilation reaction and travel in opposite directions.

The fundamental physical principle underlying PET imaging is coincidence detection. Detectors within the PET device, oriented at 180" to each other, detect these

DEPRESSION CLINICAL AND BIOCHEMICAL ASPECTS 203

511 KeVphotons almost simultaneously. To be counted, the circuitry of the PET device requires that the photons come from opposing directions within a time limit of about 5 to 20 nanoseconds. In addition, they must possess sufficient energy (usually 100 to 350 KeV) to activate the detectors. This is the process called coincidence detection.

Using a system of multiple detectors arranged in rings, these coincident events can be recorded with high efficiency and without further collimation. The nature of the PET device is such that it is able to recover a high percentage of the emitted activity and therefore is able to achieve a high definition image that is a visual representation of quantified data regarding the concentration of a labeled chemical in a given region.

For PET, the technique of image reconstruction is a mathematical method known as “linear superimposi- tion of filtered back projection.” This is the superimpo- sition of detected radiation, obtained from different angles, using appropriate mathematical processing to form tomographic images. It is the fundamental principle underlying all forms of computed tomography (including x-ray computed tomography [CT scanning]). These filtered back-projection algorithms permit the reconstruction of a comprehensive set of radioactivity distribution images covering the whole brain. Current spatial resolution is usually on the order of approximately 6 mm in all planes.

What these images represent depends upon the physical and chemical properties of the radiolabeled tracer and on the mathematical model used to quantify the physiological phenomena of interest. The concentration of radioactivity measured by PET can be converted into numerical estimates of the biochemical and physiologic processes governing the distribution of the labeled com- pound. This process of quantification requires tracer kinetic models-mathematical representations of the distribution of the tracer and its quantification by the images. The fluorodeoxyglucose (FDG) technique that is utilized for the determination of local cerebral glucose metabolism in humans is analogous to the 2-deoxyglucose autoradiographic method used in animals. FDG has unique properties different from glucose. After intravenous injection, FDG enters the brain and is sub- sequently phosphorylated by brain hexokinase. The metabolic product, 18 F-labelled 2-deoxyglucose-6- phosphate (FDG-6-P04), remains trapped in neurons with only slow dephosphorylation. This partly metabo- lized analog of glucose allows assessment of brain metabolism over a brief interval while the image is recorded by the scanner. Under nonstarvation conditions, along with oxygen, glucose is the predominate energy substrate for the brain; its metabolism reflects functional demand.

In a typical study, the patient receives an intravenous injection of FDG. For 30 minutes after the FDG is injected, blood sampling is performed. This blood is

either arterial or “arterialized venous blood. Blood is withdrawn in small samples for measurement of FDG. After approximately 30 minutes, between 85 and 95% of the neuronal uptake of FDG has occurred. It remains present in proportion to neuronal metabolism. As part of the process of image quantification, the time course of 18F activity and glucose concentration is measured in plasma. Brain activity is determined by scanning. With the knowledge of predetermined rate constants, an operational equation allows calculation of local cerebral metabolic rate for glucose distributed on a cross-sec- tional picture (Huang et al., 1980; Phelps et al., 1979).

The patient is moved into the scanner and imaging begins. Because little FDG is converted or further me- tabolized during the next 40-120 minutes when image acquisition occurs, the metabolic rate during the first 30 minutes after FDG injection remains stable for acquisition by the scanner. Typically, 2 million counts per plane are required to achieve an adequate image. From these measurements and a mathematic model developed by Sokoloff in 1977, the images composed of raw radioactivity counts are converted into the actual glucose metabolic rate. This is usually expressed in micromoles of glucose per 100 gm of brain tissue per minute.

The PET technique can be useful for studying brain functioning in patients with mood disorders. Although few of these studies have been performed, a review follows on what has been published on glucose metabolism, amino acid pools, brain circulation, and serotonin receptors in mood disorders.

PET studies of glucose metabolism Baxter et al. (1985) examined patients with affective

disorders and normal age-matched controls using the PET-FDG technique. Five bipolar depressed patients had documented prior manic episodes observed by the patient’s physicians. Three bipolar-mixed phase subjects likewise met DSM 111 criteria for bipolar-mixed state. Whole-brain metabolic rates were significantly lower in bipolar depressed patients than in all other groups except the bipolar mixed. The rates for bipolar depressed and mixed patients were similar. Whole brain metabolic rates for patients with bipolar depression increased, going from depression or a mixed state to a euthymic or manic state. In the regional analyses, bipolar depressed patients had significantly lower metabolic rates for glucose in frontal, temporal, occipital, and parietal lobes, as well as the cingulate, caudate, and thalamus, in both hemispheres, when compared with unipolar depressed patients. Normal controls had a significantly higher glucose metabolic rate than the bipolar depressed group in frontal lobes, caudate, and thalamus.

Buchsbaum (1984), in contrast, measured glucose metabolism in 10 bipolar and 1 unipolar patient. All patients were in the depressed state. There were 19 normal control subjects. Both normal subjects and pa-


tients showed an anterior to posterior gradient in absolute glucose use with the lowest metabolic rates in the occipital poles and the highest metabolic rates in the frontal poles. The bipolar depressives, in contrast to controls, showed a pattern of relative hypofrontality with, therefore, relatively diminished anteroposterior gradients.

In addition to their work with bipolar patients, Baxter et al. (1985) also examined cerebral metabolic rates for glucose in 11 patients with unipolar depression. Pa- tients showed psychomotor activity comparable with mania, but they had a markedly depressed mood or a mood that altered rapidly over a few hours between depression and irritability.

These patients showed a significantly lower ratio of the metabolic rate of the caudate nucleus divided by that of the hemisphere as a whole when compared to normal controls and patients with bipolar depression. When the five patients with unipolar depression were rescanned in the recovered state, the mean of this ratio increased for those patients who improved. Further- more, whole brain metabolic rates did not differ in unipolar patients as compared to controls.

In 1986 a NIMH group headed by Buchsbaum studied 16 patients with bipolar depression and 4 with unipolar depression, along with 24 normal controls, using PET FDG (Buchsbaum, 1986). Many of these were the same patients reported in Buchsbaum’s (1984) other publication. Data was analyzed using several techniques.

The first method of data analysis was the “cortical peel method.” In this approach, the outer contour ofeach PET slice was outlined with a computer boundary- finding technique. This outlined area was divided to create a total of eight pie-shape sectors. Bipolar patients had significantly lower frontal to occipital cortex glucose metabolic ratios than normal controls. This has been termed a relative hypofrontality in bipolar illness. Furthermore, the ratio ofthe glucose metabolic rate in a sector to that of the whole slice showed bipolars to have lower ratios than normal in frontal cortex but no significant differences in occipital cortex. Thus relative hypofrontality in bipolars was attributable to relative frontal decreases rather than relative occipital increases.

The next method of data analysis was the region of interest method, analyzing metabolic rate in each of the following structures: caudate nucleus, putamen, anterior ventrolateral thalamus, median ventrolateral thalamus, posterior ventrolateral thalamus, superior frontal gyrus, globus pallidus, inferior putamen. When metabolic rates in individual brain structures were expressed as a ratio to whole slice metabolic rate, both unipolar and bipolars had decreased glucose use in the basal ganglia as a whole and in the caudate as compared to normal controls. This relative reduction was most striking in the right caudate. Clinical depression rat- ings correlated negatively with whole slice metabolic

rate. As mentioned previously, Baxter et al. (1985) also reported reduced ratios of caudate to whole brain metabolism in unipolar depressives. Unlike Buchsbaum, Baxter did not find this result in bipolar patients.

The next method of data analysis was global glucose consumption. Values of glucose use were calculated according to the Sokoloff model. Global metabolism was found to be significantly higher in subjects with affective illness (both unipolar and bipolar depressive) compared to normal controls. This finding contrasts with the report by Baxter et al. (1985) who found decreased global metabolism in depressed bipolar patients compared to normal controls or unipolar depressives. The latter group also found that the low global metabolism of the depressed state increased to normal levels in the euthymic state. One factor that may account for the differences in global metabolic findings was the psychological state or cognitive task that the subject was engaged in at the time of FDG uptake. Three other studies that measured cerebral blood flow without the PET technique found decreased global blood flow in patients with unipolar depression (Gustafson et al., 1981; Mathew et al., 1980; Rush et al., 1982).

Bipolar and unipolar depressives could be differenti- ated by anteroposterior gradients of cortical glucose metabolism (Buchsbaum et al., 1986). Bipolar depressives had the lowest gradient; normals had an interme- diate gradient, and unipolar depressives had the great- est gradient of the three groups. The primary cause of this change in gradient was metabolism in the frontal pole. Bipolars showed reduced metabolism in this pole as compared to normals or unipolars. Baxter et al. (1985) initially also found a greater decrease in frontal lobe glucose metabolism in bipolars relative to normal controls or unipolars. Later, Baxter et al. (1989) found that unipolar and bipolar depressives both had decreased lateral left prefrontal lobe metabolism as compared to normals. Different environmental conditions and stimuli during uptake may account for some of the differences in results.

Specifically, in the 1989 study Baxter et al. examined regional cerebral glucose metabolism in 10 patients with unipolar depression, 10 with bipolar depression, 10 suffering from obsessive-compulsive disorder (OCD) with secondary depression, 14 afflicted with OCD without major depression, and 12 normal controls (Baxter et al., 1989). Six patients with mania were also studied. Depressed patients were matched for depression on the Hamilton Depression Rating Scale, and subjects with OCD without depression and OCD with depression had similar levels of OCD pathology. The middle frontal gyrus (dorsal anterolateral prefrontal cortex) was identified in all tomographic planes for each subject. This value was “normalized by dividing it by the metabolic rate of the ipsilateral hemisphere.

The authors had hypothesized a priori that a cerebral dysfunction would need to fulfill the following criteria in

205 DEPRESSION: CLINICAL AND BIOCHEMICAL ASPECTS

order to be considered a true state-dependent cerebral abnormality in depression.

1. The dorsal anterolateral prefrontal cortex to hemisphere ratio (ALPFC-hem) would be similar in both bipolar and unipolar depressions of equivalent severity and significantly different from that observed for normal controls and subjects with OCD but without major depression.

2. The ALPFC-hem would distinguish OCD without depression from OCD with secondary major depression and, likewise, would distinguish bipolar mania from bipolar depression.

3. The degree of ALPFC-hem deficit would correlate significantly with standard measures of depression severity.

4. This ratio would change significantly in the direction of normal values in the same individuals on resolution of the depressed state. They decided, again a priori, to examine the ALPFC-hem because of previous 133 Xe blood flow studies and other evidence.

The left ALPFC-hem satisfied all of these hypothesized criteria for a cerebral abnormality in depression, whereas the right ALPFC-hem did not satisfy three of these criteria completely. The ALPFC-hem was similar in the primary depressions (unipolar depression, bipolar depression) and were significantly lower than those in normal controls or OCD without depression. Results for the right hemisphere were similar for these comparisons. Values in subjects with OCD with depression were also significantly lower than in subjects with OCD without depression, but only on the left. Values in subjects with bipolar depression were lower than those in manic subjects on this measure in the left hemisphere, although results were not significant for the right hemisphere.

There was a significant negative correlation between the clinical depression severity scores (Hamilton De- pression Rating Scale, or HAM-D (Hamilton, 1967)) and the left ALPFC-hem. With improvement of depression, the left ALPFC-hem metabolic rate increased significantly, and the percentage change in the Hamilton score gave a significant correlation with the percentage in the left ALPFC-hem. For normal controls, bipolar depressives and unipolar depressives, the diagnostic groups on which the HAM-D was originally standardized, there was a significant negative correlation with the left and right ALPFC-hem. Likewise, the HAM-D and the left ALPFC-hem showed a significant negative correlation for a group combining patients with OCD with and without depression. There was not a significant correlation on the right.

Finally, a study by Post et al. (1987) reported on glucose utilization in the temporal lobes of 13 affectively ill patients as compared with 18 normal volunteer controls. These, too, were patients who were part of the initial NIMH series and some have, therefore, been

reported before. The patients consisted of 5 depressed, 6 euthymic, and 2 manic individuals. They were not the same patients studied across different states. Maximum glucose use relative to any other glucose maximums in the same PET image plane were calculated. In depressed patients, this relative measure was reduced on the right temporal lobe compared to normal volunteers. Temporal lobe glucose metabolism also showed progres- sive increases with improvement or euthymia and further increases in the two manic patients studied. These findings were consistent with those of Baxter et al. (1985). Alternations in left-right temporal lobe metabolic rates were felt to be an area for further investigation in light of the wealth of data, utilizing a variety of methodologies, indicating a role for hemispheric later- ality in affective disorders.

PET studies of brain amino acid pools in mood disorders

The PET carbon 11 (1lC) glucose method has been reported to image amino acid pools (especially glutamic acid derived from a-ketoglutaric acid in the citric acid cycle) in the brain. Three groups of researchers (Hara and Izuchi, 1983; Ilio et al., 1983; Jones et al., 1983) found that this radioactive label was primarily localized in brain amino acids at 30 minutes following the oral administration of 11C-Glucose in rats. Kishimoto et al. (1987) applied this technique to image amino acid pools in the brains of patients with mood disorders. In this project, 9 unipolar depressed patients were compared to 3 manic patients and 3 patients in remission from depression. There were 7 normal control subjects. The carbon 11 glucose counts in the brain of manic patients were increased about 25% compared to controls. The PET images of unipolar depressed patients indicated decreased amino acid pools compared to controls; carbon 11 counts in the brains of unipolar depressed patients were decreased by about 35% compared to controls. The carbon 11 counts in the brain of patients in remission were equivalent to those in normal controls. Amino acid metabolism was examined in frontal, temporal, parietal, and occipital cortices.

There were no differences in the 11C blood counts for controls or patients. Indeed, the uptake of 11C glucose from the blood to the brain was similar in all populations. Therefore, this did not appear to be an explanation for the differences observed among the groups.

There was a disturbed amino acid pool in the cortex of unipolar and bipolar patients. C11 glucose when em- ployed in the method of Kishimoto et al. (1987) shows amino acid pools, whereas the F18-deoxyglucose method shows the utilization of glucose itself. These differences in the behavior of the isotopically labelled compounds would explain the difference between Kish- imoto’s results and those of Buchsbaum and of Baxter where the FDG technique was used.


PET studies of brain circulation RaichIe et al. (1985) studied 5 patients with major

depression where one patient had depressed bipolar disorder. The control subjects for the study were normal volunteers.

Cerebral whole hemisphere blood flow (CBF) and cerebral whole hemisphere metabolic rate for oxygen (CMRO2) were significantly decreased in the patients with depression, compared to normal controls. Cerebral hemispheric blood volume (CBV) did not differ significantly from the normal controls. PET studies of serotonin receptors

Spiperone is both a dopamine-2 receptor (D2) ligand and a serotonin-2 (5HT-2) ligand. When spiperone is tagged with 18F and used in patients, the identities of the receptors that are measured are assumed from prior knowledge of whether dopamine or serotonin receptors predominate in the brain region of interest. Mayberg et al. (1988) used spiperone to study serotonin receptors of stroke victims. Higher 5HT-2 receptor binding was reported in the right parietal and temporal cortex of 9 patients with right hemisphere strokes when compared to 8 patients with left hemisphere strokes or to 17 normal control subjects. The ratio of binding in the ipsilateral to the contralateral cortex showed a significant negative correlation with the severity of depression in the left temporal cortex. Many of these patients were suffering from an organic mood disorder-poststroke depression. These findings suggest that the biochemical response of the brain may be different depending upon which hemisphere is injured and that some depressions may be a consequence of the failure to up-regulate serotonin receptors after stroke.

Problems with the PET-mood disorder studies To date, PET findings in mood disorders come from

relatively few studies that have used widely different methods, making cross-study interpretations difficult. In fact, only in the case of FDG has there been more than one group reporting anything even approaching similar experiments. Although the investigations of Buchs- baum et al. (1986) and Baxter et al. (1985) could be interpreted as showing similar findings in lateral prefrontal cortex and both groups report basal ganglia glucose metabolic dysfunction in various types of depression, a clear picture does not emerge. In all fairness it should be pointed out that all the reports reviewed here were of experiments designed to examine different questions. Nevertheless, it is lamentable that there have been no true replication attempts of any PET mood disorders experiments published to date. This is in contrast to studies of schizophrenia (Buchsbaum et al., 1982; Cohen et al., 1987; Farkas et al., 1984; Jernigan et al., 1985; Kling et al., 1987; Volkow et al., 1986; Wolkin et al., 1988), and even obsessive-compulsive

disorder (Baxter et al., 1987,1988; Nordahl et al., 1989; Swedo et al., 1989) where replication studies within groups and near replication studies between groups have been reported, allowing a better estimation of “the truth” than is possible for PET mood disorder studies at this time.

There are problems associated with PET in particular and with studies of psychiatric problems in general that are obstacles in all of these studies of mood disorders. These problems can be divided into: (1) patient selection and classification, (2) assessment of patient internal and external conditions at time of study, (3) PET instrumentation and biochemical calculations, and (4) statistical problems.

Patient selection and classification. Mood symptoms of many types are common in the general population, and selecting subjects with homogeneous syndromes on one salient variable (primary diagnosis) while excluding an uncontrolled representation of comorbid pathology (secondary diagnosis) in the same sample is difficult. All study samples examined have been “convenience samples”-drawn from those presenting at a particular institution-rather than true random samples of the general disease population. Thus results cannot be generalized safely, and significant site-to-site sample differences are likely. In this regard, lifetime drug use of all types may be a major limiting factor.

Further, different depressive diseases may manifest similar symptoms. In this regard, an individual may go for many years classified as “unipolar” only to have a late-life manic episode and thus switch polarity. Selec- tion of “normals” is also a problem with the high prevalence of psychiatric disorders in the general population and the relative ease with which many symptoms can be concealed. Those presenting for studies may be sicker than the general population (Amori and Lenox, 1989). Structured diagnostic interviews and diagnostic criteria help but do not eliminate the problem, especially when it comes to the question of meaningful subtypes.

Finally, individuals with mood disorders often show spontaneous state changes over time-more so than is typical of other illnesses such as schizophrenia. Season of the year and even time of day (Sack et al., 1987) may also be important variables, not to mention history of drug exposure. Subtle differences in diagnostic groups samples, as well as mood and other state variations, even within the same individual, may well give differing results.

Assessment ofpatient internal and external conditions at time ofstudy. It is well demonstrated that alterations in sensory input from the environment to the subject can produce large focal changes in glucose metabolism (Phelps et al., 1986). The studies reported here used a wide range of stimulation states-eyesfears, oped closed; nonspecific resting state/electrical shocks to the arm, etc. At first glance, the solution might seem to


employ specific behavioral tasks, with a quantifiable behavioral output with which to measure the subject’s involvement with the process. However, it has been demonstrated that even in the same behavioral task (music memory) and with the same behavioral outcome (scores on memory for a musical sequence), normal individuals using different internal, covert strategies to solve a problem (silent humming versus musical nota- tion) activate different brain areas (right vs. left temporal cortex) (Mazziotta et al., 1982). Thus a psychiatric patient doing a task such as the Wisconsin Card Sort, which usually activates the lateral prefrontal cortex in normals (Weinberger et al., 1986), may fail to activate this same area, not because of any problem in the brain structure itself, but because the psychiatric subject chooses to think about the problem in an unusual way. The difference obtained on testing between patient and control subjects may have nothing to do with the basic neural dysfunction in the disease but rather may reflect a difference in psychology determined by another, uni- dentified brain regon. Even the question of what would be a relevant task for both mood-disordered patients and controls-one that elicits the critical pathology, involving the relevant circuits, and gives an accurate and meaningful behavioral measure-is not clear.

Regardless of the range of states represented in a subject group, simply documenting the subject’s mood state at the time of PET scanning can itself by problematic. Most studies report mood scores for the episode on an instrument like the Hamilton Depression Rating scale, but exact mood at time of scanning is rarely documented. One approach might be to have subjects completing various mood inventories at the time of scanning, but this maneuver in and of itself might induce cerebral processes, and in different ways in different experimental and control groups by processes that have to do with psychological mind-set but not primary psychopathology.

PET instrumentation and biochemical calculations. The intrinsic resolution and time-response capacities of various tomographs, the merits of measured versus calculated attenuation corrections, arterial versus “arterialized venous blood, advantages and disadvantages of the particular tracers used, etc., are beyond the scope of this brief review. Interested readers are referred to Phelps et al. (1986). Suffice it to say that all of these factors vary across the studies reviewed.

More fundamental, the meaning of several measures of spiperone binding relative to certain brain reference structures has been questioned even for the D2 receptor, where there is an affinity much higher than that for the serotonin receptor. Since both serotonin and D2 receptors are found in the temporal cortex, the relevance of spiperone binding to actual densities of serotonin receptors is open to question. Even with PET-FDG, the best validated and most used of all present PET methods

(Bartlett et al., 1988; Brooks et al., 1987; Phelps et al., 1986) has uncertainties. Although there is general agreement on glucose metabolic rate values for the normal human brain, calculations of absolute metabolic rates vary by approximately lo%, even in the same individual in the nonspecific stimulation state done at different times on consecutive days with the same ma- chine and technique (Bartlett et al., 1988). When regional values are “normalized” to whole brain rates, however, the variation is on the order of only 1% (Bartlett et al., 1988). Whether the time-to-time intra- subject variations observed for absolute metabolic rates are a true reflection of nature or are due to inaccuracies of the method is not clear. However, it must be acknowl- edged that many of the most problematic variables in the calculation of absolute glucose metabolic rates are eliminated when normalization is carried out. However, even though the brain is an interactive organ, the exact meaning of normalized metabolic rates is itself not clear (Clark et al., 1989). This would be a problem even if all groups used the same normalization procedure, which they do not.

Statistical problems. Comparing experiments, if the same results are found in different studies, the chance of false-positive (Type I) error may be low, but obviously the risk of false-negative (Type 11) error is great in the small “n” studies, which characterize a complicated, expensive technique such as PET. Within individual experiments, multiple comparisons of many brain regions that are not totally independent are often under- taken, making the probability of Type I error hard to calculate, whereas multivariate methods used to control for multiple differences among subjects increase Type I1 error. Most PET studies of psychiatric illnesses to date are a statistician’s nightmare-many approaches to the data analyses can be argued, all giving different results.

PET studies of mood disorders are still in their in- fancy. Given their high prevalence and their great costs in both human and economic terms, there will surely be much more of this work in the next few years. Studies of neuroreceptor systems would seem particularly promising.

REFERENCES Abbott, D.H., and Hearn, J.P. (1978) Physical, hormonal, and behav-

ioral aspects of sexual development in the marmoset monkey, Cal- lithrix jacchus. J. Reprod. Fert., 53:155-166.

Abramson, F.P. (1974). Femtomole level of analysis of biogenic amines and amino acids using functional group mass spectrometry. Anal. Biochem., 51:482499.

American Ps chiatric Association (1987) DSM-IIIR: Diagnostic and Statistical Lanual of Mental Disorders, 3rd ed. Washindon D.C.: - American Psychiatric Association.

trials? J. Clin. Psychopharm. 9:321-327.

In: Neurobiolo eds. Williams Wilkms, . Baltimore, pp. 407431.

influence of stress. Behav. Brain Sci., 589-137.

Amori, G., and Lenox, R.H. (1989) Do volunteer subjects bias clinical

Anisman, H. (1984) Vulnerability to depression: Contribution of stress of Mood Disorders. R.M. Post and J.C. Ballenger,

Anisman, H., and Zacharko, R.M. (1982) Depression: the predisposing

208 S. CALDECOV-HAZARD ET AL.

Anisman, H., Irwin, J., and Sklar, L.S. (1979) Deficits of esca e erformance following catecholamine de letion. Implications f%

gehavioural deficits induced by uncontrolfable stress. Psychophar- macolo 64:163-170.

Arana, G%. , and Baldessarini, R.J. (1987) Clinical use of the dexamethasone suppression test in psychiatry. In: Psychopharmacology: The Third Generation of Progress. H.Y. Meltzer, ed. Raven Press, NewYork, p 609416.

Asberg, M., $)e;tilsson, R., Rydin, E., Schalling, D., Thoren, P., and Traskman-Bendz, L. (1980) Monoamine metabolites in cerebrospinal fluid in relation to depressive illness, suicidal behavior, and personality. In: Recent Advances in Neuropsycho harmacology. B. Angrist, G.D. Burrows, M. Lader, 0. Lingjaerde, 8. Sedvall, and D. Wheatley, eds. Pergamon, New York, p 257-271.

Asberg, M., Eriksson, B., Martensson, b.; Traskman-Bendz, L., and Wagner, A. (1986) Therapeutic effects of serotonin uptake inhibitors in depression. J. Clin. Ps chiatry, 47 Suppl.:23-35.

Asberg, M., Schalling, D., daskman-Bendz, L., and Wa er, A (1987) Psychobiology of suicide, impulsivity, and related pgnomena. In: Psychopharmacology: The Third Generation of Progress. H.Y. Melt- zer, ed. Raven Press, New York, pp. 655-668.

Babinet, C., Morello, D., and Renard, J.P. (1989) Transgenic mice. Genome, 31:938-949.

Banki, C.M., Molnar, G., and Vojnik, M. (1981) Cerebrospinal fluid aminemetabolites, t r ptophan, and clinical parameters in depression. 11. Psychopathoggical s

Bartlett, E.J., Brodie, J.D., Wolg .P . , Christman, D.R., Laska, E., and Meissner, M. (1988) Reproducibility of cerebral glucose metabolic measurements in resting human subjects. J. Cereb. Blood Flow Metab., 8:502-512.

Baxter, L.R., Phelps, M.E., Mazziotta, J.C., Guze, B.H., Schwartz, J.M., and Selin, C.E. (1987) Local cerebral glucose metabolic rates in obsessive-com ulsive disorder. A comparison with rates in unipolar depression an: in normal controls. Arch. Gen. Psychiatry, 441211- 218.

Baxter, L.R., Phelps, M.E., Mazziota, J.M., Schwartz, R.H., Gerner, R.H., Selin, C.E., and Sumida, R.M. (1985) Cerebral metabolic rates for h o s e in mood disorders studied with ositron emission tomog- rapf y and (F-18) fluoro-deoxyglucose (FDC!). Arch. Gen. Psychiatry, 42:441-447.

Baxter, L.R., Schwartz, J.C., Mazziotta, J.C., Phel s M E , Pahl, J.J., Guze, B.H., and Fairbanks, L. (1988) Cerebraf&cose metabolic rates in non-depressed obsessive-compulsives. Am. J. Psychiatry, 1451560-1563.

Baxter,L.R., Schwartz, J.M., Phelps,M.E., Mazziotta, J.C., Guze, B.H., Selin, C.E., Gerner, R.H., and Sumida, R.M. (1989) Reduction of prefrontal cortex h o s e metabolism common to three types of depression. Arch. &n. Psychiatry, 46:243-250.

Beckmann, H., and Goodwin, F.K. (1980) Urinary MHPG in subgroups of depressed patients and normal controls. Neuropsychobiology, 6:91-100.

Benfield, P., Heel, R., and Lewis, S.P. (1986) Fluoxetine: A review ofits pharmacod amic and harmacokinetic properties, and therapeutic efficacy in gpressive ilfness. Drugs, 32:481-508.

Bennett, J.P., and Yamamura, H.I. (1985) In: Neurotransmitter Rece tor Binding, 2nd ed. H.I. Yamamura, S.J. Enna, and M.J. Kuhar, e&i Raven Press, New York, pp. 61-89.

Bowlby, J . (1952) Maternal and mental health. W.H.O. Monograph Series N. 2.

Bowlby, J . (1961) Childhood mourning and its implications for psychiatr Am J Psychiatry, 118:481-498.

Box, h. (1975) A social developmental study of young monkeys (Cal- lithrix ‘acchus) within ca tive family groups. Primates, 16:419435.

Brack, d., and Rothe, H. 8981) Chronic tubulointerstitial nephritis and wasting disease in marmosets (Callithrix jacchus). Veterinary Pathology Suppl., 6:45-54.

Breier, A., Albus, M., Pickner, D., Zahn, T.P., Wolkowitz, O.M. and Paul, S.M. (1987) Controllable and uncontrollable stress in humans: Alterations in mood and neuroendocrine and psychophysiological function. Am. J. Psychiatry, 1443419-1425.

Bricca, G., Dontenwill, M., Molines, A., Feldman, F., Belcourt, A,, and Bousquet, P. (1989). The imidazoline preferring receptor: binding studies in bovine, rat, and human brainstem. Eur. J . Pharmacol.,

Brooks, R.A., Hatazawa, J., Chiro, G.D., Larson, S.M., and Fishbein, D.S. (1987) Human cerebral glucose metabolism determined by i$0sitron emission tomography: A revisit. J . Cereb. Blood Flow

Brown, R.P., Stoll, P.M., Stokes, P.E., Frances, A., Sweeney, J., Kocsis, J.H., and Mann, J.J. (1988) Adrenocortical hyperactivity in de res sion: Effects of agitation, delusions, melancholia, and other i lkess variables. Psychiatry Res., 23:167-178.

ptoms J Affect. Disord., 3:91-99.

162:l-9.

etab. 7:427432.

Buchsbaum, M.S., Wu, J., DeLisi, L.E., Holcomb, H., Kessler, R., Johnson, J., King, A.C., Hazlett, F., Langstron, K., and Post, R.M. (1986) Frontal cortex and basal banglia metabolic rates assessed by positron emission tomography with [18F] 2-deoxyglucose in affective illness. J. Affective Disord. 10:137-152.

Buchsbaum, M.S., DeLisi, L.E., Holomb, H.H., Cap elletti, J , King, A.C., Johnson, J., Hazlett, E., Zimmerman, S.D., Fost, R.M:, Mori- hisi, J., Carpenter, W., Cohen, R., Picka, R.D., Weinberger, D.R., Margolin, R., and Kessler, R.M. (1984) Anteroposterior gradients in cerebral glucose use in schizophrenia and affective disorders. Arch. Gen. Psychiat. 41:1159-1166.

Buchsbaum, M., Ingvar, D., Kessler, R., Waters, R.N., Ca pelletti, J., van Kammen, D.P., King, A.C., Johnson, J., Manning, K.G., Flynn, R.W., Mann, L.S., Bunney, W.E., J r , and Sokoloff, L. 11982) Cerebral glucography with ositron tom0 ra hy Use in normal subjects and in patients with sckizophrenia. l r c k Gen. Psychiatry, 39:251-259.

Caldecott-Hazard, S., Mazziotta, J., and Phelps, M. (1988) Cerebral correlates of depressed behavior in rats, visualized using 14C- 2-deoxyglucose autoradiography. J. Neuroscience 8(6): 1951-1961.

Carroll, B.J., Feinberg, M., Greden, J.F., Tarika, J., Albala, A.A., Haskett, R.F., James, N.M., Dronfol, A,, Lohr, N., Steiner, M., de Vigne, J.P., and Young, E. (1981) A specific laboratory test for the diagnosis of melancholia. Arch. Gen. Ps chiatry, 38:15-22.

Cassens G., Actor C., Kling M., and Schihkraut J.J. (19811 & het amine withdrawal: effects on threshold of intracranial reinirce: ment. Psychopharmacol., 73:318-322.

Clark, C., and Stoess, L.A. (1986) Glucose use correlations: A matter of inference. J . Cereb. Blood Flow Metab. 6511-512.

Clark, C., Klonoff, H., Tyhurst, J.S., Li, D., Martin, W., and Pate, B.D. (1989) Re ’onal cerebral glucose metabolism in three sets of identical twins witfpsychotic symptoms. Can. J . Psychiatry, 34:263-270.

Clark, C., Carson, R., Kessler, R., Margolin, R., Buchsbaum, M., DeLisi, L., Kmg, C., and, Cohen, R. (1985) Alternative statistical models for the examination of clinical PET-FDG data. J. Cereb. Blood Flow Metab. 5:142-150.

Cohen, R., Sem le, W., Gross, M., Nordahl, T.E., DeLisi, L.E., Holcomb, H.H., King, RC., Morihisa, J.M., and Pickar, D. (1987) Dysfunction in a refrontal substrate of sustained attention in schizophrenia. Life Jciences, 40:2031-2039.

Cornfeldt, M., Fisher, B., and Fieldin S (1982) Rat internal capsule lesion: a new test for detecting anti% ressants Fed Proc 41.1066

Coryell, W., Gaffney, G., and Burkharg, P.E. (1982) DSM-fII melan: cholia and the primary-secondary distinction: a comparison of con- current validity by means of the dexamethasone suppression test. Am. J. Psychiatry, 139:120-122.

Costa, E., Garattini, S., and Valzelli, L. (1960) Interactions between reserpine, chlorpromazine, and imipramine. Experientia, 16:461- 463.

Crawley, J.M., Hattox, S.E., Maas, J.N., and Roth, R.H. (1978) 3- Methoxy-4-hydroxyphenethyleneglycol increase in plasma after stimulation of the nucleus locus coeruleus. Brain Res., 141:380-384.

Danysz, W., Plaznik, A., Kostowski, W., Malatynska, E., Jarbe, T.U.C., Hiltunen, A.J., and Archer, T. (1988) Comparison of desi ramine, amitriptyline, zimelidine, and alaproclate in six animal moiels used to investigate antidepressant drugs. Pharmacol. Toxicol. 62:42-50.

DeLisi, L. Holcomb, H., Cohen, R., Pickar, D., Car enter, W., Morihisa, J., King, A.C., Kessler, R., and Buchsbaum, h.S. (1985) Positron emission tomography in schizophrenic atients with and without neurole tic medication. J. Cereb. Blood h o w Metab., 5201-206.

Desan, P.R., Silbert, L.H., and Maier, S.F. (1988) Long-term effects of inescapable stress on daily running activity and antagonism by desipramine. Pharmacol. Biochem. Beh. 30:21-29.

Dorworth, T.R., and Overmier, J.B. (1977) On “learned helplessness”: The therapeutic effects of electroconvulsive shock. Physiol. Beh.

Egeland, J.A., Gerhard, D.S., Pauls, D.L., Sussex, J.N., Kdd, K.K., Allen, C.R.., Hostetter, A.M., and Housman, D.E. (1987) Bipolar affective disorders linked to DNA markers on chromosome 11. Nature, 325:783-787.

Enna, S.J. (1980) Radioreceptor assafs. In: Physico-Chemical Method- oloaes in Psychiatric Research. Hanin and S.H. Koslow, eds. Raven Press, New York, pp. 83-101.

Enna, S.J. (1985) Radioreceptor assays. In: Neurotransmitter Receptor Bindin 2nd ed. H.I. Yamamura, S.J. Enna, and M.J. Kuhar, eds, Raven press, New York, pp. 177-188.

Epple, G. (1975) The behavior of marmoset monkeys (Callitrichidae). In: Primate Behavior, Vol. 4. LA. Rosenbloom, ed. Academic Press, New York, pp. 195-239.

Evans, R.M., and Arriza, J.L. (1989) A molecular framework for the actions of glucocorticoid hormones in the nervous system. Neuron,

Evans, S., and Poole, T.B. (1984) Long-term changes and maintenance

41355-358.

2:1105-1112.

DEPRESSION CLINICAL AND BIOCHEMICAL ASPECTS 209

of the pair-bond in common marmosets, Callithrix jacchus jacchus. Folia Primatologxa, 42:3341.

Extein, J., Tallma, J., Smith, C.C., and Goodwin, F.K. (1979) Changes in lymphoc te beta adrenergic receptors in depression and mania. Psychiatr. Jes., 1:191-197.

Farkas, T., Wolf, A., Jaeger, J., Brodie, J.D., Christman, D.R., and Fowler, J.S. (1984) Regional brain glucose metabolism in chronic schizophrenia. Arch. Gen. Psychiatry, 41:293-300.

Freud, S. (1917) Mourning and melancholia. In: Collected Papers, Vol. 4. Hogarth Press, London, p 152 172

Gershon, E.S., Berrettini, W., kurncerger, J., and Goldin, L.R. (1987) Genetics of affective illness. In: Psychopharmacology: The Third Generation of Progress. H.Y. Meltzer, ed. Raven Press, New York, P I . 48,1492.

Gi in, J C , Sitaram, N., Wehr, T., Duncan, W., Post, R., Mur hy, D.L., Mendelson, W.B., Wyatt, J., and Bunne W.E. (1984) &eep and affective illness. In: Neurobiology of MoodiDisorders. R.M. Post and J.C. Ballenger, eds. William & Wilkins, Baltimore, p 157 189

Glazer, H.I., and Weiss, J.M. (1976) Long-term intedrence effect: an alternative to “learned helplessness.” J . Exp. Psychol.: h i m . Beh. Proc., 2201-213.

Gold, P.W., Goodwin, F.K., and Chrousos, G.P. (1988a) Clinical and biochemical manifestations of de ression: Relation to the neurobiology of stress (part 1 of 2 parts). J e w Engl. J. Med., 319:348-353.

Gold, P.W., Goodwin, F.K., and Chrousos, G.P. (1988b) Clinical and biochemical manifestations of de ression. Relation to the neurobiol-

of stress (part 2 of 2 parts). d w Engl. J. Med., 319:413420. G t3on , J.W. (1989) Transgenic animals. Int. Rev. Cytol., 115121-229. Gower, A.J., and Marriott, A.S. (1980) The inhibition of clonidine-

induced sedation in the mouse by antidepressant drugs. Br. J. Pharmacol., 69:287P-288P.

Green, A.R., Heal, D.J., Lister, S., and Molyneux, S. (1982) The effect of acute and repeated desmethylimi ramine administration on clonidine-induced hy oactivity in rats. %r. J. Pharmacol., 75:33P.

Gustafson, L., Risterg, J., and Silfverskiold, P. (1981) Regional cerebral blood flow in organic dementia and affective disorders. Adv. Biol. Psychiatry, 6:109-116.

Halbreich, U. (1987) Hormones and depression- conce tual transi- tions. In: Hormones and Depression. U. Halbreich, ed. gave, Press, New York, pp. 1-20.

Hamilton, M. (1967) Development of a rating scale for primary depressive illness. Br. J. SOC. Psychol. 6:278-296.

Hara, T., and Izuchi, R. (1983) The study of local cerebral lucose metabolism with positron CT. Bull. of Ministry Health and #elfare, Japan, 15:15.

Harris, J.C. (1989) Experimental animal modeling of depression and anxiet Ps chiatric Clin. N. h e r . 122315-836.

Henry, &., darman , B., Johnson, D., and Williams, R. (1975) A sensitive radio-enzymatic assay for norepinephrine in tissue and plasma. Life Sci., 16:375-384.

Hinrichs, S.H., Vo el, J , Rhim, M.J., and Jay, G. (1988) Use of transgenic anim3s to study human retroviruses. Cancer Metast. Rev., 7:311-320.

Hrdina, P.D., von Kulmiz, P., and Stretch, R. (1979) Pharmacolo ’cal modification of experimental depression in infant macaques. %:sy- chopharmacolo 64:89-93. ~

Huang, S.C., Pheg i , M.E., Hoffman, E.J., Sideris, K., Selin, C.J., and Kuhl. D.E. (1980) Non-invasive determination of local cerebral metabolic rate of glucose in man. Am. J . Physiol., 238:E69-E82.

Hughes, C.W., Kent, T.A., Campbell, J . , Oke, A., Croskill, H., and Preskorn, S.H. (1984) Cerebral blood flow and cerebrovascular permeability in an inescapable shock (learned helplessness) animal model of depression. Pharmacol. Biochem. Behav., 21:891-894.

Ilio, M., Suzuki, T., and Hattori, H. (1983) The measurement of cerebral lucose metabolic rate with 11C-glucose. Nuc. Med., 20:998.

Jensen, C . 8 , and Tolman, C.W. (1962). Mother-infant relationship in the monke , Macaca Nemestrina: the effect of brief separation and mother-infYant specificity. J . Comp. Physiol. Psychol., 55:131-136.

Jernigan, T., Sargent, T., and Pfefferbaum, A. (1985) Fluorodeoxyglu- cose PET in schizophrenia. Ps chiatry Res., 16:317-329.

Johnson, E.O. (1990) The biobezavioral consequences of psychogenic stress in the common marmoset (Callithrix jacchus jacchus). Doc- toral dissertation, Universit of Mar land

Jones, S.C., Ackerman, R.H., d o p , B., 6aron, J.C., Brownell, G.L., and Taveras, J.M. (1983) Brain uptake and or an distribution of 11C from 11C-labeled elucose. Int. J. Nuc. Med. fiiol.. 10:173.

Kelsoe, J.R., Ginns, E.I., Egeland, J.A., Gerhard, D.S., Goldstein, A.M., Bale, S.J., Pauls, D.L., Lon , R.T., Kidd, K.K., Conte, G., Housman, D.E., and Paul, S.M. (1989)ie-evaluation ofthe linkage relationship between chromosome l l p loci and the gene for biiolar affective disorder in the Old Order Amish. Nature, 342:238-243.

Kiloh, L., and Garside, R. (1963) The inde endence ofneurotic depression and endo enous depression. Br. J. Fsychiatry, 109:P451463.

Kishimoto, H., ‘kkazu, O., Ohno, S., Yamapchi,.T., Fujita, H., Kuwa- hara, H., Ishii, T., Matsushita, M., Yo 01, S , and 110, M. (1987) 1 IC-glucose metabolism in manic and depressed patients. Psychia- try Res., 2231-88.

Kleiman, D. (1977) Monogamy in mammals. Quart. Rev. Biol., 52:39- 69.

Klerman, G.C. (1983) The scope of depression. In: The Ori ‘ns of De ression: Current Concepts and Approaches. J. Angst, ed. &ringer-berla ,Berlin,p 5 25

Kling, A,, Ferryman, k., a n d Parks, T. (1984) A telemetry system for the study of cortical-amygdala relationships in the squirrel monkey (S. Sciureus). In: Modulation of Sensorimotor Activity During Alter- nations in Behavioral States. Alan R. Liss, New York, p. 351-365.

Klin , A,, Metter, J., Riege, W., and Kuhl, D.E. (1986) dmparison of P#T measurement of local brain glucose metabolism and CAT measurement of brain atrophy in chronic schizophrenia and depression. Am. J. Psychiatry, 143375-180.

Kling, M.A., Perini, G.I., Demitrack, M.A., Geracioti, T.D., Linnoila, M., Chrousos, G.P., and Gold, P.W. (1989) Stress-res onsive neuro- hormonal systems and the symptom complex of affective illness. Psychopharmacol. Bull., 25312-318.

Kokkinidis, L., Zacharko, R.M., and Predy, P.A. (1980) Post-amphetamine depression of respondin from the substantia nigra: Reversal by tricyclic antide ressants Pfarmacol. Bioch. Beh., 13:379-383.

Kolata, G. (1987) ibanic-depression gene tied to chromosome 11. Science, 235:1139-1 140.

K o p , I.J., Gordon, E.K., Jinerson, D.C., and Polinsky, D.J. (1983) elation between plasma and cerebrospinal fluid levels of 3-meth-

oxy-4-h droxyphenyl lycol Science 219:73-75. Koslow, ZH., Mass, J.B., Bowden, C.L., Davis, J.M., Hanin, I., and

Javaid, J. (1983) CSF and urinary biogenic amines and metabolites in depression and mania. Arch. Gen. Psychiat., 40:999-1010.

Kraemer, G.W., and McXlnney, W.T. (1979) Interactions of pharmaco- logcal agents which alter biogenic amine metabolism and depression. J. Affective Disord. 1:33-54.

Kraepelin, E. (1921) Manic-Depressive Insanity and Paranoia. E.&S. Livingstone, Edinburgh.

Kuhl, D.E. (1984) Imaging local brain function with emission computed tomography. Radiology, 150:625-631.

Lefkowitz, R., Roth, J., and Pastan, 1. (1970) Radioreceptor assay of adrenocorticotropic hormone: new approach to assay of polypeptide hormones in plasma. Science, 170:633-635.

Lehnert, H., Reinstein, D.K., Strowbridge, B.W., and Wurtman, R.J. (1984) Neurochemical and behavioral consequences of acute, uncontrollable stress: Effects of dietary tyrosine. Brain Res., 303:215-223.

Leith, N.J., and Barrett, R.J. (1976) Amphetamine and the reward s stem. evidence for tolerance and post-drug depression. Psycho- piarmacology, 4639-25.

Leonard, B.E., and Tuite, M. (1980) Anatomical, Physiological, and Behavioral Aspects of Olfactory Bulbectomy in the rat. International Review of Neurobiology, 22:251-287.

Leonhard, K., Korff, I., and Shulz, H. (1962) Die temperamente in den Familien der monopolaren und bipolaren phasischen Psychosen. Psychiatr. Neurol., 143:416-434.

Leshner, A.J., Remler, H., Bi eon, A., and Samuel, D. (1979) Effects of desmethylimipramine (DdI) on learned helplessness. Psychophar- macolo 66:207-213.

Lewis, J.!?‘: and McKnney, W.T. (1976) Effects of electroconvulsive shock on the behavior of normal and abnormal rhesus monkevs. Behav. Psychiatg,.37:687-693.

Liebowitz. M.R.. ein. D.F.. Quitkin. F.M.. Stewart. J.W.. and Mc- Grath, P.J. (1984) Clinical implications of diagnostic subtypes of de ression. In: Neurobiolo of Mood Disorders. RM Post and JC Baylenger, eds. Williams &%ilkins, Baltimore, pp. 107-120.

Lloyd, C. (1980) Life events and depressive disorder revisited. I. Events as predisposing factors. Arch. Gen. Psychiatry, 37:529-535.

Loranger, A.W., and Levine, P.M. (1978) Age at onset of bipolar affective illness. Arch. Gen. Psychiatr , 3 5 1345-1348.

Lynch, M.A.. and Leonard, B.E. (1978) Jffect of chronic amphetamine Katz, R.J. (1981) L i m a 1 models and human depressive disorders

Neuroscience and Biobehavioral Reviews, 5:241-246 Katz, R.J. (1982) Animal model of depression: pharmacological sensi-

tivit of a hedonic deficit. Pharmacol. Biochem. Behav., 16:965-968. Katz, J .J . , Roth, K.A., and Carroll, B.J. (1981)Acute and chronic stress

effect on open field activity in the rat: implications for a model of depression. Neuroscience and Biobehavioral Reviews, 5:247-251.

norepinephrine in thecentral nervous s stem JPET, 163:147-162. Maas, J.W., Dekirmen’ian, H , Garver, d., Redmond, D.E., Jr, and

Landis, D.H. (1973) ixcretion of catecholamine metabolites follow-

-adm’inistration on the behavior of rats in the open field apparatus: Reversal of post-withdrawal depression by two antidepressants. J Pharm. Pharmacol.. 30:79€&799.

Maas. J.W.. and Landis. D.H. (19681 In vivo studies of metabolism of


ing intraventricular injection of 6-hydroxydopamine in the Macaca speciosa. Eur. J . Pharm., 23:121-130.

Maier, S.F. (1984) Learned helplessness and animal models of depression. Prog Neuro-Psychopharmacol & Biol Psychiatry, 8:435446.

Maier, S.F., and Jackson, R.L. (1979) Learned helplessness: all of us were right (and wrong): inescapable shock has multiple effects. In: The Psychology of Learning and Motivation, Vol. 13, G. Bower, ed. Academic Press, New York, 155-218.

Maier, S.F., and Seligman, M.f!P. (1976) Learned helplessness: The- ory and evidence. J. Exp. Psychol. Gen., 106:4446.

Maguire, G.P. (1988) The psychiatric interview. In: Companion to Psychiatric Studies. R.E. Kendell and A.K. Zealley, eds. Churchill Livin stone, Edinbur h, p 200

Mann, f.J., Brown, R.$., i a i e r , ' J P , Sweeney, J.A., Kocsis, J.H., Stokes, P., and Bilezikian, J$. (1985jReduced sensitivity of lympho- cyte beta-adrenergic receptors in atients with endogenous depression and psychomotor agitation. 8 En 1 J Med., 313:715-720.

Marcusson, J . , Backstrom, I., and Ross, S% (1986) Single-site model of the neuronal 5-hydroxytryptamine uptake and imipramine binding site. Mol. Pharmacol., 30:137-145.

Mardh, G., Sjoquist, B., andhggard , E. (1981)Norepinephrine metabolism in man using deuterium labelling: The conversion of 4-hydroxy-3-methoxyphenylglycol to 4-hydroxy-3-methoxymandelic acid. J. Neurochem., 36:1181-1185.

Martin, J.B., and Reichlin,. S. (1987) Regulation of adrenocorticotropic hormone (ACTH) secretion and its disorders. In: Clinical Neuroen- docrinology. J.B. Martin and S.A. Reichlin, eds. Davis, Philadelphia, pp. 159-191.

Mathew, R.J., Mayer, J.S., Frances, D.J., Semshuk, K.M., Mortel, K., and Cleghorn, J.L. (1980) Cerebral blood flow in depression. Am. J. Psychiatry, 137:1449-1450.

Matsui, T., and Hirano, A. (1978) An Atlas of the Human Brain for Corn uterized Tomography. Igaku-Shoin, Tokyo,

MaxweTl, D.R., and Palmer, H.T. (1961) Demonstration of anti-depres- sant or stimulant properties of imipramine in experimental animals. Nature, 191:84-85.

Ma berg, H.S., Robinson, R.G., Wong, D.F.: Parikh, R., Bolduc, P., darkstein, S.E., Price, T., Dannals, R.F., Links, J.,M., Wilson, A.A., Ravert, H.T., and Wagner, H.N. (1988) Pet imagmg of cortical S2 serotonin receptors after stroke: Lateralized changes in relationship to depression. Am. J. Psychiatry, 145:937-943.

Mazziotta, J.C., Phelps, M.E., and Carson, R.E. (1982) Tomographic mapping of human cerebral metabolism: Auditory stimulation. Neu- rology, 32:921-937.

hlazziotta, J.C., Phelps, M.E., Carson, R.E., and Kuhl, D.E. (1982) Tomographic mapping of human cerebral metabolism: Sensory deprivation. - h n . NeurolT, 12:435444.

McKinnev, W.T., Jr.. and Bunnev, W.E. 11969) Animal model of depresiion: review of evidence and implications for research. Arch. Gen. Psychiatry, 21:240-248.

McKinney, W.T. Jr., Moran, E.G., and Kraemer, G.W. (1983) Effects of drugs on the response to social separation in rhesus monkey. In: Hormones, Drugs, and Social Behavior in Primates. H.D. Stek is and Arthur S. Kling, eds. Spectrum, New York, Pp. 249-270.

Meltzer, H.Y. (1987) An overview of hormones and monoamine receptors. In: Hormones and Depression. U. Halbreich, ed. Raven Press, New York, pp. 49-58.

Meltzer, H.Y., and Lowy, M.T. (1987) The serotonin hypothesis of depression. In: Psychopharmacology: The Third Generation of Progress, H.Y. Meltzer, ed. Raven Press, New York, pp. 513-526. .

Meltzer, H.Y., Arora, R.C., Tricou, B.J., and Fang, V.S. (1983) Seroto- nin uptake in block platelets and the dexamethasone suppression test in depressed patients. Psychiat. Res., 8:41-48.

Mendels, J., and Cochrane, C. (1968) The nosology of depression: The endogenous reactive concept. Am. J. Psychiatr., 124 (Suppl): 1-11.

Minor, T.R., Jackson, R.L., and Maier, S.F. (1984) Effects of task- irrelevant cues and reinforcement delay on choice-escape following inescapable shock: evidence for a deficit in selective attention. J . Exp. Ps chol Anim. Beh. Proc., 10:543-556.

Minor, T.&, Pel'l mounter, M.A., and Maier, S.F. (1988) Uncontrolla- ble shock, fore<rain norepinephrine, and stimulus selection during choice-escape learning. Psychobiology, 16:135-145.

Morin, M. (1982) A different approach in examining the wasting syndrome. Lab. Animal, 16:36-39.

Nadi, N.S., Nurnberger, J.I., and Gershon, E.S. (1984) Muscarinic cholinergic rece tors on skin fibroblasts in familial affective disorder. New En 1 j. Med., 311:225-230.

Nordahl, T.E., henkelfat, C., and Semfle, W:E. (1989) Cerebral glucose metabolic rates in obsessive-compu sive disorder Neuropsychophar- macology, 2:23-28.

Nurnberger, J.I., and Gershon, E.S. (1984) Genetics of affective disorders. In: Neurobiology of Mood Disorders. R.M. Post and J.C. Bal- lenger. eds. Williams & Wilkins, Baltimore, pp. 7E-101.

Osman, T.O., Rudorfer, M.V., and Potter, W.Z. j1989) Idazoxan: A selective a1 ha2 antagonist and effective sustained antidepressant in two bipofar depressed patients. Arch. Gen. Psychiatry, 46:958- 959.

Overmier, J.B., and Seligman, M.E.P. (1967) Effects of inesca able shock on subsequent escape and avoidance learning. J. Comp. Fhy s- iol. Psycho., 163:23-33.

Overstreet, D.H. (1986) Selective breeding for increased cholinergic function: Development of a new animal model of depression. Biol. Ps chiatry 21 49 58

Pandky, G.N'., Dysken; M.W., Gamer, D.L.,. and Davis, J.M. (1979) Beta-adrenereic receotor function in affective illness. Am. J. Psvchi- atry, 136:675-678. A

Parini, A. Cou ry, I., Graham, R.M., Uzielli, I., Atlas, D., and Lanier, S.M. (1489) Eharacterization of an imidazoline/guanidinium recep- tive site distinct from the alpha-2-adrenergic receptor. J. Biol. Chem., 264:11874-11878.

Pawlik, G., Beil, C:, Hebold, I:, Herholz, K., Wjenhard, K., and Heiss, W.D. (1986) Positron emission tomography in depression research: Principles-results-perspectives. Psychopathology, 19(Suppl), 2~85-93.

Paykel, E.S. (1983) Recent life events and depression. In: The Origins of Depression: Current Concepts and Approaches. J. Angst, ed. Springer-Verlag, Berlin, p 91 106

Petty, F., and Sherman, A.L. '(1980) Reversal of learned helplessness by imi ramine. Comrnun. Psycho harmacol., 3:371-373.

Phelps, k E . , Mazziotta, J.C., a n f Schelbert, H.R. (1986) Positron Emission Tomography and Autoradiography: Principles and Appli- cation for the Brain. Raven Press, New York.

Phelps, M.E., Mazziotta, J.C., Baxter, L., Schwartz, J., and Gerner, R. (1985) Study design in the investigation ofmood disorders with PET. In: Brain Imaging and Brain Function. L. Sokoloff, ed. Raven Press, New York, pfi1227-243.

azziotta, J.C., Gerner, R., Baxter, L., and Kuhl, D.E. (1983) Human cerebral ducose metabolism in affective disorders:

Phelps, M.E.,

Drug-free states and ph~rmacologic effects. J. Cereb. Blood Flow Metab. 3(Su pl) S7 S8

Phelps, M.E., iuang,s.C:, Hoffman, E.J., Selin, C.J., Sokoloff, L., and Kuhl, D.E. (1979) Tomographic measurement of local cerebral glucose metabolic rate in humans with (F18) 2-fluoro-2-deoxyglucose: Validation of method. Ann. Neurol., 6:371-388.

Pickar, D., Cowdry, R.W., Zis, A.P., Cohen, R.M., and Murphy, D.L. 11984) Mania and hypomania during antidepressant pharmacotherapy: clinical and research implications. In: Neurobiolo of Mood Disorders. R.M. Post and J.C. Ballenger, eds. Williams% Wilkms, Baltimore, pp. 836-845.

Porsolt, R.D. (1983) Pharmacological Models of Depression. In: The Ori ins of De ression: Current Concepts and Approaches. J. Angst, ed. ipringer-ferlag, New York p 313 330

Porsolt. R.D.. Anton. G.. Blavet. @. and Yalfre. M. (1978) Behavioral despair in rats: A new model sensitive to antidepressant treatments. Euro ean Journal of Pharmacology, 47:379-391.

Poschet B., and Ninternan, R.W. (1966) Hypothalamic self-stimulation: its suppression by blockade of norepinephrine synthesis and reinstatement by metham hetamine. Life Sci., 5:ll-16.

Post, R.M., DeLisi, L.E., Horcomb, H.H., Uhde, T.W., Cohen, R., and Buchbaum. M.S. (1987) Glucose utilization in the temuoral cortex of affectively Ill patients: Positron emission tomographi. Biol. Psychi- atrv. 22545-553.

d l ~~ - ~ - - ~ - Potter, W.Z., Schleinin, M., Golden, R.N., Rudorfer, M.V., Cowdry,

R.W., Calil, H.M., Ross, R.J., and Linnoila, M. 11985) Selective antidepressants and cerebrospinal fluid: lack of specificity on nor- e p i n e h i n e and serotonin metabolites. Arch. cen . Psvchiatrv. - . 42:11?1-1177.

Prien. R.F.. Kuufer. D.J.. Manskv. P.A.. Small. J.G.. Tuason. V.B.. Voss, C.B'., an& Johnson, W.E. (1984) Drug therapyin the prevention of recurrences in unipolar and bipolar affective disorders. Arch. Gen. Ps chiatr 41 1096-1104.

Raiczle, M.$:, Ta lor, J.R., Herscovtch, P., and Guze, S.R. (1985) Brain circulation anc?metabolism and depression. In: The Metabolism of the Human Brain Studied with Positron Emission Tomography. T. Greitz et al., (eds.) Raven Press, New York, pp. 453456.

Raiei h, M J , Brammer, G.L., McGuire, M.T., Yuwiler. A,, Geller, E., a n f Johnson, C.K. (1983) .Social status related differences in the behavioral effects of drugs in vervet monkeys. In: Hormones, Dru s and Social Behavior in Primates. H.D. Steklis and S. Kling, ejs; S ectrum, New York, pp. 83-105.

Re&, R.H., Borys, H.K., and Moore, K.E. (1966)Alterations in behavior and brain catecholamine levels in rats treated with alpha-methyl- para-tyrosine. J. Pharmacol. Exp. Ther., 153:412419.

Redmond, D.E., Jr. (1983) Social effects of alterations in brain noradrenergic functions on untreated grou members. In: Hormones, Drugs, and Social Behavior in Primates. h.D. Steklis and A.S. Kling,

21 1 DEPRESSION: CLINICAL AND BIOCHEMICAL ASPECTS

eds. Spectrum, New York, 271-279. Redmond, D.E., Hinrichs, i.!., Maas, J.W., and Kling, A. (1973)

Behavior off free ranging macaques after intraventricular 6-hydroxydopamine. Science, 1813256-1258.

Redmond, D.E., Maas, J.W., Iiling, A., and Dekirmenjian, H. (1971a) Changes in primate social behavior after treatment with alpha- methyl-para-tyrosine. Psychosomatic Medicine, 33:97-113.

Redmond, E.E. Jr . , Maas, J.W., Kling, A., Graham, C.W., and Dekir- menjian, H. (1971b) Social behavior of monkeys selectively depleted of monoamines. Science, 174:428-431.

Reite, M.L., Short, R., Seiler, C., and Pauley, J.D. (1981) Attachment, loss, and depression. J. Child Psychol. Psychiat., 22:141-169.

Reite, M.L., Anders,T.F., Greil, W., Heilhammer, D., Henn, F.A., Katz, R.J., Kaufman, I.C., Kraemer, G.W., Linden, M., McGuire, M.T., McKinney, W.T., Nissen,G., andPorsoft, R.D. (1983lAnimal models: group re ort In The Origins of De ression: Current Concepts and Approacies. J . Angst, ed. S ringer-beria , New York, pp. 405423.

Robson, R.D., Antonaccio, M.J., Saelens, J.E., and Liebman, 3. (1978) An tagonism by mianserin and classical alpha-adrenoceptor blocking drugs of some cardiovascular and behavioral effects of clonidine. Eur. J. Pharmacol., 47:431442.

RosenfeId, M.G., Crenshaw, E.B. III, and Lira, S.A. (1988) Transgenic mice: applications to the study of the nervous system. Ann. Rev. Neurosci., 11:353-372.

Rush, D.K., Mineka, S., and Suomi, S.J. 11983) Therapy for helpless monkevs. Behav. Res. Ther.. 21:297-301.

Rush, A.J., Schlessor, M.A., Stokely, E., Bonte, F.R., and Altshuler, K.Z. (1982) Cerebral blood flow in depression and mania. In: Brain Imaging in Psychiatry and Neurolo Positron Emission Tomo a ghy and Other Techniques. M.S.%uchsbaum, E. Eusdin, #EI

unne , Jr , and D. Ingvar, eds. Boxwood Press, Pacific Grove, CA. Sack, D.J., Rosenthal, N.E., Parry, B.L., and Wehr,T.A. (1987)Biolog-

ical Rhythms in Psychiatry. In: Psychopharmacology: The Third Generation of Progress. H.Y. Meltzer, ed. Raven Press, New York,

Schdkraut, J.J., Orsulak, P.J., Schatzberg, A.F., Gudeman, J.E.,Cole, J.O., Rohode, W.A., and LaBrie, R.A. (19781 Toward a biochemical classification of depressive disorders: I. Differences in urinary excretion of MHPG and other catecholamine metabolites in clinically defined subtypes of depression. Arch. Gen. Psychiatry, 35:1427- 1433

p. 669-685.

Schmidt, A.W., and Peroutka, S.J. (1989) 5-Hydroxytryptamine receptor "families." FASEB J., 3:2242-2249.

Schwartz, J.M., Baxter, L.R., Mazziotta, J.C., Gerner, R.K.H., and Phelus. M.E. (19871 The differential diagnosis of denression. Rele- vanck of positron emission tomography &dies of cerebral glucose metabolism to the bipolar-unipolar dichotomy. JAMA, 258:1368- 1374.

Seay, B., Hansen, E., and Harlow, H.F. (1962) Mother-infant separation in monkeys. J . Child Ps chol Psycbiat., 3:123-132.

Seligman, E.P., and Weiss, I M . (1980) Coping behavior: Learned helplessness, physiological change and learned inactivity. In: Behav- ioral Research and Therap , Vol18. M. Weinraub and A. Schulman, eds. Pergamon Press, Lodon, 459-512.

Seltzer, V., and Tonge, S.R. (1975Pbethylamphetamine withdrawal as a model for the depressive state: Anta onism of post-amphetamine depression by imiprarnine. J. Pharm. P%armacoL, 27:16P.

Sherman, A.D., Sacquitne, J.L., and Petty,,F. (1982) Specificity of the learned helplessness model of depression. Pharmacol. Biochem. Behav., 16:649-454.

Siever, L.J. (1987) Role of noradrenerqc mechanisms in the etiology of the affective disorders. In: Psvchoagarrnacoloet: The Third Genera- tion of Progress. H.Y. Meltier, ed. Raven P"cess, New York, pp.

Silvestrini, B. (1982) Trazodone-a new txpe of. antidepressant: A discussion of pharmacological data and t eir clinical implications In: Typical and Atypical Antidepressants: Molecular Mechanisms E. Costa and G. Racagni, eds. Raven Press, New York, pp. 327-340.

Sokoloff, L. (1977) Relation between physiolo .cal function and energy metabolism in the central nervous system. FNeurochem., 29:13-26.

Spitz, R.A. (1946) Anaclitic depression. Psychoanal. Study Child,

Stahl, S.M. (1977) The human platelet: A diagnostic research tool for the study of biogenic amines in psychiatric and neurologc disorders. Arch. Gen. Psychiatry, 34:509-516.

Stanley,M., Mann, J.J . , and Cohen, L.S. (1986) Roleofthe serotonergic system in the postmortem analysis of suicide. Psychopharmacol. Bull., 22:735-740.

Stedman, T.L. (1978) Stedman's Medical Dictionary, 23rd ed. Williams and Wilkins, Baltimore.

Strother, S., Allard, C., and Moeller, J. (1987) Methodological factors effecting patterns of regional cerebral glucose metabolism as determined by 18F-fluorodeoxyg~ucose positron emission tomography. J. Cereb. Blood Flow Metab.. 7:S443.

493-504.

2: 3 13-342.

Suomi,S.J.,Kraemer,G.W.,Baysinger,C.M., andDeLizio,R.D.(1981) Inherited and experiential factors in anxious behavior displayed by rhesus monkeys. In: Anxiety: New Research and Chan 'ng Con- cepts. D.F. Klein and J. Rabkin, eds. Raven Press, New%ork, pp.

Suomi, S.J., Seaman, S.F., Lewis, J.K., DeLizio, R.B., and Mchnney, W.T. (1978) Effects of imipramine treatment on separation-induced social disorders in rhesus monkeys. Arch. Gen. Psychiatry, 35:321- 325.

Swedo, S.E., Schapiro, M.B., Grady, C.L., Chesiow, D.L., Leonard, H.L., Kumar, A., Friedland, R., Rapoport, S.F., and Rap0 ort, J L (1989) Cerebral lucose metabolism in childhood-onset ogsessive: compulsive disorfer. Arch. Gen. Psychiatry, 46518423.

Szewczak, M.R., Fielding, S., and Cornfeldt, M. (1982) Rat internal capsule lesion: further characterization of antidepressant screening potential. SOC Neurosci. Abstr., 8:465.

Theobald, W., Buch, O., Kunz, H., Morsurgo,,C., Stenger, E.G., and Wilhelmi, G. (1964) Com arative stu ies with Tofranil, Pertofran, and Endison. Arch. Inc. Piarmacodyn. Ther., 148:560-569.

Van Riezen, H. i 1972) Different central effects of the 5-HT antagonists mianserin and cyproheptadine. Arch. Int. Pharmacodyn. Ther., 1 98:256-269.

Volkow, N., Brodie, J . , Wolf, A., Gomez-Mont., F., Cancro, R., Van Gelder, P., Russell, J.A., and Overall, J. (1986): Brain organization in schizophrenia. J. Cereb. Blood Flow Metab. 6:441446.

Von Euler, U.S. (1947) A specific sympathomimetic ergone in adrenergic nerve fibers (sym athin) and its relations to adrenaline and noradrenaline. Acta Pgysiol. Scand., 12:73-97.

Wang, Y.C., Pandey, G.N., Mendels, J., and Frazer, A: 11974) Platelet adenylate cyclase responses in depression: Implications for a receptor defect. Ps chopharmacolo 'a 36 291 300

Watson, R.E., dartman, E., and!&ildkraG, J.J. (1972) Amphetamine withdrawal: affective state, sleep patterns, and MHPG excretion. Am. J . Psychiatry, 129:263-269.

Wehr, T.A., and Goodwin, F.K. (1979) Rapid cycling in manic-depressives induced by tricyclic antidepressants. Arch. Gen. Psychiatry,

Weinberger, D.R., Berman, K.F., and Zee, R.F. (1986) Physiologc dysfunction of dorsolateral prefrontal cortex in schizophrenia: I. Regional blood flow evidence. Arch. Gen. Psychiatry, 43:114-124.

Weiss, J.M., Goodman, P A , Losito, P.G., Corrigan, S., Charry, J., and Bailey, W. (1982) Behavioral depression produced by an uncontrolled stressor: relation to norepinephrine, dopamine, and serotonin levels in various reBons of the rat brain. Brain Res. Rev., 3:167-205.

Weiss, J.M., Bailey, W.H., Goodman, P.A. Hoffman, L.J., Ambrose, M.J., Salman, S., and Charry, J.M. (1982) A model for neurochemical study of d e y s i o n . In:,Behavioral Models and the Analysis of Drug Action. M. Spiegelstein and A. Levy, eds. Elsevier, Amsterdam, pp.

Weissman, M.M., and Boyd, J.H. (1984) The epidemiolo of affective disorders. In: Neurobiolo of Mood Disorders. R.M. Kst and J.C. Bailenger, eds. Williams gWilkins, Baltimore, pp. 60-75.

Weissman, M.M., and merman, G.L. (1977) Sex differences in the e idemiology of depression. Arch. Gen. Psychiatry, 34:98-111.

Wilen, L., Blomquist, G., Greitz, T., Litton, J.E., Bergstrom, M., Ehrin, E., Ericson, K., Eriksson, L., In ar, D H , Johansson, ,L., Nilsson, J.L.G.. Stone-Elander. S.. SedvarG.. Wiese1.F.. and W1k.G. 119831

179-200.

36:555-559.

195-223.

Pet studies of glucose mktabolism in patients with schizophrenia. Am. J . Neuroradiology, 4:550-552.

Willner, P. (1984) The validity of animal models of depression. Psy- chopharmacolo 833-16.

Willner. P. (1990)%tnal models of denression: an overview. Pharmac. Ther.; 4k425-455.

Wolkin, A,, Angrist, B., Wolf, A., Brodie, J.D., Wolkin, B., Jaeger, J. , Cancro. R.. and Rotrosen. J . (1988) Low frontal ducose utilization in chronic schizophrenia. A replication study. -Am. J . Psychiatry,

Wolkin, A., Jaeger, J., and Brodie, J. (1985) Persistence of cerebral metabolic abnormalities in chronic schizophrenia as determined by positron emission tomography. Amer. J. Psychiatry, 142564471.

Woodcock, A.J. (1982) The first weeks of cohabitation of newly formed- heterosexual pairs of common marmosets (Callithrix jacchus). Folia Primatolo ca 37.228 254

Zacharko, R S . , BowersrW.J:, Kokkinidis, L., and Anisman, H. (1983) Region-s ecific reductions of intracranial self-stimulation after un- controllagle stress: Possible effects on reward processes. Behav. Brain Res., 9:129-141.

Zacharko, R.M., Bowers, W.J., and Anisman, H. (1984) Res onding for brain stimulation: Stress and desmethyiimipramine. Log. Neu- ro sychopharmacol. & Biol. Psychiatry, 8:601406.

Zie re,, M., Lake, C., Fogpen,.F., Shoulson,,I., and Kopin, I. (1976) Iforepinephrine in cere rospinal fluid Brain Res., 108:436440.

145 :25 1-253.

Date post:	15-Nov-2023
Category:	Documents
Upload:	independent
View:	0 times
Download:	0 times

Clinical and biochemical aspects of depressive disorders: III. Treatment and controversies

Documents