AJP In Advance 1Reviews and Overviews [630431]

AJP In Advance 1Reviews and Overviews
ajp.psychiatryonline.orgThe Evolution of the Cognitive Model of Depression
and Its Neurobiological Correlates
Aaron T. Beck, M.D. Although the cognitive model of depres-
sion has evolved appreciably since its first
formulation over 40 years ago, the poten-
tial interaction of genetic, neurochemical,
and cognitive factors has only recently
been demonstrated. Combining findings
from behavioral genetics and cognitive
neuroscience with the accumulated re-
search on the cognitive model opens new
opportunities for integrated research.
Drawing on advances in cognitive, per-
sonality, and social psychology as well as
clinical observations, expansions of the
original cognitive model have incorpo-
rated in successive stages automatic
thoughts, cognitive distortions, dysfunc-
tional beliefs, and information-processing
biases. The developmental model identi-
fied early traumatic experiences and the
formation of dysfunctional beliefs as pre-
disposing events and congruent stressors
in later life as precipitating factors. It isnow possible to sketch out possible ge-
netic and neurochemical pathways that
interact with or are parallel to cognitivevariables. A hypersensitive amygdala is
associated with both a genetic polymor-
phism and a pattern of negative cognitivebiases and dysfunctional beliefs, all ofwhich constitute risk factors for depres-sion. Further, the combination of a hyper-
active amygdala and hypoactive prefron-
tal regions is associated with diminishedcognitive appraisal and the occurrence of
depression. Genetic polymorphisms alsoare involved in the overreaction to thestress and the hypercortisolemia in thedevelopment of depression—probablymediated by cognitive distortions. I sug-
gest that comprehensive study of the psy-
chological as well as biological correlatesof depression can provide a new under-standing of this debilitating disorder.
(Am J Psychiatry Beck; AiA:1–9)
I was privileged to start my research on depression at a
time when the modern era of systematic clinical and bio-
logical research was just getting underway. Consequently,
the field for new investigatio ns was wide open. The cli-
mate at the time was friendly for such research. The Na-tional Institute of Mental He alth had only recently been
funding research and providing salary support for full-
time clinical investigators. The Group for Advancement of
Psychiatry, under the leadership of pioneers like David
Hamburg, was providing guidelines as well as the impetus
for clinical research.
Caught up with the contagion of the times, I was
prompted to start something on my own. I was particularlyintrigued by the paradox of depression. This disorder ap-
peared to violate the time-honored canons of human na-
ture: the self-preservation inst inct, the maternal instinct,
the sexual instinct, and the plea sure principle. All of these
normal human yearnings were dulled or reversed. Even vi-
tal biological functions like eating or sleeping were attenu-
ated. The leading causal theory of depression at the time
was the notion of inverted host ility. This seemed a reason-
able, logical explanation if translated into a need to suffer.
The need to punish one’s self could account for the loss of
pleasure, loss of libido, self-criticism, and suicidal wishes
and would be triggered by guilt. I was drawn to conductingclinical research in depression because the field was wide
open—and besides, I had a testable hypothesis.
Cross-Sectional Model of Depression
I decided at first to make a foray into the “deepest” level:
the dreams of depressed patients. I expected to find signsof more hostility in the dream content of depressed pa-tients than nondepressed patients, but they actuallyshowed less hostility . I did observe, however, that the
dreams of depressed patients contained the themes ofloss, defeat, rejection, and abandonment, and thedreamer was represented as defective or diseased. At first Iassumed the idea that the negative themes in the dreamcontent expressed the need to punish one’s self (or “mas-ochism”), but I was soon disabused of this notion. Whenencouraged to express hostility, my patients becamemore, not less, depressed. Further, in experiments, theyreacted positively to success experiences and positive re-
inforcement when the “masochism” hypothesis predictedthe opposite (summarized in Beck [1]).
Some revealing observations helped to provide the basis
for the subsequent cognitive model of depression. I notedthat the dream content contained the same themes as thepatients’ conscious cognitions—their negative self-evalu-ations, expectancies, and memories—but in an exagger-AJP in Advance. Published July 15, 2008 (doi: 10.1176/appi.ajp.2008.08050721)
Copyright © 2008 American Psychiatric Association. All rights reserved.

2 AJP In AdvanceEVOLUTION OF THE COGNITIVE MODEL OF DEPRESSION
ajp.psychiatryonline.orgated, more dramatic form. The depressive cognitions con-
tained errors or distortions in the interpretations (ormisinterpretations) of experience. What finally clinchedthe new model (for me) was our research finding thatwhen the patients reappraised and corrected their misin-
terpretations, their depression started to lift and—in 10 or12 sessions—would remit (2).
Thus, I undertook the challenge of attempting to inte-
grate the different psychological pieces of the puzzle of de-pression. The end product was a comprehensive cognitivemodel of depression. At the surface, readily accessiblelevel was the negativity in the patients’ self-reports, in-
cluding their dreams and their negative interpretations of
their experiences. These variables seemed to account forthe manifestations of depression, such as hopelessness,loss of motivation, self-criticism, and suicidal wishes. Thenext level appeared to be a systematic cognitive bias in in-
formation processing leading to selective attention to neg-ative aspects of experiences, negative interpretations, andblocking of positive events and memories. These findingsraised the question: “What is producing the negativebias?” On the basis of clinical observations supported byresearch, I concluded that when depressed, patients hadhighly charged dysfunctional attitudes or beliefs about
themselves that hijacked the information processing andproduced the negative cognitive bias, which led to thesymptoms of depression (1, 3, 4).
A large number of studies have demonstrated that de-
pressed patients have dysfunctional attitudes, show a sys-tematic negative attentional and recall bias in laboratoryexperiments, and report cogn itive distortions (selective
abstraction, overgeneralizing, personalization, and inter-pretational biases [4]). Dysfunctional attitudes, measuredby the Dysfunctional Attitudes Scale (5), are representedby beliefs such as “If I fail at something, it means I’m a totalfailure.” During a full-blown episode of depression, the hy-persalient dysfunctional attitudes lead into absolute nega-tive beliefs about the self, their personal world, and the fu-ture (“I am a failure”). I suggested that these dysfunctionalattitudes are embedded within cognitive structures, orschemas, in the meaning assignment system and thushave structural qualities, such as stability and density aswell as thresholds and levels of activation. The degree ofsalience (or “energy”) of the schemas depends on the in-tensity of a negative experience and the threshold for acti-vation at a given time (successive stressful experiences, forexample, can lower the threshold [1]).
When the schemas are activated by an event or series of
events, they skew the information processing system,
which then directs attentional resources to negative stim-uli and translates a specific experience into a distortednegative interpretation. The hypersalience of these nega-tive schemas leads not only to a global negative percep-
tion of reality but also to the other symptoms of depres-sion, such as sadness, hopelessness, loss of motivation,and regressive behaviors such as social withdrawal and in-activity. These symptoms are also subjected to negative
evaluation (“My poor functioning is a burden on my fam-ily” and “My loss of motivation shows how lazy I am”).Thus, the depressive constellation consists of a continu-ous feedback loop with negative interpretations and at-
tentional biases with the subjective and behavioral symp-toms reinforcing each other.
Developmental Model of Depression
Cognitive Vulnerability
What developmental event or events might lead to the
formation of dysfunctional attitudes and how these events
might relate to later stressful events leading to the precip-itation of depression was another piece of the puzzle. Inour earlier studies, we found that severely depressed pa-
tients were more likely than moderately or mildly de-pressed patients to have experienced parental loss inchildhood (6). We speculated that such a loss would sensi-tize an individual to a significant loss at a later time in ad-olescence or adulthood, thus precipitating depression.Brij Sethi, a member of our group, showed that the combi-
nation of a loss in childhood with an analogous loss inadulthood led to depression in a significant number of de-pressed patients (7). The meaning of the early events (such
as “If I lose an important person, I am helpless”) is trans-formed into a durable attitude, which may be activated bya similar experience at a later time. A recent prospective
study observed that early life stress sensitizes individualsto later negative events through impact on cognitive vul-nerability leading to depression (8).
The accumulated research findings have supported the
original cognitive vulnerability model derived from clini-cal observations (4). As shown in Figure 1, early adverseevents foster negative attitudes and biases about the self,which are integrated into the cognitive organization in theform of schemas; the schemas become activated by lateradverse events impinging on the specific cognitive vulner-ability and lead to the systematic negative bias at the coreof depression (1, 9).
Much of the early research by others on the cognitive
model overlooked the role of stress in activating previ-ously latent dysfunctional schemas. Scher and colleagues(10) provided a comprehensive review of the diathesis-stress formulation based on prospective studies and prim-ing methodologies to test the cognitive model.
Although the original cognitive model proposed that se-
vere life events (e.g., death of a loved one or loss of a job)were the usual precipitants of depression (1), more recentresearch has suggested that milder stressful life eventsprovide an alternate pathway to depression in vulnerableindividuals (11, 12). Moreover, the triggering events of suc-cessive episodes of depression become progressivelymilder, suggesting a “kindling” effect (13, 14).
Cognitive vulnerability to the experience of depressive
symptoms following stress has been reported in children,

AJP In Advance 3AARON T. BECK
ajp.psychiatryonline.orgadolescents, and adults (10, 12, 15). For example, students
showing cognitive vulnerability were more likely to be-come depressed following negative outcomes on collegeapplications than were students not showing cognitivevulnerability (12, 16). It should be noted that these studies
generally described minor depressive episodes ratherthan full-blown major depression.
What relevance do dysfunctional attitudes have to the
vulnerability to recurrence of depression? Segal and col-
leagues (17) showed that the muted dysfunctional atti-tudes of recovered depressed patients could be primed by anegative mood induction procedure. Furthermore, the ex-tent to which the mood induction activated the dysfunc-
tional attitudes during the nondepressed period predictedfuture relapse and recurrence. This activation of the dys-functional attitudes was more likely to occur with patientswho had received pharmacotherapy than those receivingcognitive therapy. The prospective and priming studiesthus indicated that dysfunctional attitudes could be re-garded as a cognitive vulnerability factor for depression.
A more recent refinement of the cognitive vulnerability
model has added the concept of cognitive reactivity , ex-
pressed clinically as fluctuations in patients’ negative atti-tudes about themselves in response to daily events (18).Cognitive reactivity has been demonstrated experimen-tally by a variety of priming interventions (or “mood inter-ventions”) such as sad music, imaging of sad autobio-graphical memories, social rejection film clip, or contrivedfailure. Following these priming interventions, clinicallyvulnerable subjects report more dysfunctional attitudes,negative cognitive biases, and erosion of normal positivebiases than do other subjects (10). Clinical vulnerabilitywas defined in terms of high-risk variables (e.g., a personalor family history of depression).
Studies have shown that the presence of cognitive reac-
tivity before a stressful life event predicts the onset of de-pressive episodes (10). The importance of the meaning as-signed to a stressful event as a crucial component ofcognitive reactivity was borne out by the finding (19) thatthe daily negative appraisals of daily stressors predicteddaily depressive symptoms. The addition of the concept ofcognitive reactivity to the cognitive model suggested thatthe predisposition to depression may be observable in thedaily cognitive-emotional reactions of the depression-
prone individual (18). However, the model left unan-swered why certain individuals were more reactive to dailyevents (or were more likely to develop dysfunctional atti-tudes and cognitive biases) than others.
As indicated previously, the experience of episodes of
depressive symptoms is different from the total immersion
of the personality in a full-blown major depression. Severedepression is characterized not only by a broad range of in-tense symptoms but also “endogenous” features, such asrelative insensitivity to external events. To account for thecomplex characteristics of the fully expressed depression, Iproposed an expanded cognitive model (4, 20). I presentedthe concept of the mode , a network of cognitive, affective,
motivational, behavioral, and physiological schemas, to
account for the profound retardation, anhedonia, andsleep and appetite disturbance, as well as the cognitive ab-errations. The activation of this mode (network) produces
the various phenomena of depression.
In the formation of the mode, the connections among
the various negatively oriented schemas become strength-ened over time in response to negatively interpreted
events. Successive symptom-producing events or a majordepressogenic event locks these connections into place. Ina sense, the cognitive schemas serve as the hub and theother schemas as nodes with continuous communication
among them. A major stressful event or events symbolizinga loss of some type trigger the cognitive schemas that acti-vate the other (affective, motivational, etc.) schemas. Whenfully activated, the mode becomes relatively autonomousand is no longer as reactive to external stimuli; that is, pos-itive events do not reduce the negative thinking or mood.Attentional resources are di sproportionate ly allocated
from the external environment to internal experiencessuch as negative cognitions and sadness, manifested clini-cally as rumination. Also, resources are withdrawn fromadaptive schemas such as coping and problem solving.The mode presumably would correspond to a complexneural network, including multiple relevant brain regionsthat are activated or deactivated during depression.
The negatively biased cognitive schemas function as
automatic information processors. The biased automatic
processing is rapid, involuntary, and sparing of resources.The dominance of this system (efficient but maladaptive)in depression could account for the negative attentionaland interpretational bias. In contrast, the role of the cogni-tive control system (consisting of executive functions,problem solving, and reappraisal) is attenuated during de-pression. The operation of this system is deliberate, reflec-tive, and effortful (resource demanding), can be reacti-vated in therapy, and, thus, can be used to appraise thedepressive misinterpretations and dampen the salience ofthe depressive mode.
The concept of the two forms of processing can be
traced back to Freud’s model of primary and secondaryprocesses (21) and has been reformulated many timessince (22). Recently, Beevers (23) suggested a similar for-mulation of a two-factor processing in depression. He pro-posed that cognitive vulnerability to depression occurswhen negatively biased associative processing is uncor-rected by reflective processing.
The expanded cognitive model includes the following
progression in the development of depression: adverse
early life experiences contribu te to the formation of dys-
functional attitudes incorporated within cognitive struc-tures, labeled cognitive schemas (cognitive vulnerability).When activated by daily life events, the schemas producean attentional bias, negatively biased interpretations, andmild depressive symptoms (cognitive reactivity).

4 AJP In AdvanceEVOLUTION OF THE COGNITIVE MODEL OF DEPRESSION
ajp.psychiatryonline.orgAfter repeated activation, the negative schemas become
organized into a depressive mode, which also includes af-
fective, behavioral, and motivational schemas (cognitivevulnerability). Accumulated ne gative events or a severe
adverse event impacts on the mode and makes it hypersa-
lient. The hypersalient mode takes control of the informa-tion processing, reflected by increased negative appraisalsand rumination. The cognitive control of emotionally sig-nificant appraisals is attenuated and, thus, reappraisal ofnegative interpretations is limited. The culmination ofthese processes is clinical depression. Crick and Dodge(24) point out that with repeated activation, maladaptiveinformation-processing patterns become routinized andresistant to change. Thus, cognitive schemas, after re-
peated activation before and during depressive episodes,become more salient and more ingrained over time, con-sistent with the “kindling” phenomenon.
Although supportive evidence for this model was mea-
ger in early years, support for the fundamental hypotheseshas accumulated over the past 40 years. In 1999, David A.Clark and I (4) reviewed over 1,000 publications relevant tothe cognitive model of depression and found substantialresearch support for the various facets of the cognitivetheory. Several more recent reviews have provided addi-tional research support (10, 12, 15, 25).
As the cognitive model of depression buttressed by
years of systematic research has grown to maturity, itseems timely and appropriate to compare it with the bur-geoning findings in neurogenetics and neuroimaging.
Biological Correlates of the Cognitive
Model
Genetic Vulnerability
I still had an unsolved problem: how can we account for
the observation that only a proportion of individuals sub-
jected to child abuse, other adverse events, and majortraumatic experiences become depressed? Many of us hadspeculated about the existence of a “blue gene,” but the
technology for identifying it had not been available. Also,although there was considerable support for the cognitivemodel of depression at the psychological and clinical lev-els (4), there were minimal data from neurophysiologicalstudies to correlate with these findings. It was not until
this century that these problems could be addressed, as aresult of investigations by researchers in behavioral genet-ics and cognitive neuroscience. The spectacular techno-logical advances in genetics and functional neuroimaginghave enabled researchers to demonstrate that geneticvariations and their impact on neural functioning play amajor role in the hyperreactivi ty to negative experiences
leading to depression. A number of studies have provideda structure for understanding the relationships betweenlife events, neural dysregulation, cognitive processes, anddepression. This research has also provided a preliminarybasis for formulating the neurobiological correlates ofsuch psychological constructs as cognitive vulnerability,cognitive reactivity, and cognitive biases.
These advances have facilita ted a breakthrough in un-
derstanding the relationship between cognitive, biologi-cal, and experiential factors in the development of depres-sion. The genetic and neurobiological findingsilluminated some probable causal pathways to depressionas well as suggesting biological correlates of the cognitivemodel. The pioneering paper by Caspi and colleagues (26)suggested that individuals possessing either one or twocopies of the short variant of the 5-HTTLPR (serotonintransporter) gene, which is not transcriptionally as effec-tive as the long form, experienced higher levels of depres-sion and suicidality following a recent life stressor. Thestudy by Caspi and colleagues (26) has since been sup-ported by a large number of other studies (27).
The types of stressful life events moderated by the 5-
HTTLPR gene varied considerably in these studies, rang-ing from mild stressors to a single, large traumatic event.Also, in some studies, adverse experiences, for example,abuse in childhood (28), appear to represent a distal pre-disposition to depression, whereas in others, 5-HTTLPRgenotype moderates the depressogenic effects of more
proximal events (11). Studies of both predispositional (bi-
ological and psychological) and precipitating eventswould provide a test of this aspect of the cognitive model.Kilpatrick and colleagues (29) found that carriers of thelow-expression variant of the short-form 5-HTTLPR poly-morphism were prone to develop major depression andpostdisaster posttraumatic stress disorder under condi-tions of high exposure to hurricanes and low social sup-
port. Kaufman and colleagues (28) have also found evi-
dence that social support buffers against depressivereactions to stressful experiences among genetically vul-nerable individuals. Although questions have been raisedregarding the generalizability of the 5-HTTLPR findings(e.g., 30), they are useful to illustrate parallels between bi-ological and psychological concepts. The more exhaustive
analyses would include other variants such as the CREB
and COMT genes (31, 32).
Investigators have also found that the brain-derived
neurotrophic factor genotype interacted with the 5-HTTLPR gene to predict depression in children (28) andFIGURE 1. A Developmental Model of Depression Based on
Vulnerability Diathesis and Stressful Life Events
Activation by
stressful eventsDysfunctional attitudes
(schemas):
cognitive vulnerability
Pervasive negative
cognitive bias:
depressionAdverse developmental
experiences

AJP In Advance 5AARON T. BECK
ajp.psychiatryonline.orgolder adults (33). Of interest, variants of the brain-derived
neurotrophic factor predicted ruminations and depres-
sion differently in adolescent girls and their mothers (34).
Specifically, Hilt and colleagues (34) found that girls withthe Val/Val genotype had higher rumination scores and
exhibited more symptoms of depression than girls withthe Val/Met genotype. In contrast, mothers with adult -on-
set depression and the Val/Met genotype exhibited more
symptoms of depression and rumination. Of interest,mothers with childhood -onset depression were more
likely to have the Val/Val genotype. Kaufman and col-
leagues (28) found that in maltreated children, depressiveseverity was predicted in part by an interaction of the 5-
HTTLPR (short allele) with the brain-derived neu-
rotrophic factor ( Val/Met ) genotype, particularly among
children with low social su pport. A variant of the HTR-2A
gene has been found to potentiate the effect of maternalnurturance in mitigating th e experience of depressive
symptoms in children (35). As research proceeds in thisarea, it seems likely that a variety of other gene-gene andgene-environment interactions will be discovered. Ingeneral, the established relationships between environ-mental events, biological pr edisposition, and depression
appear to run parallel to the findings of the cognitive
model regarding environmental events, cognitive factors,
and depression.
Genetic Diathesis and Cognitive Bias
While the genetic studies have pointed to the innate bi-
ological vulnerability to stress leading to depression, the
relation of biological to cognitive vulnerability needed to
be clarified. The gene-by-environment findings for de-
pression have prompted an interest in uncovering their re-
lationships to cognitive variables. A variety of experimen-tal procedures have been used to test for cognitive bias inindividuals at genetic risk for depression. The negative at-tentional, recall, and interpretative biases are generally
elicited by mood induction procedures, such as viewing
sad movie clips or imagining sad experiences (36). Geneticantecedents for these observed cognitive biases have been
identified. A number of studies indicate that negative cog-nitive processing and negative cognitions are associated
with the presence of the 5-HTTLPR short allele (37–40). Of
particular relevance to possi ble cognitive predisposition,
Hayden and colleagues (40) found that nondepressed chil-dren homozygous for the short allele showed greater neg-ative processing on a self-referential encoding task follow-ing a negative mood induction than did children with
other genotypes. Thus, the accumulating evidence sug-
gests the genetic predisposition to depression is associ-ated with biases in the processing of information.
Neurophysiological and Cognitive Factors/Bias
What neurophysiological processes are related to the
cognitive biases? Multiple findings have tied amygdala hy-
peractivity to depression (41, 42). However, the findingsneed to be considered within a broader framework, in-
cluding many brain regions implicated in depression (43,44). A specific line of inquiry has tied the 5-HTTLPR vari-ant to activation in brain regions critical for processingnegative stimuli. Hyperreactivity of the amygdala in theshort 5-HTTLPR variant carriers is associated with in-
creased sensitivity to negative stimuli (45) and leads tonegative bias in the processing or interpretation of emo-tional stimuli (46, 47). Since the amygdala is involved inthe evaluation and storage of emotionally charged events(48), its hyperreactivity to negative stimuli in predisposedindividuals would appear to represent a neurophysiologi-cal correlate of cognitive bias.
The systematic bias in information processing in de-
pression is reflected not only in selective attention and ex-aggerated reaction to negative stimuli (49, 50) but also inthe expectancy of aversive events (51). Abler and col-
leagues (51) found that the anticipation of noxious stimuliproduced excessive amygdala activation in geneticallyprone individuals. Further, there is evidence that the 5-HTTLPR gene interacted with children ’ s attentional andinferential biases to predict depressive symptoms; infer-ential bias alone predicts lifetime diagnosis of depressionamong carriers of the 5-HTTLPR short allele, but notamong those homozygous for the long allele (unpublishedwork by Gibb BE, Benas JS, Grassia M, McGeary J andunpublished work by Gibb BE, Uhrlass DJ, Grassia M,McGeary J).
The pathway from genetic and cognitive predisposition
to depression may be clarified by studies of the impact ofstress on neural functioning. Gotlib and colleagues (52)have found that carriers of the short-form serotonin trans-porter gene (5-HTTLPR) show elevated cortisol response,cognitive biases, and activation of the amygdala during amood repair procedure. Adverse circumstances engagethe hypothalamic-pituitary-adrenal (HPA) axis, whichleads to the secretion of excessive “stress hormones” suchas cortisol (52). Presumably, the continual secretion ofcortisol culminates in the hypercortisolemia characteris-tic of most depressed individuals (1, 53).
Several converging findings suggest that the cognitive
appraisal of a stressor plays a role in the evocation of corti-
sol response and the generation of depressive symptoms.In a review of relevant literature, Dickerson and Kemeny(54), for example, noted consistent findings that experi-mental manipulations appraised as threat of social rejec-tion produced an elevated co rtisol response. These results,
combined with the findings by Hankin and colleagues (19)
on the negative cognitive responses to daily stressfulevents leading to depressive symptoms, suggest a pathwayto depressive symptoms: stress → distorted appraisal → en-
gagement of the HPA axis → cortisol → depressive symp-
toms. Of course, there are un doubtedly feedback loops in-
volving both psychological and biological variables.
A further elaboration of this hypothesis is suggested by
Gotlib (unpublished work by Gotlib IH), who proposed a

6 AJP In AdvanceEVOLUTION OF THE COGNITIVE MODEL OF DEPRESSION
ajp.psychiatryonline.orgreciprocal model involving the dysregulated HPA axis (in
response to specific stressors) leading to increased cortisol
secretion, which affects the serotonergic system. He also
proposes a more complex theory that involves the impactof increased cortisol secretion on the short form of the 5-
HTTLPR gene, leading to an alteration in the transmission
of serotonin and, consequently, negative feedback to the
HPA axis and increased cortisol secretion.
Gotlib’ s theory can be amplified to take into account re-
search findings demonstrating interactions between mea-sures of cognitive reactivity and serotonin. Studies have
shown that increases or decreases in serotonin activity are
related to self-assessment of dysfunctional attitudes or
cognitive reactivity. Meyer and colleagues (55) reported
that acute tryptophan depletion of serotonin increases
dysfunctional attitudes, while Booij and colleagues (56)
found that depletion of serotonin increased cognitive re-activity. The finding that individuals with the short variant
of the 5-HTTLPR gene show depressive symptoms after
experimental depletion of serotonin provides evidence ofthe association of genomic with neurochemical vulnera-bility. Finally, Meyer and colleagues (57) reported in-
creased dysfunctional attitudes in depressed patients withlow intracellular serotonin. Thus, the preliminary evi-
dence suggests linkages between cognitive vulnerabilityand genetic vulnerability expressed as a hyperreactive se-rotonergic system.
Recent neurophysiological research is pertinent to an-
other aspect of the expanded cognitive model, mainly theformulation that during depression of the cognitive con-trol system, top-down processing is dysregulated whilethe bottom-up schematic processing is prepotent. Siegleand colleagues (41) found that nearly all depressed pa-tients have reduced prefrontal function and about one-half have increased amygdala activity. Thus, the balance oftheir respective activity is relevant to cognitive control.Banks and colleagues (58) found that the reducedamygdala coupling with the orbitofrontal cortex and thedorsal medial prefrontal cortex predicts the extent of at-
tenuation of negative affect following reappraisal.Johnstone and colleagues (59) point out that a key featureunderlying the pathophysiology of major depression is thedysfunctional engagement of the right prefrontal cortexand the lack of engagement of the left lateral ventral me-
dial prefrontal circuitry, important for the down-regula-tion of amygdala responses to negative stimuli. They sug-gest that the top-down process of reappraisal is defectivein depressed individuals; this may account for the impor-tance of reappraisal in the cognitive therapy of this disor-der. Thus, two concurrent processes are involved in emo-tional processing in depression: diminished cognitivecontrol from prefrontal and cingulate regions and in-creased activity in the amygdala and other regions.
Deconstructing Depression
Interpretation of the rese arch comparing components
of the cognitive model with the neurophysiological inves-
tigations of depression poses a philosophical problem.How can one reconcile two totally different levels of ab-straction: mentalism and materialism? The cognitive and
neurophysiological approaches use different concepts, re-search strategies, and technical procedures. Given thisphilosophical problem, is there justification for mixing thetwo models in terms of causation or interaction (for exam-ple, reduction of serotonin causes an increase in dysfunc-tional attitudes; see reference 57) or are the neurophysio-logical and cognitive processes simply “different sides ofthe same coin,” as I once argued (60)? According to myearlier notion, the cognitive processes are parallel to but
do not interact with the biological processes.
Notwithstanding this philosophical problem, I believe
that it is possible to present a pragmatic formulation of the
interaction of the two levels (Figure 2). In deconstructing
the phenomenon of depression, I propose a hypotheticalpathway starting with genetic vulnerability (including pre-dispositional but not protective genes). The 5-HTTLPRFIGURE 2. A Developmental Model of Depression Based on
Anomalous Genesa
aMultiple interactions are not shown. Genetic pathways leading toreduced prefrontal activity have not been determined as yet. In-creased limbic activity overrides prefrontal control.Reactive amygdala
Cognitive biases
Exaggeration of
stressful events
Activation of hypothalamic-
pituitary-adrenal axis
Dominance of limbic
activity over
prefrontal function
Deficient reappraisal
of negative cognitions
Depressive symptomsGenetic diathesis

AJP In Advance 7AARON T. BECK
ajp.psychiatryonline.orgpolymorphism leads to excessive reactivity of the
amygdala (45). The heightened limbic reactivity to emo-
tionally significant events triggers deployment of increased
attentional resources to such events, manifested by nega-tive attentional bias and recall (cognitive reactivity). Theselective focus on the negative aspects of experience re-sults in the familiar cognitive distortions such as exaggera-
tion, personalization, and overgeneralization (6) and, con-
sequently, in the formation of dysfunctional attitudesregarding personal adequacy, acceptability, and worth.Frequent reiterations of negative interpretations shape thecontent of the cognitive schemas (unlovable, inadequate,worthless). Concurrently, the negative interpretations of
experience have an impact on the HPA axis and set in mo-
tion the previously described cycle involving the overreac-tive serotonergic system and consequently lead to depres-sion. Care needs to be taken in the interpretation of gene-environment studies, including those of the 5-HTTLPRvariant, given the numerous possible methodological pit-
falls in these sorts of analyses (Kenneth Kendler, personal
communication, May 11, 2008). Consequently, this partic-ular formulation is tentative, subject to future research.
The theory of the consolidation of the negative attitudes
and the core negative self-concept along with the associ-
ated affective, motivational, and behavioral factors into
the depressive mode is speculative but is useful as an ex-planatory construct (20). The progression from depressiveproneness to a full-blown depression would involve not
only the hyperactivation of this conglomerate but also the
diminution of reality testing. The biological counterpart ofthis theoretical model includes complex circuits involvingmultiple brain regions. Mayberg (61), for example, definesa major depressive episode as a “pattern of dysfunctional
interactions among specific cingulate, paralimbic, sub-
cortical, and frontal regions critical to maintaining emo-tional homeostasis under conditions of exogenous or en-dogenous stress” (p. 258). At present the most directparallel involves the predominance of negatively biasedprocessing and reduced reality testing on the one hand
and amygdala activation and disengagement of executive
(especially prefrontal) regions on the other.
Future Perspectives
The accumulation of studies of the psychological and
biological aspects of depression has reached a critical
mass warranting a new synthesis. The findings of relation-ships among the diverse genetic, neurophysiological, en-vironmental, and cognitive aspects of the vulnerability to
and development of depression call for future studies inte-
grating these findings into a comprehensible formulation.A series of multiple wave prospective studies could iden-tify the relevant variables and their interrelationships. Thestudies ideally would be complementary to each other sothat each finding would contribute to the overall formula-
tion of the theory of depres sion. The clarification of spe-cific associations among the relevant variables should
yield valuable information. In order to simplify the sug-gested plan, I have limited the genetic diathesis to the 5-HTTLPR gene. Obviously, other genes, including thosethat have a protective effect, as well as the individuals’ so-cial support should be included in any formulations.Moreover, as indicated previously, the research based onthis gene is not totally reliable.
The proposed psychobiolog ical model of depression
poses a number of problems and questions that need to beaddressed. The developmental model presupposes thatthe 5-HTTLPR gene in some as yet unidentified way leadsto hyperreactivity of the amygdala to external stimuli. Thisactivity is associated with a negative cognitive bias. In addi-tion, increased reactivity of the HPA axis is also associatedwith this genetic variant (52). A longitudinal study startingin childhood could investigate the causal sequence. In onepathway, the hyperactive amygdala leads initially to atten-tional bias that gradually progresses into significant nega-tive cognitive distortions of daily experiences that engagethe HPA axis. Alternatively, the amygdala and HPA overac-tivity may contribute independently to the cognitive andphysiological impact of life events.
A number of other problems warrant further investiga-
tion. The association of increased dysfunctional attitudeswith serotonin depletion poses an interesting question: isthe association due to a common linkage to some unde-fined third factor, such as amygdala reactivity? Does sero-tonin deficiency cause an increase in dysfunctional
attitudes (55), is the converse true, or is there another ex-planation for this association? The finding that depressedindividuals treated with cognitive therapy, compared topharmacotherapy treated patients, do not show an in-
crease in depressive symptoms (generally associated withdysfunctional attitudes) following tryptophan depletioncalls for further investigation (62). These results indicate amore complex relationship between serotonin depletionand dysfunctional attitudes. The neurobiological mecha-nisms involved in the reduced reality testing of negativebeliefs deserve further attention. The specific brain areasinvolved in dysconnectivity need to be spelled out. Is thepreemption of amygdala activity over prefrontal and cin-
gular executive functions related to reduction of serotonininhibition of the amygdala (63)? Work relating poor errordetection in depression to specific dysfunctions in variousbrain regions (64) needs to be followed.
The integrative developmental model postulates that
various genetic variants sensitize individuals to life experi-
ences that make them vulnerable to depression. Specifi-cally, the studies would address the biological mecha-nisms that contribute to depression through the tendencyto construe events in an excessively negative way. A seriesof waves, starting in early childhood, should examine thevariables associated with the polymorphism: assessmentsof information processing biases (36, 37, 40), negative cog-nitions (38), and dysfunctional attitudes (65) following

8 AJP In AdvanceEVOLUTION OF THE COGNITIVE MODEL OF DEPRESSION
ajp.psychiatryonline.orgnegative mood inductions. These findings would then be
compared with studies of brain activity to determine theirassociations with the limbic system as well as prefrontalcingular and other regions. The early studies would deter-mine whether automatic cognitive processing precedes
the development of negative cognitions and dysfunctionalattitudes. Another wave could examine diary records ofdaily dysfunctional cognitions in response to stressful sit-uations and relate these to cortisol responses to specificstimuli situations. Overall, these assessments over a longtime span would integrate findings from neuroimagingand neuroendocrine responses to stressors with cognitiveresponses to daily stressful activities as well as to major lifeevents. The association of in creased dysfunctional atti-
tudes with serotonin depletion would also be determined
in these studies, and in addition the attempt to find thespecific relationships among these variables could be de-termined.
I have reason to hope that future research will perhaps
provide a new paradigm which for the first time can inte-grate findings from psychological and biological studies tobuild a new understanding of depression.
Received May 15, 2008; accepted May 20, 2008 (doi: 10.1176/
appi.ajp.2008.08050721). From the Department of Psychiatry, Uni-
versity of Pennsylvania. Address correspondence and reprint re-quests to Dr. Beck, Department of Psychiatry, University of Pennsyl-vania, 3535 Market St., Rm. 2032, Philadelphia, PA 19104;abeck@mail.med.upenn.edu (e-mail).
Dr. Beck reports no competing interests.
The author thanks Brad Alford, Andrew Butler, Brandon Gibb, Ken-
neth Kendler, Helen Mayberg, John O’Reardon, and Greg Siegle for
criticism, comments, and suggestions and Brianna Mann and Allison
Fox for technical support.
References
1. Beck AT: Depression: Clinical, Experimental, and Theoretical
Aspects. New York Harper & Row, 1967.
2. Rush AJ, Beck AT, Kovacs M, Hollon SD: Comparative efficacy of
cognitive therapy and pharmacotherapy in the treatment of
depressed outpatients. Cognit Ther Res 1977; 1:7–37
3. Beck AT: Cognitive models of depression. J Cogn Psychother:
An Int Quarterly 1987; 1:5–37
4. Clark DA, Beck AT: Scientific Foundations of Cognitive Theory
and Therapy of Depression. New York, John Wiley & Sons, 1999
5. Beck AT, Steer RA, Brown GK, Weissman A: Factor analysis of
the Dsyfunctional Attitude Scale in a clinical population. Psy-
chological Assessment 1991; 3:478–483
6. Beck AT, Sethi B, Tuthill R: Childhood bereavement and adult
depression. Arch Gen Psychiatry 1963; 9:295–302
7. Sethi BB: The relationship of separation to depression. Arch
Gen Psychiatry 1964; 10:486–496
8. Harkness KL, Lumley MN: Child abuse and neglect and the de-
velopment of depression in children and adolescents, in Hand-
book of Depression in Children and Adolescents. Edited by
Abela JR, Hankin BL. New York, Guilford, 2008, pp 466–488
9. Beck AT: Cognitive Therapy and the Emotional Disorders. New
York, Meridian, 1976
10. Scher C, Ingram R, Segal Z: Cognitive reactivity and vulnerabil-
ity: empirical evaluation of construct activation and cognitive
diatheses in unipolar depression. Clin Psychol Rev 2005; 25:
487–51011. Kendler KS, Kuhn JW, Vittum J, Prescott CA, Riley B: The interac-
tion of stressful life events and a serotonin transporter poly-
morphism in the prediction of episodes of major depression: a
replication. Arch Gen Psychiatry 2005; 62:529–535
12. Abela JRZ, Hankin BL: Cognitive vulnerability to depression in
children and adolescents: a developmental psychopathologyperspective, in Handbook of Depression in Children and Ado-lescents. Edited by Abela JRZ, Hankin BL. New York, Guilford,
2008, pp 35–78
13. Kendler KS, Thornton LM, Gardner CO: Stressful life events and
previous episodes in the etiology of major depression inwomen: an evaluation of the “kindling” hypothesis. Am J Psy-
chiatry 2000; 157:1243–1251
14. Monroe SM, Harkness KL: Life stress, the “kindling” hypothesis,
and the recurrence of depression: considerations from a lifestress perspective. Psyc hol Rev 2005; 112:417–445
15. Jacobs RH, Reinecke MA, Gollan JK, Kane P: Empirical evidence
of cognitive vulnerability for depression among children and
adolescents: a cognitive science and developmental perspec-tive. Clin Psychol Rev 2008; 28:759–728
16. Abela JRZ, D’Alessandro DU: Beck’s cognitive theory of depres-
sion: a test of the diathesis-stress and causal mediation com-ponents. Br J Clin Psychol 2002; 41:111–128
17. Segal ZV, Gemar M, Williams S: Differential cognitive response
to a mood challenge following successful cognitive therapy orpharmacotherapy for unipolar depression. J Abnorm Psychol
1999; 108:3–10
18. Butler AC, Hokanson JE, Flynn HA: A comparison of self-esteem
lability and low trait self-esteem as vulnerability factors for de-pression. J Pers Soc Psychol 1994; 66:166–177
19. Hankin BL, Fraley C, Abela JRZ: Daily depression and cognitions
about stress: evidence for a trait-like depressogenic cognitive
style and the prediction of depressive symptoms in a prospec-
tive daily diary study. J Pers Soc Psychol 2005; 88:673–685
20. Beck AT: Beyond belief: a theory of modes, personality, and
psychopathology, in Frontiers of Cognitive Therapy. Edited by
Salkovskis P . New York, Guilford, 1996, pp 1–25
21. Freud S: Collected Works of Sigmund Freud. New York and
Washington, DC, BiblioBazaar, originally published in 1920
22. Chaiken S, Trope Y: Dual-Process Theories in Social Psychology.
New York, Guilford, 1999
23. Beevers CG: Cognitive vulnerability to depression: a dual pro-
cess model. Clin Psychol Rev 2005; 25:975–1002
24. Crick NR, Dodge KA: A review and reformulation of social infor-
mation-processing mechanisms in children’s social adjust-ment. Psychol Bull 1994; 115:74–101
25. Dozois DJA, Beck AT: Cognitive schemas, beliefs and assump-
tions, in Risk Factors for Depression. Edited by Dobson KS,
Dozois DJA. Oxford, UK, Elsevier (in press)
26. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington HL:
Influence of life stress on depression: moderation by a poly-
morphism in the 5-HTT gene. Science 2003; 301:386–389
27. Uher R, McGuffin P: The moderation by the serotonin trans-
porter gene of environmental adversity in the aetiology of
mental illness: review and methodological analysis. Molecular
Psychiatry 2008; 13:131–146
28. Kaufman J, Yang B, Douglas-Palumberi H, Grasso D, Lipschitz D,
Houshyar S: Brain-derived neurotrophic factor 5-HTTLPR geneinteractions and environmental modifiers of depression inchildren. Biol Psychiatry 2006; 59:673–680
29. Kilpatrick DG, Koenen KC, Ruggiero KJ, Acierno R, Galea S,
Resnick HS: The serotonin tran sporter genotype and social
support and moderation of posttraumatic stress disorder and
depression in hurricane-exposed adults. Am J Psychiatry 2007;
164:1693–1699
30. Gillespie NA, Whitfield JB, Williams B, Heath AC, Martin NG: The
relationship between stressful life events, the serotonin trans-

AJP In Advance 9AARON T. BECK
ajp.psychiatryonline.orgporter (5-HTTLPR) genotype and major depression. Psychol
Med 2005; 35:101–111
31. Zubenko GS, Hughes HB III: Effects of the G(-656), a variant on
CREB1 promoter activity in a neuronal cell line: interactions
with gonadal steroids and stress. Mol Psychiatry 2008 (Epub
ahead of print)
32. Wichers M, Aguilera M, Kenis G, Krabbendam L, Myin-Germeys
I, Jacobs N, Peeters F, Derom C, Vlietinck R, Mengelers R, Deles-
paul P , van Os J: The catechol-O-methyl transferase Val(158)Met
polymorphism and experience of reward in the flow of dailylife. Neuropsychopharmacology 2007 (Epub ahead of print)
33. Kim J, Stewart R, Kim S, Yang S, Shin I, Kim Y: Interactions be-
tween life stressors and susceptibility genes (5-HTTPR andBDNF) on depression in Korean elders. Biol Psychiatry 2007;
62:423–428
34. Hilt LM, Sander LC, Nolen-Hoeksema S, Simen AA: The BDNF
Val66Met polymorphism predicts rumination and depressiondifferently in young adolescent girls and their mothers. Neuro-
sci Lett 2007; 429:12–16
35. Jokela M, Keltikangas-Järvinen L, Kivimäki M, Puttonen S,
Elovainio M, Rontu R: Serotonin receptor 2A gene and the in-fluence of childhood maternal nurturance on adulthood de-
pressive symptoms. Arch Gen Psychiatry 2007; 64:356–360
36. Joormann J, Talbot L, Gotlib IH: Biased processing of emotional
information in girls at risk for depression. J Abnorm Psychol
2007; 116:135–143
37. Beevers CG, Gibb BE, McGeary JE, Miller IW: Serotonin trans-
porter genetic variation and biased attention for emotionalword stimuli among psychiatric inpatients. J Abnorm Psychol
2007; 116:208–212
38. Beevers CG, Scott WD, McGeary C, McGeary JE: Negative cogni-
tive response to a sad mood induction: associations with poly-morphisms of the serotonin transporter (5-HTTLPR) gene. Cog-nition and Emotion (in press)
39. Canli T, Lesch K: Long story short: the serotonin transporter in
emotion regulation and social cognition. Nat Neurosci 2007;10:1103–1109
40. Hayden EP , Dougherty LR, Maloney B, Olino TM, Durbin CE,
Sheihk HI: Early-emerging cognitive vulnerability to depressionand the
serotonin transporter promoter region polymorphism.
J Affect Disord 2008; 107:227–230
41. Siegle GJ, Thompson W, Carter CS, Steinhauer SR, Thase ME: In-
creased amygdala and decreased dorsolateral prefrontal bold
responses in unipolar depression: related and independent
features. Biol Psychiatry 2007; 61:198–209
42. Surguladze S, Brammer MJ, Keedwell P , Giampietro V, Young
AW, Travis MJ: A differential pattern of neural response toward
sad versus happy facial expressions in major depressive disor-der. Biol Psychiatry 2005; 57:201–209
43. Mayberg HS: Modulating dysfunctional limbic-cortical circuits
in depression: towards development of brain-based algo-rithms for diagnosis and optimised treatment. Br Med Bull
2003; 65:193–207
44. Phillips ML, Drevets WC, Rauch SL, Lane R: Neurobiology of
emotion perception II: implications for major psychiatric disor-ders. Biol Psychiatry 2003; 54:515–528
45. Munafò M, Brown S, Hariri A: Serotonin transporter (5-HTTLPR)
genotype and amygdala activation: a meta-analysis. Biol Psy-chiatry 2008; 63:852–857
46. Dannlowski U, Ohrmann P , Bauer J, Kugel H, Arolt V, Heindel W:
Amygdala reactivity to masked negative faces is associatedwith automatic judgmental bias in major depression: a 3T fMRI
study. J Psychiatry Neurosci 2007; 32:423–429
47. Monk CS, Klein RG, Telzer EH, Schroth EA, Mannuzza S, Moul-
ton JL III, Guardino M, Masten CL, McClure-Tone EB, Fromm S,Blair RJ, Pine DS, Ernst M: Amygdala and nucleus accumbensactivation to emotional facial expressions in children and ado-
lescents at risk for major depression. Am J Psychiatry 2008;
165:90–98
48. LeDoux JE: The Emotional Brain: The Mysterious Underpin-
nings of Emotional Life. New York, Simon & Schuster, 1996
49. Siegle GJ, Steinhauer SR, Thase ME, Stenger VA, Carter CS: Can’t
shake that feeling: fMRI assessment of sustained amygdala ac-tivity in response to emotional information in depressed indi-
viduals. Biol Psychiatry 2002; 51:693–707
50. Siegle GJ, Carter CS, Thase ME: Use of fMRI to predict recovery
from unipolar depression with cognitive behavior therapy. Am
J Psychiatry 2006; 163:735–738
5 1 . A b l e r B , E r k S , H e r w i g U , W a l t e r H : A n t i c i p a t i o n o f a v e r s i v e
stimuli activates extended amygdala in unipolar depression. JPsychiatr Res 2007; 41:511–522
52. Gotlib IH, Joormann J, Minor K, Hallmayer J: HPA axis reactivity:
A mechanism underlying the associations among 5-HTTLPR,
stress, and depression. Biol Psychiatry 2008; 63:847–851
53. Parker KJ, Schatzberg AF, Lyons DM: Neuroendocrine aspects of
hypercortisolism in major depression. Hormones & Behavior
2003; 43:60–66
54. Dickerson SS, Kemeny ME: Acute stressors and cortisol re-
sponses: a theoretical integration and synthesis of laboratory
research. Psychol Bull 2004; 130:355–391
55. Meyer JH, McMain S, Kennedy SH, Korman L, Brown GM, DaSilva
JN, Wilson AA, Blak T, Eynan-Harvey R, Goulding VS, Houle S,
Links P: Dysfunctional attitudes and 5-HT
2 receptors during de-
pression and self-harm. Am J Psychiatry 2003; 160:90–99
56. Booij L, Van der Does AJW: Cognitive and serotonergic vulnera-
bility to depression: convergent findings. J Abnorm Psychol
2007; 116:86–94
57. Meyer JH, Houle S, Sagrati S, Carella A, Hussey DF, Ginovart N:
Brain serotonin transporter binding potential measured with
carbon 11–labeled DASB positron emission tomography: ef-fects of major depressive episodes and severity of dysfunc-
tional attitudes. Arch Gen Psychiatry 2004; 61:1271–1279
58. Banks SJ, Eddy KT, Angstadt M, Nathan PJ, Phan KL: Amygdala-
frontal connectivity during emotion regulation. Soc Cogn
Affect Neurosc 2007; 2:303–312
59. Johnstone T, van Reekum CM, Urry HL, Kalin NH, Davidson RJ:
Failure to regulate: counterproductive recruitment of top-down prefrontal-subcortical circuitry in major depression. J
Neurosc 2007; 27:8877–8884
60. Beck AT: Cognitive therapy, behavior therapy, psychoanalysis,
and pharmacotherapy: a cognitive continuum, in Psychotherapy
Research: Where Are We and Where Should We Go? Edited byWilliams JBW, Spitzer RL. New York, Guilford, 1984, pp 114–135
61. Mayberg HS: Defining neurocircuits in depression: insights
from functional neuroimaging studies of diverse treatments.
Psychiatr Ann 2006; 4:258–267
62. O’Reardon JP , Chopra MP , Bergan A, Gallop R, DeRubeis RJ,
Crits-Christoph P: Response to tryptophan depletion in major
depression treated with either cognitive therapy or selectiveserotonin reuptake inhibitor antidepressants. Biol Psychiatry
2004; 55:957–959
63. Stutzmann GE, McEwen BS, LeDoux JE: Serotonin modulation
of sensory inputs to the lateral amygdala: dependency on cor-
ticosterone. J Neurosci 1998; 18:9529–9538
64. Holmes AJ, Pizzagalli DA: Spatiotemporal dynamics of error
processing dysfunctions in major depressive disorder. Arch GenPsychiatry 2008; 65:179–188
65. Abela JRZ, Skitch SA: Dysfunctional attitudes, self-esteem, and
hassles: cognitive vulnerability to depression in children of af-
fectively ill parents. Behav Res Ther 2007; 45:1127–1140