Science Brief

Treatment changes in the depressive self-schema

Research examines effects of cognitive therapy and medication on the structure of the self-schema.

By David J. A. Dozois, PhD and Lena C. Quilty

David J. A. DozoisDavid J. A. Dozois is director of the clinical psychology graduate program and full professor of psychology at the University of Western Ontario. He is a fellow of the Canadian Psychological Association, Section on Clinical Psychology, and the Academy of Cognitive Therapy and is a former Beck Institute Scholar at the Beck Institute for Cognitive Therapy and Research. Dozois’ research focuses on cognitive vulnerability to depression and cognitive-behavioral theory and therapy. He is editor of Cognitive-Behavioral Therapy: General Strategies (in press, Wiley) and co-editor of Prevention of Anxiety and Depression: Theory, Research and Practice (2004, American Psychological Association) and Risk Factors in Depression (2008; Elsevier/Academic Press). Dozois has also been involved administratively in various organizations, most recently serving as president of the Canadian Psychological Association (2011-12). He also maintains a small private practice. Author website.

Lena C. QuiltyLena C. Quilty is an independent scientist in the clinical research department at the Centre for Addiction and Mental Health (CAMH) in Toronto. Quilty received her doctoral degree from the University of Waterloo in 2006, and subsequently completed a postdoctoral fellowship funded by the Canadian Institutes of Health Research at CAMH and the University of Toronto in 2009. She is a registered clinical psychologist, with active involvement in clinical research, service delivery and training. Quilty’s research examines personality and cognitive mediators and moderators of illness course and outcome, with a particular focus on the role of dimensional personality traits and thinking styles in cognitive behavioral therapy for depression. Quilty further studies interventions for co-occurring mood disorder and behavioral difficulties, such as binge eating and alcohol use. Author website.

The depressive self-schema

The depressive schema is a well-organized and interconnected negative internal representation of self. Believed to develop through early life experiences and to remain dormant until triggered by negative life events (Beck, Rush, Shaw, & Emery, 1979), a depressive self-schema has long been identified as a key factor associated with depression risk (see Beck & Dozois, 2011; Dozois & Beck, 2008).

This deep level of cognition, once activated by negative life events (such as loss, failure or rejection) purportedly impacts surface-level cognitions (e.g., information processing, dysfunctional attitudes, automatic thoughts). As such, the schema is considered a crucial variable in vulnerability models of depression. The schema is believed to be characterized by both its content (e.g., negative absolutist beliefs) and its structure or organization (Ingram, Miranda, & Segal, 1998). Although many researchers have examined its content, few studies have attempted to examine the structure of the self-schema.

The Psychological Distance Scaling Task (PDST; Dozois & Dobson, 2001a, 2001b) was developed to measure the structure of the schema. Participants are presented with a square grid divided into four quadrants on the computer screen. The x-axis pertains to self-descriptiveness and is anchored with "Very much like me" on the right and "Not at all like me" on the left. The y-axis taps into the valence of the word and is anchored with "Very positive" at the top and "Very negative" at the bottom. Adjectives are displayed in the middle of the grid. Using the computer mouse, participants consider both axes, and place each adjective on the grid in terms of where it fits in psychological space for them. After each response, a new grid and new adjective are displayed on the screen, until all adjectives are presented. The  computer records the X and Y coordinate point for each adjective and calculates the interstimulus distance among the positive schematic adjectives and among the negative schematic adjectives.  The computations used to derive self-relevant distances entail dividing the sum of squared positive or negative self-relevant distances by the total number of possible self-descriptive positive or negative distances (see Seeds & Dozois, 2010). An assumption of this task is that less distance among adjectives is indicative of greater interconnectedness or consolidation of self-referent content, whereas greater distance among adjectives is indicative of less interconnectedness or consolidation.

A number of studies have shown that a well-organized negative representation of self (i.e., the organization of the self-schema) meets sensitivity (Dozois & Dobson, 2001b, Lumley, Dozois, Hennig, & Marsh, 2012; Seeds & Dozois, 2010), specificity (Dozois & Dobson, 2001b, Dozois & Frewen, 2006; Lumley et al., 2012) and stability (Dozois, 2007; Dozois & Dobson, 2001a) criteria as a vulnerability factor for depression. People with depression, for instance, show well-interconnected negative content and loosely connected positive content. Studies using this method have also been successful in differentiating the self-schema structures observed in depression from those seen in anxiety (Dozois & Dobson, 2001b; Dozois & Frewen, 2006). The interaction of cognitive organization and negative life events also predicts depression prospectively (Seeds & Dozois, 2010).

The stability of negative cognitive organization has also been supported. An early trial that examined this idea followed 45 depressed individuals over a six month period. The hypothesis was that individuals who remitted from depression would show a significant cognitive shift in information-processing (e.g., deactivation of negative processing — for example, attention and memory biases) but that temporal stability would be found on the PDST. Individuals who improved from a depressive episode showed an increase in positive processing and a decrease in negative processing over time. As predicted, however, there was no significant change over time in negative interpersonal structure (i.e., the organization of negative adjective content; see Dozois & Dobson, 2001a). This finding was replicated in an independent sample of 54 patients (Dozois, 2007). In this replication study, interpersonal content remained well-organized even as patients moved from a depressed to a remitted state. Together, these studies support one of the central tenets of Beck’s cognitive theory of depression that negative self-schemas may be present but latent and that, once activated they may proceed to impact various processing biases associated with depressed mood (Beck et al., 1979; see Dozois & Beck, 2008).

Thus, cognitive structure or organization for interpersonal content appears to be a stable vulnerability factor for depression. Stability, however, doesn’t imply that a vulnerability factor is impermeable to change. It is possible, for example, that cognitive therapy (CT) is able to alter these negative cognitive structures. CT is comparable in effectiveness to behavior therapy, other bona fide psychological treatments and antidepressant medication for an acute episode of depression, with each treatment producing superior results compared to placebo (see Beck & Dozois, 2011). CT also carries an advantage, relative to antidepressant medication, for the prevention of relapse (Glogcuen, Cottraux, Cucherat, & Blackburn, 1998).

The self-schema and treatment

The precise mechanisms underlying the prophylactic power of CT are not presently known. One possibility is that CT and antidepressant medication may both change certain aspects of negative thinking (such as information processing, automatic thoughts, dysfunctional attitudes) but that cognitive therapy also alters the “deeper” cognitive structures that give rise to relapse (DeRubeis, Webb, Tang, & Beck, 2010; Garratt, Ingram, Rand, & Sawalani, 2007).

Consistent with this idea, Segal, Gemar and Williams (1999) compared patients who had successfully completed either CT or pharmacotherapy. After remission, participants were administered the Dysfunctional Attitude Scale (DAS), a self-report measure of negative beliefs and attitudes concerning self. They were subsequently induced into a dysphoric mood state and then administered a parallel form of the DAS. Individuals who received antidepressant medication showed an elevation of DAS scores whereas individuals in the CT group did not. Segal et al. (2006) showed that this activation was predictive of subsequent relapse. Thus, it is conceivable that CT changes an individual’s core negative structures and that this shift may be responsible for lasting therapeutic gains. 

Also consistent with this idea, are the findings from a trial which compared the combination of cognitive therapy and pharmacotherapy (CT+PT) to pharmacotherapy (PT) alone (Dozois et al., 2009). Patients were randomly assigned to one of the two conditions. CT was provided for 15 individual sessions (one hour/week) and administered according to the empirically-supported protocol outlined by Beck and his colleagues (Beck et al., 1979). PT involved medication plus clinical management (SSRI or SNRI plus augmentation if needed, following the Canadian Network for Mood and Anxiety Treatment [CANMAT] guidelines; see Kennedy et al., 2009).

There were no significant between-group differences on age, education, marital status, ethnicity, previous depressive episodes, suicide attempts, current medications or comorbidity. Similarly, no group differences were obtained on depression or anxiety at initial assessment. No group differences were found post-treatment on symptom scores. In other words, both treatments were equally effective at treating depression.

Consistent with the hypothesis that negative thinking would improve as depression improved, automatic thoughts changed significantly in both groups — positive automatic thoughts increased significantly and negative automatic thoughts decreased significantly (with no statistically significant between-group differences). Significant changes were also evident on the Dysfunctional Attitudes Scale — both groups showed a significant decrease in dysfunctional attitudes from pre- to post-treatment, with no significant between-group differences.

Individuals treated with CT+PT, however, showed significantly greater cognitive organization of positive interpersonal content and less well-connected negative interpersonal content than did individuals treated with PT alone. Moreover, individuals in the CT+PT group showed significant pre-post differences on positive and negative cognitive organization, whereas a shift in cognitive structure was not evident in the PT group (Dozois et al., 2009). These results are intriguing in light of previous research which has shown that the organization of interpersonal negative content is stable despite the remission of depressive symptoms (Dozois, 2007; Dozois & Dobson, 2001a). It appears that cognitive therapy is able to modify these stable cognitive structures, an effect that was unique to CT+PT.

These results suggest that, although both medication and CT improve depressive symptoms, automatic thoughts and dysfunctional attitudes, CT may offer more in terms of deeper structural change than medication. An important caveat is that this study examined only CT+PT compared to PT alone. It is possible that it was the combination of interventions rather than CT alone that resulted in this change. As such, there is a need to replicate this study comparing CT alone to medication alone.

Lena Quilty and her colleagues recently completed a trial that compared cognitive behavioral therapy (CBT) to pharmacotherapy on cognitive products, processes and structure. A sample of 104 patients were randomly assigned to CBT (n = 54) or PT (n = 50). The dropout rate was 9 percent and 14 percent, respectively for CBT and PT, leaving a final sample of 92. Preliminary data analyses revealed that over the course of 16 weeks of treatment, both the CBT and the PT groups showed a significant decrease in psychological distance for positive content (so positive content became more interconnected over time). There were no significant between-group differences, however, nor a significant interaction between time and treatment group. Similarly, both groups showed a significant increase in negative distance (i.e., there was less interconnectedness of negative content over the course of treatment). In contrast to the previous study, however, no group differences were found and the interaction of time and treatment group was not significant.

Conclusion and future directions

Cognitive organization appears to be an important vulnerability factor for depression. This variable, assessed via the Psychological Distance Scaling Task, appears to demonstrate sensitivity, specificity and stability. Across two independent trials, negative cognitive organization remained well interconnected even though people improved significantly from a depressive episode. 

The impact of cognitive therapy and antidepressant medication on cognitive organization has also been examined in two trials. In the first trial, cognitive therapy outperformed medication in shifting cognitive organization. However, the more recent trial found no significant differences between groups on cognitive organization. Instead, both treatments resulted in a significant shift — the positive content became significantly more interconnected and the negative content less interconnected.

Why were there no differences between groups in the second trial? What may account for the differences between studies? One argument might be that the combination of CT+PT resulted in stronger effects on cognitive change variables; however, in the second trial, CBT and pharmacotherapy each independently showed a significant shift on cognitive organization. In addition, it was the PT group in the previous trial that did not shift significantly for negative content whereas it did in the subsequent trial. The average severity of depression in the initial trial was somewhat higher than in the second trial — this may have rendered CT more powerful in changing cognitive structure in the initial trial, whereas both conditions were able to do so in the second trial.

Another possibility is that the pharmacological treatment was superior in the second trial. This explanation seems highly unlikely, however. In both trials, the pharmacotherapy provided was top of the line: rigorous CANMAT guidelines were followed closely and the psychiatrist was free to switch or augment the medication. It is also possible that the first trial was underpowered compared to the second trial — although effects were found for CT+PT, there was not sufficient statistical power to detect effects in the PT alone arm.

These more recent findings are just preliminary and should be treated as such. Quilty and colleagues will be conducting more nuanced lagged analyses to see if the causal pathways between psychological distance and depression are different across treatment groups.

Regardless of the exact reason, these findings call into question the idea that cognitive organization or structure shifts uniquely in CT. The more recent trial suggests that pharmacotherapy may also be capable of shifting these stable cognitive structures — however, this does not rule out the possibility that the cognitive shift is perhaps the final common pathway.


Beck, A. T., & Dozois, D. J. A. (2011). Cognitive therapy: Current status and future directions. Annual Review of Medicine, 62, 397-409.

Beck, A.T., Rush, A.J., Shaw, B.F., & Emery, G. (1979). Cognitive therapy of depression. New York: Guilford.

DeRubeis, R. J., Webb, C. A., Tang, T. Z., & Beck, A. T. (2010). Cognitive therapy. In K. S. Dobson (Ed.), Handbook of cognitive-behavioral therapies (3rd ed., pp. 277-316). New York: Guilford.

Dozois, D. J. A. (2007). Stability of negative self-structures: A longitudinal comparison of depressed, remitted, and nonpsychiatric controls. Journal of Clinical Psychology, 63, 319-338.

Dozois, D. J. A., & Beck, A. T. (2008). Cognitive schemas, beliefs and assumptions. In K. S. Dobson & D. J. A. Dozois (Eds.), Risk factors in depression (pp. 121-143). Oxford, England: Elsevier/Academic Press.

Dozois, D. J. A., Bieling, P. J., Patelis-Siotis, I., Hoar, L., Chudzik, S., McCabe, K., & Westra, H. A. (2009). Changes in self-schema structure in cognitive therapy for major depressive disorder: A randomized clinical trial. Journal of Consulting and Clinical Psychology, 77, 1078-1088.

Dozois, D. J. A., & Dobson, K. S. (2001a). A longitudinal investigation of information processing and cognitive organization in clinical depression: Stability of schematic interconnectedness. Journal of Consulting and Clinical Psychology, 69, 914-925.

Dozois, D. J. A., & Dobson, K. S. (2001b). Information processing and cognitive organization in unipolar depression: Specificity and comorbidity issues. Journal of Abnormal Psychology, 110, 236-246.

Dozois, D. J. A., & Frewen, P. A. (2006). Specificity of cognitive structure in depression and social phobia: A comparison of interpersonal and achievement content. Journal of Affective Disorders, 90, 101-109.

Garratt, G., Ingram, R. E., Rand, K. L., & Sawalani, G. (2007). Cognitive processes in cognitive therapy: Evaluation of the mechanisms of change in the treatment of depression. Clinical Psychology: Science and Practice, 14, 224-239.

Glogcuen, V., Cottraux, J., Cucherat, M., & Blackburn, I. (1998). A meta-analysis of the effects of cognitive therapy in depression. Journal of Affective Disorders, 49, 59-72.

Ingram, R. E., Miranda, J., & Segal, Z. V. (1998). Cognitive vulnerability to depression. New York: Guilford Press.

Kennedy, S. H., Lam, R. W., Parikh, S. V., Patten, S. B., & Ravindran, A. V. (2009). Canadian Network for Mood and Anxiety Treatments (CANMAT) Clinical guidelines for the management of major depressive disorder in adults. Journal of Affective Disorders, 117, 51-52.

Lumley, M. N., Dozois, D. J. A., Hennig, K., & Marsh, A. (2012). Cognitive organization, perceptions of parenting and depression symptoms in early adolescence. Cognitive Therapy and Research, 36, 300-310.

Seeds, P. M., & Dozois, D. J. A. (2010). Prospective evaluation of a cognitive vulnerability-stress model for depression: The interaction of schema self-structure and negative life events. Journal of Clinical Psychology, 66, 1307-1323.

Segal, Z. V., & Gemar, M., & Williams, S. (1999). Differential cognitive response to a mood challenge following successful cognitive therapy or pharmacotherapy for unipolar depression. Journal of Abnormal Psychology, 108, 3-10.

Segal, Z. V., Kennedy, S., Gemar, M., Hood, K., Pedersen, R., & Buis, T. (2006). Cognitive reactivity to sad mood provocation and the prediction of depressive relapse. Archives of General Psychiatry, 63, 749-755.

The views expressed in Science Briefs are those of the authors and do not reflect the opinions or policies of APA.