Copyright 2001 by the American Psychological Association, Inc. 0021-843X/01/S5.00 DOI: 10.1037//0021-843X. 110.2.236 Information Processing and Cognitive Organization in Unipolar Depression: Specificity and Comorbidity Issues David J. A. Dozois University of Western Ontario Keith S. Dobson University of Calgary This study investigated information processing and cognitive organization in clinical depression. The specificity of various cognitive mechanisms to depression was also examined. Twenty-six depressed/ anxious individuals, 24 pure depressives, 25 never-depressed anxious controls, and 25 nonpsychiatric controls completed a modified Stroop task, the Self-Referent Encoding Task, and two tasks designed to assess cognitive structure. Comorbid depressed/anxious, depressed, and anxious groups performed similarly to one another but differed significantly from nonpsychiatric controls, on the processing and organization of negative content. Specificity to depression was also obtained, as both depressed groups endorsed and recalled less positive information and organized positive self-relevant information with less interconnectedness than anxious individuals and nonpsychiatric controls. These results suggest that depressed individuals have an interconnected negative self-representational system and lack a well- organized positive template of self. These findings are discussed in terms of cognitive models of depression and the tripartite model of depression and anxiety. Cognitive features related to unipolar major depression continue to generate considerable interest as a focus for both its conceptu- alization and its amelioration (D. A. Clark, Beck, & Alford, 1999; Ingram, Miranda, & Segal, 1998). Unquestionably, the dominant cognitive paradigm for depression has been Beck's model (Beck, Rush, Shaw, & Emery, 1979; Salkovskis, 1996). According to this theory, maladaptive cognition is implicated in both the etiology and maintenance of the depressive process. At the crux of Beck's theory of depression is the notion of self-schemata, which are seen as enduring internal templates of the self, the world, and the future derived from past experiences (Beck et al., 1979). Ingram (1984; Ingram et al., 1998) has further operationalized this construct by distinguishing between its structural and prepositional elements. The self-system of depressed individuals is believed to consist of a highly interconnected network of negative content. This inter- connectedness, in turn, is purported to influence information pro- cessing (e.g., attention, encoding, retrieval). The cognitive mechanisms postulated to underlie depressive biases and distortions have, for a variety of reasons, not been adequately tested (Ingram, 1990; Z. V. Segal & Ingram, 1994; Z. V. Segal, 1988). For example, the preponderance of research related to cognition in depression has emphasized either the prod- David J. A. Dozois, Department of Psychology, University of Western Ontario. London, Ontario, Canada; Keith S. Dobson, Department of Psy- chology, University of Calgary, Calgary, Alberta, Canada. This research was supported by the Medical Research Council of Can- ada, the Killam Foundation, and the Alberta Heritage Foundation for Medical Research (AHFMR) and is gratefully acknowledged. Correspondence concerning this article should be addressed to David J. A. Dozois, Department of Psychology, University of Western Ontario, London, Ontario N6A 5C2. Canada. Electronic mail may be sent to ddozois@uwo.ca. ucts of the cognitive system (e.g., via self-report) or its operations (e.g., attentional processes; D. A. Clark et al., 1999; D. A. Clark & Steer, 1996; Haaga, 1997). Although some researchers have re- cently attempted to measure cognitive structure (e.g., Z. V. Segal & Gemar, 1997), few studies exist to address this clinically and theoretically important construct, and no studies have investigated the interface among cognitive products, processes, and structures. Consequently, empirical research lags far behind theoretical con- jecture when it comes to understanding how schemata are hierar- chically structured, activated, and deactivated, or how this cogni- tive organization relates to biased processing and/or maladaptive thinking (Ingram, 1990; Weary & Edwards, 1994). An objective of this study was to investigate information processing and cognitive organization in clinical depression. To our knowledge, no studies have examined the functional and structural levels of cognition within the same group of depressed individuals while using iden- tical well-controlled sets of adjectives across tasks. A second purpose of this research was to determine the speci- ficity of these cognitive mechanisms to depression. Several simi- larities exist between depression and anxiety, which have raised concerns as to whether these disorders are empirically distinct entities (Cole, Truglio, & Peeke, 1997; Dobson, 1985a, 1985b; Feldman, 1993; Mineka, Watson, & Clark, 1998). Dobson's (1985b) review of the psychometric literature, for instance, re- vealed considerable overlap between measures of depression and anxiety, with an average correlation of .61. Moreover, research has consistently demonstrated that unipolar depression has high rates of comorbidity with other psychiatric syndromes, most notably the anxiety disorders. The average comorbidity rate of depression with various anxiety disorders is approximately 58% (Mineka et al., 1998; Newman, Moffitt, Caspi, & Silva, 1998). A model that perhaps best explains both the common and distinctive features of depression and anxiety was initially ad- vanced by Watson and Tellegen (1985). According to this two- 236