A Convergent–Divergent Approach to Context Processing, General Intellectual Functioning, and the Genetic Liability to Schizophrenia Angus W. MacDonald III, Vina M. Goghari, and Brian M. Hicks University of Minnesota, Twin Cities Campus Janine D. Flory Mount Sinai School of Medicine and University of Pittsburgh Cameron S. Carter University of California, Davis Stephen B. Manuck University of Pittsburgh Convergent and divergent validity are critically important in developing psychological measures that reveal interpretable deficits in disordered populations. This article reports on 2 studies that evaluated the validity of context processing measures. In Experiment 1, a confirmatory factor analysis of data from 481 healthy adults established the convergent validity of 2 context processing measures and showed that context processing accounted for significant amounts of variance in standard IQ and working memory measures. In Experiment 2, 20 schizophrenia patients, 16 of their healthy siblings, and 28 controls were evaluated using a novel, short context processing measure, the dot pattern expectancy (DPX) task. The DPX was sensitive to specific deficits in schizophrenia patients and their healthy siblings. These findings support the construct validity of context processing measures, suggest context processing is a component of intellectual functioning, and demonstrate that brief context processing measures remain sensitive to psychopathological deficits. Keywords: context processing, schizophrenia, intelligence, endophenotype The two studies reported in this article contribute to the ongoing expansion of the neuropsychologist’s toolbox to include more measures of molecular, explanatory cognitive mechanisms such as those developed in experimental cognitive psychology (Mac- Donald & Carter, 2002). The focus of these studies is context processing, a construct that has been generative in both experi- mental psychology and psychopathology. Here we evaluate several aspects of the validity of context processing measures and the sensitivity of a novel measure to specific deficits in context pro- cessing among schizophrenia patients and their healthy siblings. Context processing is conceptualized as the component of cog- nitive control that represents and actively maintains task-relevant information despite subsequent noise (Cohen & Servan-Schreiber, 1992; E. K. Miller, 2000). Task-relevant information includes the environmental stimuli, instructions, or goals that must be inte- grated to guide behavior. Such guidance is particularly needed to make novel or secondary responses. For example, on your way to work, the context of having to pick up doughnuts allows you to change your usual route accordingly. Failure to maintain this context can lead to very disappointed colleagues. Changing a habitual route requires more than simply storing this errand in short-term memory as one might a letter string; the stored infor- mation must be actively maintained despite environmental distrac- tions and transformed into support for a novel response (e.g., turning at the stoplight). This hallmark of context processing is related to earlier cognitive constructs such as selective attention (Norman & Shallice, 1986), distractibility (Oltmanns & Neale, 1975), and working memory’s central executive (Baddeley & Hitch, 1974), all of which speak to the need for manipulating incoming information to control responses on the basis of abstract task demands. This kind of context processing has been the focus of a number of connectionist models. That is, unlike many other executive functions, mathematical simulations of parallel units performing simple computations have been able to account for experimental findings across a number of (primarily verbal) context processing tasks (e.g., Braver, Barch, & Cohen, 1999; Cohen, Dunbar, & McClellend, 1990). Furthermore, functional MRI and electrophysiological experiments suggest that two aspects of context processing— context representation and con- text maintenance—are associated with activity in middle frontal gyrus (Barch et al., 1997; Dias, Foxe, & Javitt, 2003; Mac- Donald, Cohen, Stenger, & Carter, 2000; Perlstein, Dixit, Carter, Noll, & Cohen, 2003). Validity of Context Processing Measures Although formal and neurobiological components of context processing are well documented in the literature, there is little Angus W. MacDonald III, Vina M. Goghari, and Brian M. Hicks, Department of Psychology, University of Minnesota, Twin Cities Campus; Janine D. Flory, Department of Psychiatry, Mount Sinai School of Medi- cine and Department of Psychology, University of Pittsburgh; Cameron S. Carter, Center for Neuroimaging, University of California, Davis; Stephen B. Manuck, Department of Psychology, University of Pittsburgh. We thank Leslie Hart, Victor Ortega, Melissa K. Johnson, Theresa M. Becker, Michael F. Pogue-Geile, and Janet Lower. This work was sup- ported by National Institutes of Health Grants PO1 HL-40962 and RO1 HL-65137 to Stephen B. Manuck. Angus W. MacDonald was supported by National Institute of Mental Health Grant T32 MH18269. This work was conducted at the Department of Psychology, University of Pittsburgh. Computer tasks cited in this project are publicly available at http:// www.psych.umn.edu/research/tricam/ Correspondence concerning this article should be addressed to Angus W. MacDonald III, Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455. E-mail: angus@umn.edu Neuropsychology Copyright 2005 by the American Psychological Association 2005, Vol. 19, No. 6, 814 – 821 0894-4105/05/$12.00 DOI: 10.1037/0894-4105.19.6.814 814