Working Memory for Visual Features and Conjunctions in Schizophrenia James M. Gold, Christopher M. Wilk, Robert P. McMahon, and Robert W. Buchanan Maryland Psychiatric Research Center, University of Maryland School of Medicine Steven J. Luck University of Iowa The visual working memory (WM) storage capacity of patients with schizophrenia was investigated using a change detection paradigm. Participants were presented with 2, 3, 4, or 6 colored bars with testing of both single feature (color, orientation) and feature conjunction conditions. Patients performed signif- icantly worse than controls at all set sizes but demonstrated normal feature binding. Unlike controls, patient WM capacity declined at set size 6 relative to set size 4. Impairments with subcapacity arrays suggest a deficit in task set maintenance: Greater impairment for supercapacity set sizes suggests a deficit in the ability to selectively encode information for WM storage. Thus, the WM impairment in schizo- phrenia appears to be a consequence of attentional deficits rather than a reduction in storage capacity. Working Memory Capacity Working memory impairments in patients with schizophrenia have been the focus of an intensive multidisciplinary research effort in the past decade. Behavioral studies have documented reliable deficits in a wide array of experimental paradigms that are thought to tap working memory (see Keefe, 2000, for a review). Furthermore, there have been several demonstrations that WM impairments in schizophrenia are related to patient performance on more cognitively complex measures of language comprehension, free recall, problem solving, and reasoning, suggesting that WM deficits may be implicated in the broad cognitive compromise that is characteristic of the illness (Condray, Steinhauer, van Kammen, & Kasparek, 1996; Gold, Carpenter, Randolph, Goldberg, & Weinberger, 1997; Perry et al., 2001; Stone, Gabrieli, Stebbins, & Sullivan, 1998). Functional neuroimaging studies have docu- mented a range of prefrontal metabolic abnormalities in patients during the performance of several different WM tasks (Barch et al., 2001; Callicott et al., 2000; Carter et al., 1998; Manoach et al., 2000; Stevens, Goldman-Rakic, Gore, Fulbright, & Wexler, 1998). These abnormalities are most robust at higher working memory loads (e.g., 2-back vs. 1-back in n-back paradigms), suggesting that an understanding of the specific mechanisms that are impli- cated in WM deficits may provide an important window on the nature of cognitive dysfunction and the neurobiology of schizo- phrenia. The present study assesses the hypothesis that schizophre- nia involves a reduction in WM capacity. The term capacity has two distinctly different meanings in the context of WM systems. First, WM systems include a storage component with a limited storage capacity, which determines how much information can be present in the system at one time. Additional processes may be recruited to facilitate information maintenance over time. The concept of storage capacity should not be confused with retention: Storage capacity refers to an amount of information, whereas retention refers to performance over time. Second, WM systems include a processing component (e.g., the central executive in the model of Baddeley, 1986); this component may be limited in its processing capacity, which refers to the amount of information that can be processed in a given period of time. These two types of capacity are analogous to a computer’s memory capacity and processor speed, respectively, and they are conceptually independent: An individual may have a normal stor- age capacity and an impaired processing capacity, or vice versa. However, just as decreases in either the memory size or the processor speed of a computer will lead to decreased performance on most computing tasks, an impairment in working memory storage capacity may lead to performance deficits in behavioral tasks that are designed to tap processing capacity, and vice versa. For example, an individual with impaired processing capacity may exhibit deficits in digit-span tasks because impaired processing makes it difficult to encode and rapidly rehearse the digits in a manner that takes full advantage of the available storage capacity. Similarly, an individual with a reduced storage capacity may be unable to represent intermediate results in tasks such as mental arithmetic, leading to impaired performance. Deficits in either storage capacity or processing capacity would be expected to lead to impaired performance across a wide spectrum of everyday tasks, so both are important areas of inquiry. However, different neural systems may underlie the storage and processing components of WM; it is therefore important to determine whether schizophrenia involves deficits in the storage component, the processing compo- nent, or both. The present study was designed to determine the extent to which schizophrenia involves a specific deficit in storage capacity. Reduced storage capacity should lead to a decrease in the number of items in a memory set that a subject can reproduce without error. In auditory digit-span tasks, for example, partici- pants are presented with varying numbers of digits for subsequent James M. Gold, Christopher M. Wilk, Robert P. McMahon, and Robert W. Buchanan, Maryland Psychiatric Research Center, University of Mary- land School of Medicine; Steven J. Luck, Department of Psychology, University of Iowa. This research was supported in part by National Institute of Mental Health Grants MH-40279 and MH-65034. We gratefully acknowledge the computer programming contribution of Joel Bish. Correspondence concerning this article should be addressed to James M. Gold, Maryland Psychiatric Research Center, P.O. Box 21247, Baltimore, Maryland 21228. E-mail: jgold@mprc.umaryland.edu Journal of Abnormal Psychology Copyright 2003 by the American Psychological Association, Inc. 2003, Vol. 112, No. 1, 61–71 0021-843X/03/$12.00 DOI: 10.1037/0021-843X.112.1.61 61