کنترل توجه و مهار کلمه در اسکیزوفرنی

Abstract Previous studies have suggested that schizophrenia patients do not utilize contextual information efficiently to modulate attentional performance. The goal of the current study was to compare the utilization of context in modulating responses to irrelevant information on the Stroop task between a group of schizophrenia outpatients and matched controls. A single-trial version of the Stroop task was used to investigate performance on the Stroop task under three expectancy conditions. Eleven schizophrenia outpatients (on and off antipsychotic medication) and sixteen matched controls were tested. The schizophrenia patients showed: (1) augmented facilitation; (2) interference comparable to normals; and (3) normal ability to reduce interference under certain experimental circumstances. Schizophrenia patients were able to utilize contextual information under certain conditions and could modulate the magnitude of irrelevant word interference, although they were not able to overcome the prepotent tendency to read the word during the Stroop task as effectively as normals, which was reflected in greater Stroop facilitation. This suggests that the integrity or impairment of cognitive control functions in schizophrenia is related to the complexity of the context representation required to support that function.

. Introduction Cognitive deficiencies have long been reported in schizophrenia with deficits appearing in multiple domains including memory, attention and language (Bleuler, 1911/1950 and Kraepelin, 1919/1971). Inadequate performance, such as associative intrusions in schizophrenia speech, has been attributed mainly to attentional deficits (Cornblatt and Erlenmeyer-Kimling, 1985, Gjerde, 1983 and Nuechterlein and Dawson, 1984) and in particular to the patients’ inability to inhibit irrelevant words or phrases (Maher, 1983 and Manschreck et al., 1988). The study of language processes in schizophrenia has yielded valuable insight into the processes that underlie cognitive abnormalities in this psychiatric disorder (Andreasen et al., 1982 and Salzinger et al., 1970). In addition, many studies have examined the breakdown of attentional processes in schizophrenia patients using lexical stimuli (Nuechterlein and Dawson, 1984, for review). Although many studies of memory and attention in schizophrenia employ lexical stimuli, only a few have measured the impact of context and expectancy in guiding these processes (Cohen et al., 1999, Cohen and Servan-Schreiber, 1992 and Salzinger et al., 1970). The Stroop task is one test that has employed lexical stimuli to examine attentional inhibition. In the Stroop task, reaction time deficits on color naming are often interpreted as an inability to inhibit the irrelevant lexical stimulus (i.e. color-word). The Stroop task is especially relevant for investigating the ability of schizophrenia patients to utilize context to guide attentional inhibition because healthy subjects show attentional modulation of words in the Stroop task under different expectancy conditions (Tzelgov et al., 1992). The present study investigated the ability of schizophrenia patients to suppress irrelevant word processing under various expectancy conditions of the Stroop task. 1.1. The Stroop effect The Stroop task (Stroop, 1935) has been employed extensively in studies of word processing (MacLeod, 1991, for a recent review). In this task, subjects are asked to name the ink color of a word and ignore the meaning of the word (e.g. when presented with the word GREEN in red ink, they are supposed to say ‘red’). Ink-color-naming is slower on incongruent trials (e.g. GREEN in red ink) than on congruent trials (e.g. RED in red ink) or on neutral trials (e.g. XXX or a non-color word in red ink). That is, the to-be-ignored word interferes with reporting the ink color. With young normals (e.g. college undergraduates) it is usually the case that the interference effect (the reaction time, RT, difference between the neutral and the incongruent conditions) is relatively large and stable, whereas the facilitatory effect (RT difference between the congruent and the neutral conditions) is small and may not reach significance. Although other conventions are sometimes used to measure Stroop effects, they can be problematic to interpret as no baseline or neutral stimulus is employed (Henik, 1996). Some researchers have suggested that the Stroop effect measures the automaticity of word processing (Logan, 1980, Logan, 1985, Posner, 1978 and Stroop, 1935). Namely, word reading is initiated without intention and occurs in spite of the subject's effort to suppress it. 1.2. Stroop and expectancy effects Recent work has shown that subjects may reduce the influence of the irrelevant word. Such reductions are dependent upon language proficiency (Tzelgov et al., 1990) and expectations (Tzelgov et al., 1992). In order to study the effect of expectancy, Tzelgov and his colleagues (Tzelgov et al., 1992) manipulated the proportion of neutral trials in order to induce different expectations as to the forthcoming stimulus. They reasoned that if subjects could suppress reading the irrelevant words, they should be in a better position to do so when a large proportion of words required application of such suppression. Tzelgov and his colleagues found that the greater the proportion of color words (relative to neutral words) in a block of trials, the smaller the interference effect. Interestingly, this manipulation affected interference but did not affect facilitation. It is important to note that the neutral proportion effect does not influence performance in the Stroop task through supplying different opportunities to practice the suppression of the irrelevant word. Instead, modulation of the Stroop effect is achieved by creating expectations, which in turn induce subjects to modulate their attentional resources (Tzelgov et al., 1992). In addition, these findings demonstrated that performance on the Stroop task could be dissociated into two distinct components with only one component (i.e. interference) being affected by strategies. The other component producing facilitation was not affected by expectancy and appears to be automatic or reflexive. 1.3. Stroop effect and schizophrenia Attentional deficits in schizophrenia patients are thought to be reflected in a decreased ability to select the appropriate dimension of a stimulus and to inhibit the irrelevant dimension. Since the suppression of the content of the irrelevant word is essential for efficient performance on the Stroop task, this task has been employed in many studies that examine attentional processing in schizophrenia patients. Most researchers have measured the interference component and predicted that it would be larger for schizophrenia patients compared to controls. The results, however, as discussed in greater detail below, were not always consistent. Several investigators have used the serial version of the Stroop task (Abramczyk et al., 1983, Buchanan et al., 1994, Everett et al., 1989, Golden, 1976, Mirsky et al., 1984, Schwartz and Shagass, 1960, Wapner and Krus, 1960 and Wysocki and Sweet, 1985). Most studies reported that schizophrenia patients exhibited slower color naming both on incongruent and neutral cards compared to controls, but did not test directly the interference effect (i.e. the difference between incongruent and neutral cards). However, two studies (Abramczyk et al., 1983 and Buchanan et al., 1994) tested the interference effect directly and reported a significant difference in the predicted direction (i.e. increased interference in the schizophrenia patients). Carter and colleagues (Carter et al., 1992) employed the single-trial version of the Stroop task with unmedicated patients and found a different pattern of results. Schizophrenia patients showed augmented facilitation but no difference in interference compared to control subjects. Later Carter and his colleagues (Carter et al., 1993), also with unmedicated patients, found that the augmented facilitation characterized an undifferentiated subtype of the disorder whereas interference characterized a paranoid subtype. In a number of recent studies that used the single-trial version (Salo et al., 1996, Salo et al., 1997 and Taylor et al., 1996), no difference in Stroop interference between schizophrenia patients and normal controls was found. The various studies mentioned above differed in methods, instructions, and general design so that various aspects of the task and experimental design may be responsible for differences among these studies. In most of the single-trial studies cited above, equal proportions of the Stroop wordtypes were employed. This balanced design eliminated the possibility that the subjects had any expectancy about which word type would appear on the next trial. When such a balanced design was employed, equivalent RT interference was found between schizophrenia patients and matched controls. Other studies have manipulated expectancy or contextual situations in non-Stroop tasks by changing the inter-stimulus interval (ISI) between trials. On a visual continuous performance task, Cohen and Servan-Schreiber (1992) found that at longer ISIs, the performance of the schizophrenia patients suffered compared to matched controls. Their interpretation of this finding was that the context that guided performance on the subsequent trial was no longer available to the schizophrenia patient. This finding of a performance deficit at longer delays has also been reported using an auditory continuous performance task (Cohen et al., 1988) and a lexical disambiguation task (Cohen and Servan-Schreiber, 1992 and Salzinger et al., 1970). Because the ability to use expectancy in schizophrenia patients has been studied in large part by manipulating temporal distance between trials, we were interested in examining whether or not patients with schizophrenia could take advantage of other contextual information such as the proportion of neutral or control conditions (Tzelgov et al., 1992). 1.4. Rationale The goals of the present study are (1) to investigate the effects of expectancy and attentional control on Stroop facilitation and interference; (2) to compare the utilization of context to modulate attention between schizophrenia patients and matched controls; and (3) to test the effects of neuroleptic medication on attentional modulation. Previous studies with healthy college aged subjects (Tzelgov et al., 1992 and Tzelgov et al., 1990) have shown that expectancy, generated by manipulation of proportions of neutral trials, could help subjects reduce the Stroop effect. Accordingly, we manipulated the proportion of neutral trials relative to the proportion of congruent and incongruent trials (the number of congruent and incongruent trials was always equal). We hypothesized that control subjects would show greater interference under larger proportions of neutral trials in conjunction with no change in facilitation. A similar effect for the schizophrenia patients would document their ability to utilize contextual information (% of neutrals) and to control inhibition of word processing in a normal way. As mentioned previously, there is evidence that the facilitatory and the interference components of the Stroop effect are dissociable (MacLeod, 1991 and Tzelgov et al., 1992). Thus, we expected the effect of neutral proportion manipulations to be limited to the interference component of the Stroop effect in both the schizophrenia patients and the matched controls.

Conclusion Although many studies have suggested that schizophrenia patients have impaired context processing, the present results suggest that this may be a function of the complexity of the context representation involved. The schizophrenia patients tested in this study were able to reduce Stroop interference as a result of different proportions of neutral trials, suggesting that when the task representation involved merely adjusting the degree to which attention was allocated to word reading or color naming in order to produce a speeded response, they are able to do so. On the other hand, their ability to use context to overcome the prepotent tendency to read the word during the Stroop task was quite impaired, as reflected by their increased facilitation effects. In a previously published semantic priming study (Henik et al., 1995); Henik et al. reported that although schizophrenia patients showed augmented priming effects, they were still able to anticipate and prepare for the upcoming stimuli. Similar to finding in control subjects, the patients’ responses were speeded up when the prime to target interval [stimulus onset asynchrony (SOA)] was relatively long. Moreover, under the long SOAs the patients coped better with certain aspects of cognitive load. The SOA effect is similar to the foreperiod effect found in reaction time studies in schizophrenia (Bohannon and Strauss, 1983 and Greiffenstein et al., 1981). Thus, at least under certain circumstances, schizophrenia patients are able to maintain task-oriented set and utilize context to prepare for responding. On the other hand Meiran et al. (2000) investigated task switching in schizophrenia patients and concluded that while schizophrenia patients were able to use a preparatory interval in order to update a task set and reduce the magnitude of task-switching RT costs, their overall pattern of performance suggested that they were impaired in representing the task context. In a follow-up experiment these investigators were able to show that increasing the complexity of the task context in normal subjects produced a pattern of performance comparable to that seen in the schizophrenia patients. These results and those of the present study suggest that many control processes that depend upon relatively simple context representations are intact in schizophrenia but that the ability to represent and maintain more complex task contexts, such as those required to overcome a prepotent response tendency or those that involve a highly complex or unconstrained set of stimulus response mappings, is impaired in this illness. The present results have implications not only for research on cognitive processes in schizophrenia but possibly also for work on word processing and the Stroop effect in normal cognition. These results may be interpreted as supporting the hypothesis that the facilitatory and interference components are dissociable parts of the task (Henik et al., 1993, Lindsay and Jacoby, 1994 and Tzelgov et al., 1992). The data suggest that the facilitatory component of the Stroop task involves early processes, those that are more automatic, operate prior to or at the point of lexical entry and thus are less subject to strategic control. If facilitation were due to processes operating post-lexically (after the word was recognized), it should have been influenced by neutral proportion in the same way that the interference component was. In contrast, interference may be due to post-lexical processes that operate after the word has been recognized. Finally, we would like to comment on the subtype issue. The current study was not designed to examine the paranoid and the undifferentiated subtypes separately. The number of patients in each subtype was relatively small, limiting the statistical power needed to investigate this issue.