Objective: The Wisconsin card sorting test (WCST) is a neuropsychological test widely used to investigate prefrontal dysfunctions and cognitive deficits in schizophrenia. It consists of 128 cards and requires 10 consecutive correct responses to achieve one category. It may be too difficult for schizophrenic patients, i.e. many patients finished the test without achieving any one category. A few studies evaluated correct response percentages on the first four trials of WCST. It was suggested that performance on the first four cards were decreased in schizophrenia related to impairments in evaluating negative feedback. This study evaluated consecutive response patterns on the first four trials of WCST, not just correct % of each card, in deficit and non-deficit schizophrenic patients.

Methods: The subjects were 145 Schizophrenic outpatients diagnosed by DSM-IV criteria and the control group were 423 medical students. The deficit (n=37) and non-deficit (n=108) subgroups of the patients were evaluated by retrospective chart review based on the Schedule for the Deficit Syndrome (SDS). A computerized version of WCST (Wang neuropsychological LAB, USA) was used. Given assign code ‘1’ to the correct responses and ‘0’ to the wrong responses of the WCST, the patients who responded as 1111, 0111, or 0011 on the first four trials were considered as Good performers (n=33) and who responded other ways, such as 1010, 0101, etc. were Poor performers (n=112). The demographic features, clinical features, and classical indices of the WCST were analyzed between groups. The sensitivity and specificity of Good / Poor classification on the first four responses predicting deficit / non-deficit schizophrenia were also calculated.

Results: There were no significant differences in clinical features except for conceptual disorganization and socioeconomic status between deficit and non-deficit subgroup of the patients. Deficit schizophrenic patients had lower outcomes in % correct, categories achieved and conceptual level responses than non-deficit schizophrenic patients. There were no significant differences in clinical features between poor and good performers. Poor performers had lower outcomes in all classical indices which were number of trials, % correct, categories achieved and conceptual level responses. Significantly more patients in the Deficit group were classified as Poor performers on the first four trials. The Poor performers failed to chase up the good performers during the first ten trials. The sensitivity and specificity of Poor / Good performance criteria predicting Deficit or Non-deficit schizophrenia was 89.2% and 26.9%, respectively.

Conclusion: The response patterns of the first four trials in WCST could be an index reflecting the whole WCST performance in schizophrenia. Further researches, for example, reviewing ROC curves of cutoff scores using the response patterns of the first four, five, six trials are necessary to improve its specificity of identifying schizophrenia from normal and deficit from non-deficit subgroup.