Effects of haloperidol on cognition in schizophrenia patients depend on baseline performance: a saccadic eye movement study.

Schizophrenic patients are heterogeneous with respect to voluntary eye movement performance, with some showing impairment (e.g., high antisaccade error rates) and others having intact performance. To investigate how this heterogeneity may correlate with different cognitive outcomes after treatment, we used a prosaccade and antisaccade task to investigate the effects of haloperidol in schizophrenic subjects at three time points: baseline (before medication), 3-5 days post-medication, and 12-14 days post-medication. We also investigated changes on the Stroop Task and the Positive and Negative Syndrome Scale (PANSS) in these same subjects. Results were compared to matched controls. When considered as a single patient group, haloperidol had no effects across sessions on reflexive and voluntary saccadic eye movements of schizophrenic patients. In contrast, the performance of the Control group improved slightly but significantly across sessions on the voluntary eye movement task. When each subject was considered separately, interestingly, for schizophrenic patients change in voluntary eye movement performance across sessions depended on the baseline performance in a non-monotonic manner. That is, there was maximal worsening of voluntary eye movement performance at an intermediate level of baseline performance and the worsening decreased on either side of this intermediate baseline level. When patients were divided into categorical subgroups (nonimpaired and impaired), consistent with the non-monotonic relationship, haloperidol worsened voluntary eye movement performance in the nonimpaired patients and improved performance in the impaired patients. These results were only partially reflected in the Stroop Test. Both patient subgroups showed clinically significant improvement over time as measured by the PANSS. These findings suggest that haloperidol has different effects on cognitive performance in impaired and nonimpaired schizophrenic patients that are not evident in clinical ratings based on the PANSS. Given that good cognitive function is important for long-term prognosis and that there is heterogeneity in schizophrenia, these findings are critical for optimal evaluation and treatment of schizophrenic patients.