Correlations between measures of attention and topographical abnormalities of evoked cortical potentials elicited during the Continuous Attention Test (CAT) were assessed in 50 schizophrenic patients, compared to 50 healthy subjects. For each group and for each CAT condition evoked responses consisted of six successive epochs (segments) of stable spatially configured potentials. Quantitative descriptors of those configurations (Lehmann 1987) were referred to the CAT data. In patients: (1) segments III-V were delayed, (2) in the non-target condition, diminished global field power (GFP) emerged, coexisting either with lower amplitude of posterior potentials in segment I and II or with lower amplitude of a central positive potential in segment V, (3) an altered topographic pattern of responses to the target stimulus occurred. In healthy subjects detection of the target (as compared to the non-target condition) was associated with a shift of the location of the positive potential in segments IV and V from a central towards the prefrontal area. In patients, in segment V a similar shift reached frontal, but not prefrontal areas, and additionally, the central areas remained active. Delayed latency and low GFP in segment V in the non-target condition in patients correlated with poor CAT performance. A more posterior location of the positive centroid in segment V during detection of the target correlated with better CAT results, and the associated GFP increase with less prolonged reaction time. The data revealed a possible compensatory role of central and frontal areas in the face of weakened prefrontal functions in schizophrenia.