Auditory event-related potential indices of fronto-temporal information processing in schizophrenia syndromes: valid outcome prediction of clozapine therapy in a three-year follow-up.

Reliability, specificity, and validity of three event-related potential (ERP) indices of non-attended and attended auditory information processing [the mismatch negativity (MMN), the novelty-P3a and the target-P3b] were assessed in 21 healthy subjects and 25 schizophrenic patients while performing a visual discrimination task (Ignore condition) and, subsequently, an auditory discrimination task (Attend condition). Re-test reliability, as measured in a subset of 10 healthy subjects and 9 patients respectively, ranged from r fl 0.4 to r fl 0.8. In both groups P3a was found to be smaller in the Ignore than in the Attend condition. Patients had significantly larger P3a amplitudes than healthy subjects while their MMN and P3b amplitudes were smaller. Two ERP factors were extracted in healthy subjects: (1) ‘novelty reaction’ with high loading scores for P3a and (2) ‘deviant detection’ with high loading scores for MMN. P3b was predominantly loading on the first factor thus confirming partly overlapping P3a and P3b generators. However, considerable variance was also shared by the second factor, particularly in patients. External validity of the ERP factors was confirmed post hoc. Particularly ERP indices of deviant detection were found to be associated with negative symptoms. In order to assess further clinical relevance, therapy response to clozapine was followed-up over 3 yr in a sub-sample of 17 patients. Poor treatment response was associated with high Novelty Factor scores and large P3a amplitudes whereas good response was associated with high Deviant Factor scores and relatively intact MMN. It is concluded that ERP abnormalities provide a severity index of brain impairment and a measure of predicting therapeutic outcome. Received 1 November 1998; Reviewed 18 January 1999; Revised 14 March 1999; Accepted 20 March 1999