Clinical application of the P3 component of event-related potentials. II. Dementia, depression and schizophrenia.

Patients with dementia, schizophrenia and depression were tested with analogous auditory and visual event-related potential (ERP) paradigms designed to elicit a large P3. The patient groups were compared to age normative predictions derived from a large control sample for a number of ERP and behavioral variables. The results were similar for the auditory and visual paradigms. P3 latency was prolonged two or more S.D.s beyond that predicted by age for less than one-half of the demented patients. This latency prolongation was significant for the group as a whole but would result in too many false negatives if used diagnostically for individuals. Furthermore, increased P3 latency was not specific, as the schizophrenic patients also had later P3s. The amplitude of P3 was reduced in the demented patients, but it was also smaller in other patient groups. The only variable which distinguished the demented patients from both controls and from the other patients was the single trial P3 latency/RT correlation. The demented patients, as a group, had significantly lower P3 latency/RT correlations, but this effect also was not sensitive enough to be diagnostic for individuals. The data from these two paradigms suggest that the P3 amplitude and latency abnormalities observed reflect a common, rather than a diagnostically specific deficit. This study is in contrast to some others which report much more sensitivity and specificity in the use of P3 latency in the diagnosis of dementia. Differences in task demands, patient samples and ERP analysis techniques might explain some of the discrepancy.