Abstract

BACKGROUND--Current evidence indicates that, on their own, neither flash visual evoked responses (FVEPs) nor event related potentials (ERPs) are sufficiently useful to the clinician in the very early stages of memory dysfunction. However, the possibilities for the combined use of these measures has not been fully explored. METHODS--This study examined the clinical utility of combined FVEP and ERP-P300 component latencies as predictive markers in 16 patients with Alzheimer's disease, 15 patients with depression, and 21 control subjects. RESULTS--There were significant group differences in FVEP P2 latency (P = 0.004) between the controls and both the depressive patients and those with very mild Alzheimer's disease. There were no statistically significant group differences for the ERP component (N2/P300) amplitudes or latencies. The P300 component latency was positively correlated with both the FVEP N2 and FVEP P2 component latencies in the patients with Alzheimer's disease but not in the control subjects or the depressed patients. A discriminant function, using two ERP and two FVEP component measures, gave an overall correct classification rate for dementia of 78%. In this study of very mildly impaired patients the FVEP latencies provided a more sensitive marker for the presence of cognitive dysfunction than P300 latency delay. CONCLUSIONS--The findings support the use of multimodal evoked potentials in the differential diagnosis of very mild Alzheimer's disease and normal aging.