A neurophysiologic test battery (consisting of a 24-hour, seven-channel electroencephalogram [EEG], EEG spectral analysis, multiple sleep latency test, visual evoked potentials, critical flicker fusion, and visual contrast sensitivity) was administered twice to 16 patients with juvenile myoclonic epilepsy (JME) in a double-blind, randomized, crossover study comparing two daily doses of sodium valproate (VPA), 1000 mg and 2000 mg. Clinical observation time was 6 months for each dose. Mean total VPA concentration during low-dose treatment was 470.4 mmol/L and during high-dose treatment was 700.0 mmol/L. Ten patients had seizures during low-dose treatment, but only three of these showed spike-wave activity on EEGs. During high-dose treatment, nine patients had seizures; five of these had spike-wave activity. EEG power spectrum did not change between doses. The other tests also showed no change between doses. Our results suggested that EEG and our selection of other neurophysiologic tests were of limited value for monitoring seizure frequency and clinical effects of VPA.