Abstracts

RATIONALE: The effect of valproate (VPA) on human EEG was studied to investigate the changes in brain complexity using nonlinear dynamics analysis. A measure, spatial linear mode complexity(SLMC) is proposed to quantify the complexity of spatial linear modes in multi-channel EEG. METHODS:_Nine patients with complex partial seizure, who were not exposed to antiepileptic drugs previously, were included. Eighteen channel-EEG data were collected before and after VPA therapy. Changes in brain complexity were examined by means of nonlinear analysis using SLMC algorithm of EEG, where SLMC value reflects the brain complexity. Fifteen normal healthy people were included as the control group. RESULTS: The patient's mean SLMC value prior to VPA therapy was significantly lower than that of the control (p=0.02). Most eigenvalues were weighted on the first few modes, and the remaining modes rapidly decay with negligible eigenvalues, which indicates reduced the brain complexity. The average value for all patients has increased after treatment (p=0.074) and came nearer to that of the control group. CONCLUSIONS: These results suggest that epilepsy patients have interictal abnormalities, which can be demonstrated by reduced brain complexity. VPA reverses this trendin part. The nonlinear analysis of EEG may be useful in evaluating the effect of antiepileptic drug.