BACKGROUND AND OBJECTIVES: The hypothesis that brain is a nonlinear dynamic system exhibiting deterministic chaos has offered new methods to the investigation of information processing in the brain by analysis and classification of EEG signals. We used the first positiveLyapunov exponent (L1) which is one of indicator of nonlinear dynamic to evaluate the brain function in chemical seizure models METHODS: Lithium-Pilocarpine induced seizure model and kainic acid induced seizure model are used. From serial EEG according to seizure stages. 32.768 sec of data (16.384 data point) were recorded and digirized by a 12-bit analog-digital converter in an IBM PC. The data from serial EEG according to seizure stageswere analyzed for determining the L1. We used the time delays calculated by the method of mutual information to reconstruct the attactor. Time delays of 46-58 msec and enbedding dimensions of 13-19 were used for chemical seizure model. The L1 were calculated for 4 channels. RESULTS: The averaged valued of L1 were serially decreased in both lithium-pilocarpin model and kainic acid model according to increasing seizure stages. CONCLUSION: Our results reveal the decrease of the chaotic activity according to increasing seizure stage. It is suggested that the brain has decreased information procedding and a less flexible neural network during seizure. However further evaluation is required because the significance of these changes are not confirmed.
Animals ; Automatic Data Processing ; Brain ; Classification ; Electroencephalography* ; Kainic Acid ; Nonlinear Dynamics ; Rats* ; Seizures ; Status Epilepticus*