Quantitative Video Analysis of Complex Epileptic Seizures During Videotelemetry: Increasing the Reliability of EEG Correlation and Behavioral Autocorrelation

The analysis of complex partial seizures constitutes a major clinical problem. Because of the partial perturbation of consciousness during such episodes, variable behavioral patterns may emerge. After seizures are captured by simultaneous video-electroencephalographic (EEG) monitoring, the offline analysis involves autocorrelation of patient behavior over successive seizures and correlation with EEG findings. To avoid attempting to quantify the relationships between behavioral observations and the EEG findings "by eye," a process performed by a neurologist that can be quite subjective, quantitative rendering of relevant movements captured by video monitoring is helpful. The resulting graphic data of relevant movement phenomena can be compared with EEG data and with patient behavior in multiple possible or actual seizure episodes. This quantitative movement data can be generated and linked to the monitoring system using a video preprocessing device. Preliminary results indicate that the preprocessor output data may be autocorrelated over multiple seizures and may also be correlated with simultaneous EEG features to achieve greater reliability in categorization of seizure type and focus of origin in some cases.