Multibit Local Neighborhood Difference Pattern Optimization for Seizure Detection of West Syndrome EEG Signals

West syndrome is a common age-related epileptic encephalopathy, which seriously threatens children’s intellectual development. The accurate detection of seizures in the electroencephalogram (EEG) signals of patients with West syndrome is helpful for doctors to make effective diagnosis. This article proposes a seizure detection algorithm based on optimized multibit local neighborhood difference pattern (MLNDP) for robust and accurate detection of ictal and interictal periods of West syndrome. Different from the traditional local domain differential pattern, MLNDP can construct multibit relationship between neighboring bits in each channel and combine the adder-based multichannel pattern. In addition, the artificial rabbits optimization (ARO) algorithm is used to select the optimal MLNDP parameters. The EEG data of 25 patients with West syndrome are recorded by Children’s Hospital, Zhejiang University School of Medicine for fivefold cross validation evaluation. The experimental results show that the average of accuracy, precision, sensitivity, and F1 score of the algorithm are 97.18%, 97.03%, 97.43%, and 97.16%, respectively.