An 8-channel scalable EEG acquisition SoC with fully integrated patient-specific seizure classification and recording processor

Tracking seizure activity to determine proper medication requires a small form factor, ultra-low power sensor with continuous EEG classification. Technical challenges arise from: 1) patient-to-patient variation of seizure pattern on EEG, 2) fully integrating an ultra-low power variable dynamic range instrumentation circuits with seizure detection processor, and 3) reducing communication overhead. Reference [1] extracted EEG features locally on-chip to reduce the data being transmitted, and saved power by 1/14 when compared to raw EEG data transmission. However, it still needs data transmission and off-chip classification to detect and to store seizure activity. This paper presents an ultra-low power scalable EEG acquisition SoC for continuous seizure detection and recording with fully integrated patient-specific Support Vector Machine (SVM)-based classification processor.