Data reduction for wireless seizure detection systems

Wireless electroencephalogram (EEG) systems have become an increasingly important tool for the diagnosis and management of epilepsy. One major consideration in using such a wireless EEG-based system is the stringent battery energy constraint at the sensor side. Different solutions to reduce the power consumption are therefore highly desired. Conventionally, the entire EEG signals acquired by the sensor nodes are continuously streamed to an external data server (where seizure detection is carried out). Such approach incurs a high power consumption, which substantially limits the battery life of the sensor node. In this study, we examine the use of data reduction techniques, including compressive sensing-based EEG compression and various low-complexity feature extraction techniques, for reducing the amount of data that has to be transmitted and thereby reducing the required power consumption at the sensor side. The performance of such techniques is evaluated in terms of power consumption and seizure detection efficacy. Results show that by extracting and transmitting only the nonlinear autocorrelation features of the EEG signals to the server, the battery life of the system is increased by 14 times relative to the conventional approach of transmitting the entire EEG signals, while the same seizure detection performance is maintained (94.1% sensitivity and 99.9% specificity).