VLSI for Ambulatory EEG Epilepsy Monitoring

DESCRIPTION (provided by applicant): This research will create a small and inexpensive multichannel physiological data acquisition system. It will allow dense array (128 or 256 channel) EEG systems to be made at a similar cost as conventional (19 or 32 channels) systems, and it will allow conventional systems to become very compact and inexpensive. The initial implementation will be an ambulatory EEG recording using EGI's Geodesic Sensor Net, an electrode system noted for its comfort and ease of installation even with very high channel counts (to 256 electrodes). Recent high-performance laboratory EEG amplifier designs are providing full DC bandwidth capability, by using a single 24-bit delta-sigma A/D converter for each channel. This capability is especially important in epilepsy studies, for prompt recovery from artifactual noise (such as from movements during seizures). In addition, very low frequency (near DC) phenomena may be clinically significant. By combining multiple low-noise input instrumentation amplifier channels and delta-sigma modulators on a single chip, and then integrating the digital filtering and signal processing functions of multiple delta-sigma converters on a single FPGA (Field Programmable Gate Array), we can implement a full-bandwidth data acquisition unit with much smaller overall size and reduced power consumption, compared to existing systems. The system will incorporate local disk storage and a wireless data link for continuous downloading to a base station. Ambulatory systems are gaining increasing acceptance in EEG for epilepsy assessment, because of lower costs of outpatient recording and better opportunities to capture epileptiform events, with longer recording times in a variety of settings. Published research on source localization of interictal epileptic foci has shown that dense array EEG provides substantial improvements in source estimation. Our own recent research at the University of Washington Harborview Hospital shows we can conduct long-term monitoring for several days with the 256-channel system, with quantifiable improvements in accuracy of localizing seizure onset. In addition to the benefits for fully ambulatory applications, the proposed research will allow dense array measures to become cost- effective, while at the same time greatly reducing the cost and size of conventional EEG recording systems. Advances in computational capacity can improve the information yield from medical devices if matched by similar advances in data acquisition technology. The proposed VLSI dense array EEG system will allow low cost, compact, wireless ambulatory acquisition of human brain activity for applications such as long-term monitoring to characterize seizure onset in epilepsy.