Hardware Implementation of an Adaptive Data Acquisition System for Real-World EEG

Collecting EEG involves digitizing a very small signal across a vast potential dynamic range, particularly within real-world neuroimaging conditions, where noise can be especially prominent. Conventional methods require highresolution, power-hungry data acquisition systems (DAQs), creating limits on usable time before manual interaction is necessary for recharge. Here, we discuss continued work on an alternative DAQ approach capable of acquiring high resolution data with ultra-low power use by adjusting parameters of the analog front end (AFE) in real time to allow use of low-resolution ADCs. This work compares signal quality of a hardware implementation of our adaptive AFE DAQ to that of an industry standard DAQ. Results demonstrate successful reconstruction of signals in both clean and noisy EEG monitoring environments at low bit-depths while maintaining high correlation and low standard deviation of error. This suggests promise for a fully integrated implementation with substantially lower power consumption.