Embedded EEG Recording Module with Active Electrodes for Motor Imagery Brain-Computer Interface

Abstract

This paper presents a novel modular, portable and low-power electroencephalography (EEG) acquisition system for a Brain-Computer Interface (BCI) application. The system is based on the versatile microcontroller MSP430F5529, to acquire from 4 to 12 EEG channels at 256 Hz with a single 3.7 V supply. The prototype supports both passive and active electrodes; these were also designed and built. Additionally, a graphic visualization interface was developed on the open-source programming language Processing. Common-mode rejection ratio, input-referred noise and magnitude and phase frequency response were measured for each analog EEG channel, these requirements comply the International Federation of Clinical Neurophysiology guidelines. Furthermore, four noise and distortion parameters for AC input signals were evaluated for each ADC channel, based on the IEEE-1241-2000 standard. Power spectrum of EEG recordings using our prototype and the commercial amplifier g.USBamp were compared in open/closed eyes conditions (alfa-band reactivity). It was verified that spectral information retrieval (EEG alfa peak) attained with the prototype is quite similar to the commercial system, and that 60 Hz noise level is reduced when using active instead of passive electrodes, as expected. It was shown that analog pre-processing and noise design considerations are effective ways to get a good quality EEG signal and to enhance certain portions of its frequency spectrum (0.1‑35 Hz). This embedded recording system improves the acquisition of EEG into the frequency bands of interest; the mu (8-13 Hz) and beta (14-27 Hz) EEG rhythms for a motor-imagery BCI application, while rejecting unwanted components (DC, >30 Hz, 60 Hz). With this design, it is possible to reduce noise and interferences in EEG signal to decrease the computational workload in digital post-processing, and thus, may increase the likelihood of achieving a completely stand-alone BCI system for clinical applications.