A bulk-driven lowpass filter for EEG signal

Abstract

© 2020 IEEE. A typical Electroencephalography (EEG) acquisition system must include a lowpass filter (LPF) unit. According to the clinical applications, EEG signal have a conventional frequency range of 0.1-Hz to 70-Hz. The cutoff frequency for an LPF in EEG acquisition system should be in the same range. Moreover, to minimize size and power-consumption of an LPF is challenging due to circuit components and schematics. The subthreshold buffer-based biquadratic cell by Thanapitak et al., which is a compact nanopower electrocardiogram (ECG) lowpass filter with low supply consumption, have recently been developed. This buffer-based biquad has a 100-Hz cutoff frequency. Based on this paper, we designed and analyzed an LPF for conventional EEG signals. Our LPF was designed with a foundation of the buffer-based biquad and supplemented by a bulk-driven method. We modified a fourth-order LPF from a pseudo-differential biquad, which combined two bulk-driven buffer-based biquads together. By simulation, our LPF has an average cutoff frequency of 10.52±0.5-Hz which is suitable for conventional EEG measurements as expected. The simulation results also shown that a dynamic range of our bulk-driven are 68.62-dB with a low output noise. After comparison with the conventional bufferbased biquad, our LPF have an improved Figure-of-Merit in nanopower filter to 11.67-aJ.