A 16-Ch CMI-Tolerant Neural AFE with Inherent CM Detection and Shared CM Suppression Achieving 0.006mm2/Ch and 3.1 $\mu \mathrm{W}/\text{Ch}$
Joan Aymerich, Chutham Sawigun, José Cisneros-Fernández, Xiaolin Yang, Carolina Mora López · 2024
This work introduces a multichannel AFE for compact, low-power bidirectional neural interfaces, capable of tolerating common-mode interferences (CMIs) and capturing action potentials (APs) and local-field potentials (LFPs) with a 10kHz bandwidth (BW). By combining noiseless capacitive commonmode (CM) coupling and intrinsic OTA CM detection, our AFE achieves power- and area-efficient CM suppression (CMS) without input impedance$(\mathrm{Z}_{\mathrm{i}\mathrm{n}})$nor noise degradations. Our design allows sharing a single CMS loop across multiple channels, significantly reducing the total area and power. The fabricated proof-of-concept 16-ch AFE prototype tolerates$400\text{mV}_{\mathrm{p}\mathrm{p}}$CMIs, consumes$3.1\mu \mathrm{W}/\text{ch}$, and occupies O.006mm2/ch, thus achieving the best CMI-tolerance vs. area tradeoff. The CMS capability has been fully demonstrated in saline.