Selecting IRN for AFE to Achieve Power-Area-Noise Efficiency in Next-Generation Neural Implants

Advanced neural implants demand a 10x increase in input channels for diverse neuroscience applications, necessitating meticulous top-level design decisions to attain optimal noise-power-area efficiency. This paper underscores the pivotal role of analog front-ends (AFEs) and the influence of input-referred noise (IRN) on the energy efficiency and area performance of intelligent, high-channel-count neural implants. We delve into the intricacies of IRN, highlighting its impact on signal-to-noise (SNR) ratio and power/area performance. Our study provides insights into selecting the optimal IRN, offering a pathway for the development of energy-efficient and precise neural implants by introducing four operating regions on how to target the IRN to achieve at least 7x power and area reduction on AFEs.