Publication
In the past decade, there has been growing interest in developing AI-enabled neural interfaces for various neurological disorders and emerging brain-machine interface (BMI) applications. The focus has shifted from raw signal acquisition and data compression for off-body processing to intelligent systems featuring on-site signal processing, neural biomarker extraction, and AI. In this tutorial, I discuss the key characteristics, design trade-offs, and recent advances in CMOS-based systems-onchip (SoCs) for diverse categories of intelligent neural prostheses. These categories include real-time symptom tracking and closed-loop stimulation, intelligent BMI systems for movement and communication recovery following paralysis, and beyond.