Event-Driven Analog-to-Digital Converter for High-Density Neural Interfaces

High-density neural interfaces are central to next-generation brain-machine interfaces (BMIs), neuroprosthetics, and large-scale neuroscience research. Modern neural probes integrate hundreds to thousands of recording channels, creating stringent requirements on power consumption, data bandwidth, and scalability. Conventional Nyquist-rate ADCs digitize neural signals continuously, resulting in redundant data, excessive power dissipation, and bandwidth bottlenecks - especially given the sparse, event-based nature of neural activity.

Event-driven (or asynchronous) signal acquisition architectures offer a promising alternative. By digitizing neural signals only when meaningful activity occurs (e.g., spikes or significant local field potential changes), event-driven ADCs can drastically reduce power consumption and data throughput while preserving critical neural information. This approach aligns naturally with high-density neural recording systems, where per-channel power budgets are extremely limited.

This internship proposes the design and evaluation of an event-driven ADC architecture optimized for high-density neural interfaces, targeting ultra-low power operation, scalability, and compatibility with advanced neural probes.