Searcharxiv⌕ Search

arXiv subjects

Johnson Song

Publications and source records attributed to Johnson Song.

1 recordsLinked to original sources

An Open-Architecture Dual-Core STM32-Based Neural Interface with Low-Latency On-Device AI for Neuromodulation Applications

We present an open-architecture bidirectional neural interface for closed-loop neuromodulation built around a dual-core STM32H747 microcontroller and commercial Intan recording and stimulation front-ends. The system supports up to 128 recording channels at 20 kSps and 64 stimulation channels while enabling low-latency on-device AI inference through a dual-core processing architecture and a two-stream memory hierarchy. To validate the machine learning pipeline, we demonstrate two neural decoding models, a compact multilayer perceptron executing entirely from on-chip memory and a larger temporal neural decoder utilizing external memory staging. Both models successfully decode fingertip velocity from neural spiking activity. As a proof of concept for closed-loop operation, we implement spike-triggered stimulation and achieve an end-to-end latency of $189 \pm 14\,μ$s, representing an improvement of three orders of magnitude over a host-PC-mediated closed-loop pathway. These results demonstrate that resource-constrained embedded platforms can support high-channel-count neural interfacing, flexible on-device AI, and sub-millisecond closed-loop neuromodulation.

eess.SP↗