Architectural Exploration of Hybrid Neural Decoders for Neuromorphic Implantable BMI

TL;DR

This paper introduces a novel, efficient decoding pipeline for neuromorphic implantable brain-machine interfaces that significantly reduces data processing and computational demands while maintaining high decoding accuracy.

Contribution

It presents a tunable event filter and a new decoding pipeline that eliminates traditional signal processing steps, improving efficiency for neuromorphic BMI systems.

Findings

Event reduction of 192X to 554X with EvFilter

Decoding performance up to R^2=0.73

SNN-Decoder reduces computation and memory by 5-23X

Abstract

This work presents an efficient decoding pipeline for neuromorphic implantable brain-machine interfaces (Neu-iBMI), leveraging sparse neural event data from an event-based neural sensing scheme. We introduce a tunable event filter (EvFilter), which also functions as a spike detector (EvFilter-SPD), significantly reducing the number of events processed for decoding by 192X and 554X, respectively. The proposed pipeline achieves high decoding performance, up to R^2=0.73, with ANN- and SNN-based decoders, eliminating the need for signal recovery, spike detection, or sorting, commonly performed in conventional iBMI systems. The SNN-Decoder reduces computations and memory required by 5-23X compared to NN-, and LSTM-Decoders, while the ST-NN-Decoder delivers similar performance to an LSTM-Decoder requiring 2.5X fewer resources. This streamlined approach significantly reduces computational and memory demands, making it ideal for low-power, on-implant, or wearable iBMIs.