BrainDistill: Implantable Motor Decoding with Task-Specific Knowledge Distillation

TL;DR

BrainDistill introduces a power-efficient, implantable neural decoder with task-specific knowledge distillation and quantization, achieving high decoding accuracy suitable for implantable brain-computer interfaces.

Contribution

It presents a novel implantable neural decoder framework with task-specific knowledge distillation and quantization-aware training for efficient brain decoding.

Findings

IND outperforms prior neural decoders on motor tasks

TSKD improves few-shot calibration performance

Quantized IND enables low-power deployment with minimal accuracy loss

Abstract

Transformer-based neural decoders with large parameter counts, pre-trained on large-scale datasets, have recently outperformed classical machine learning models and small neural networks on brain-computer interface (BCI) tasks. However, their large parameter counts and high computational demands hinder deployment in power-constrained implantable systems. To address this challenge, we introduce BrainDistill, a novel implantable motor decoding pipeline that integrates an implantable neural decoder (IND) with a task-specific knowledge distillation (TSKD) framework. Unlike standard feature distillation methods that attempt to preserve teacher representations in full, TSKD explicitly prioritizes features critical for decoding through supervised projection. Across multiple neural datasets, IND consistently outperforms prior neural decoders on motor decoding tasks, while its TSKD-distilled variant further surpasses alternative distillation methods in few-shot calibration settings. Finally, we present a quantization-aware training scheme that enables integer-only inference with activation clipping ranges learned during training. The quantized IND enables deployment under the strict power constraints of implantable BCIs with minimal performance loss.