# Real-time Closed Loop Neural Decoding on a Neuromorphic Chip

## TL;DR

This paper demonstrates a real-time neuromorphic chip for neural decoding in brain-machine interfaces, achieving high success rates and speeds comparable to manual control in non-human primates.

## Contribution

It introduces the first real-time closed-loop neuromorphic decoder chip for intra-cortical brain-machine interfaces in a non-human primate model.

## Key findings

- Neural control success rate of ~96% of joystick control
- Mean target reach speed of ~85% of joystick control
- First demonstration of real-time neuromorphic decoding in iBMI

## Abstract

This paper presents for the first time a real-time closed loop neuromorphic decoder chip-driven intra-cortical brain machine interface (iBMI) in a non-human primate (NHP) based experimental setup. Decoded results show trial success rates and mean times to target comparable to those obtained by hand-controlled joystick. Neural control trial success rates of approximately 96% of those obtained by hand-controlled joystick have been demonstrated. Also, neural control has shown mean target reach speeds of approximately 85% of those obtained by hand-controlled joystick . These results pave the way for fast and accurate, fully implantable neuromorphic neural decoders in iBMIs.