Spiking Linear Dynamical Systems on Neuromorphic Hardware for Low-Power Brain-Machine Interfaces

TL;DR

This paper presents methods for implementing linear dynamical systems on neuromorphic hardware using spiking neurons, enabling low-power brain-machine interfaces with validated accuracy and potential applications in neural decoding and robotics.

Contribution

Developed analytical and numerical methods for precise linear computations in spiking neural networks on neuromorphic hardware, including a neuromorphic Kalman filter for neural data decoding.

Findings

Validated the discrepancy between spiking and non-spiking LDS

Demonstrated low-power neural decoding of vocal pitch

Analyzed tradeoffs between accuracy, energy, and computation time

Abstract

Neuromorphic architectures achieve low-power operation by using many simple spiking neurons in lieu of traditional hardware. Here, we develop methods for precise linear computations in spiking neural networks and use these methods to map the evolution of a linear dynamical system (LDS) onto an existing neuromorphic chip: IBM's TrueNorth. We analytically characterize, and numerically validate, the discrepancy between the spiking LDS state sequence and that of its non-spiking counterpart. These analytical results shed light on the multiway tradeoff between time, space, energy, and accuracy in neuromorphic computation. To demonstrate the utility of our work, we implemented a neuromorphic Kalman filter (KF) and used it for offline decoding of human vocal pitch from neural data. The neuromorphic KF could be used for low-power filtering in domains beyond neuroscience, such as navigation or robotics.