Wearable Neural Interfaces: Real-Time Identification of Motor Neuron Discharges in Dynamic Motor Tasks

Abstract

ObjectiveRobustness to non-stationary conditions is essential to develop stable and accurate wearable neural interfaces.ApproachWe propose a novel adaptive electromyography (EMG) decomposition algorithm that builds on blind source separation methods by leveraging the Kullback-Liebler divergence and kurtosis of the signals as metrics for online learning. The proposed approach provides a theoretical framework to tune the adaptation hyperparameters and compensate for non-stationarities in the mixing matrix, such as due to dynamic contractions, and to identify the underlying motor neuron (MN) discharges. The adaptation is performed in real-time (∼22 ms of computational time per 100-ms batches).Main resultsThe proposed adaptation algorithm significantly improved all decomposition performance metrics with respect to the absence of adaptation in a wide range of motion of the wrist (80°). The rate of agreement, sensitivity, and precision were ≥ 90% in ≥ 80% of the cases in both simulated and experimentally recorded data, according to a two-source validation approach.SignificanceThe findings demonstrate the feasibility of accurately decoding MN discharges in real-time during dynamic contractions from wearable systems mounted at the wrist and forearm. Moreover, the study proposes an experimental validation method for EMG decomposition in dynamic tasks.

I-Spin live: An open-source software based on blind-source separation for real-time decoding of motor unit activity in humans

“…This explanation is in line with the lower rate of agreement observed when participants tracked a force target higher than the level of the baseline contraction (Figures 3 & 4). One way to overcome these challenges would be to dynamically update the motor unit filters and the centroids of the ‘spikes’ and ‘noise’ classes (27, 47, 48). While appealing, this approach is also computationally demanding (27, 47, 48).…”

“…One way to overcome these challenges would be to dynamically update the motor unit filters and the centroids of the ‘spikes’ and ‘noise’ classes (27, 47, 48). While appealing, this approach is also computationally demanding (27, 47, 48). We propose to update the motor unit filters and the centroids of the ‘spikes’ and ‘noise’ classes during the resting periods.…”