Dorsal Column Nuclei Neural Signal Features Permit Robust Machine-Learning of Natural Tactile- and Proprioception-Dominated Stimuli

Abstract

Neural prostheses enable users to effect movement through a variety of actuators by translating brain signals into movement control signals. However, to achieve more natural limb movements from these devices, the restoration of somatosensory feedback is required. We used feature-learnability, a machine-learning approach, to assess signal features for their capacity to enhance decoding performance of neural signals evoked by natural tactile and proprioceptive somatosensory stimuli, recorded from the surface of the dorsal column nuclei (DCN) in urethane-anesthetized rats. The highest performing individual feature, spike amplitude, classified somatosensory DCN signals with 70% accuracy. The highest accuracy achieved was 87% using 13 features that were extracted from both high and low-frequency (LF) bands of DCN signals. In general, high-frequency (HF) features contained the most information about peripheral somatosensory events, but when features were acquired from short time-windows, classification accuracy was significantly improved by adding LF features to the feature set. We found that proprioception-dominated stimuli generalize across animals better than tactile-dominated stimuli, and we demonstrate how information that signal features contribute to neural decoding changes over the time-course of dynamic somatosensory events. These findings may inform the biomimetic design of artificial stimuli that can activate the DCN to substitute somatosensory feedback. Although, we investigated somatosensory structures, the feature set we investigated may also prove useful for decoding other (e.g., motor) neural signals.

Transformation of neural coding for vibrotactile stimuli along the ascending somatosensory pathway

“…Indeed, mouse performance was comparable to that previously demonstrated by a macaque in a similar microstimulation task, only when the electrode quality was lower (Callier et al 2020). This suggests higher quality electrodes that can be chronically implanted in the brainstem, are necessary to investigate the feasibility of a brainstem neural prosthetic device (Loutit and Potas 2020a). Nevertheless, Manipulation of temporal patterns of electrical pulses through microstimulation is readily achievable, and does not require characterization of detailed response properties, nor targeting of individual neurons (Stieger et al 2022).…”

“…Precise temporal coding in the somatosensory brainstem, at both the cellular and population level, might provide a significant advantage for potential brain machine interfaces (Loutit and Potas 2020a). Our behavioral experiments show, for the first time, that mice could discriminate brainstem electrical microstimulation frequencies.…”

“…Finally, we hope our findings will help to understand how feature selectivity emerges in sensory systems, as a transformation from temporal to rate coding is also common in the auditory and visual systems (Gao et al 2016; Reinagel and Reid 2000). Moreover, detailed knowledge of the neural coding could assist the development of better neuroprosthetics in the ascending somatosensory pathway, and highlight the brainstem as a novel target for somatosensory microstimulation (Vickery et al 2020; Loutit and Potas 2020a, 2020b).…”

Peripheral direct current reduces naturally evoked nociceptive activity at the spinal cord in rodent models of pain

“…A wooden rod was used to firmly press the stimulation site at a rate of ∼1 Hz. Sets of 50 stimulations were recorded alongside analogue voltage changes at the piezoelectric sensor [39].…”

Peripheral direct current suppresses physiologically evoked nociceptive responses at the spinal cord in rodent models of pain

“…A wooden rod was used to firmly press the stimulation site at a rate of ~1 Hz. 5 sets of 10 repetitions were recorded alongside analogue voltage changes at the piezoelectric sensor (46).…”