Somatosensory responses in a human motor cortex

Ammar Shaikhouni, John P. Donoghue, Leigh R. HochbergJournal of Neurophysiology 2013 · DOI

TL;DR

This study investigates how the human motor cortex (M1) processes somatosensory information in a person with a long-standing brainstem stroke but intact sensory pathways. The findings reveal that M1 retains significant responses to passive joint manipulation, suggesting that these sensory inputs remain crucial for motor control and BCI applications, even years after severe neurological damage.

What Problem Does This Paper Address?

Brain-computer interfaces (BCIs) aim to restore motor function by decoding neural activity from the motor cortex. A key question is how sensory feedback, which is crucial for natural motor control, influences M1 activity and BCI performance, especially in individuals with neurological damage. Previous research has shown that M1 neurons remain active during attempted movements in paralyzed individuals, forming the basis for BCI control. However, the role and organization of somatosensory inputs to M1 in humans, particularly after stroke, and their potential impact on BCI design, remain less understood. This paper addresses this gap by examining somatosensory responses in M1 of a stroke survivor with intact sensory pathways, providing insights into cortical plasticity and its implications for BCI development.

Methods

The study utilized a single participant (S3), a 52-year-old woman with a chronic brainstem stroke resulting in tetraplegia but preserved somatosensory function. A 96-electrode microelectrode array was implanted in her left precentral gyrus (M1 hand/arm area).

The researchers employed two main tasks:

  • Passive Joint Manipulation: An experimenter passively moved the participant's contralateral (right) arm joints (wrist, elbow, shoulder) through their range of motion. The participant was instructed not to attempt any movement. Neural activity was recorded during both the dynamic movement phase and static holding periods.
  • Attempted Movement: The participant was cued to imagine and attempt specific joint movements. Neural activity was recorded during these attempted actions.

Neural data was acquired using a high-density microelectrode array and processed for single-unit isolation. Peristimulus time histograms (PSTHs) were used to analyze unit responses to passive movements, with statistical significance determined by random permutation tests. Decoding analyses, using linear discriminant analysis with leave-one-out cross-validation, were performed to assess the information available in neural populations for distinguishing between passive and attempted movements.

Datasets

The study used data collected from a single participant (S3) over two sessions, approximately one year apart. No external datasets were used for training or evaluation. The data itself, from a unique participant with specific neurological characteristics, can be considered a novel dataset for this type of investigation.

Key Insights and Ideas

  • Preserved Somatosensory Input: Despite a severe brainstem stroke and long-standing paralysis, M1 neurons in this participant retained robust responses to passive manipulation of contralateral limb joints. This suggests that the fundamental somatosensory processing capabilities of M1 are remarkably resilient to such damage.
  • Dominance of Wrist Representation: Similar to findings in non-human primates, wrist joint manipulation elicited the most prevalent responses across the recorded M1 population. This could reflect a topographic bias in the implanted array's location or a genuine emphasis on wrist representation in this individual's M1.
  • Multijoint Receptive Fields: A significant proportion of M1 neurons responded to passive movements of multiple joints, indicating a complex and integrated somatosensory representation rather than strictly unisensory fields. This challenges a simple, localized view of sensory processing in M1.
  • Distinct Passive vs. Active Representations: Crucially, decoding analyses revealed that neural activity patterns during passive movement were significantly different from those during attempted movement. This implies that the brain utilizes distinct neural codes for sensory feedback and motor intention, even when controlling the same limb.
  • Implications for BCIs: The findings suggest that BCIs relying solely on decoding attempted movements might benefit from incorporating somatosensory feedback. The presence of intact sensory pathways in potential BCI users (e.g., stroke survivors) offers an opportunity to leverage this information for more intuitive and effective control.

Key Results

  • High Response Rate: In session 1, 69% of recorded units responded to passive joint motion, and in session 2, 58% responded.
  • Dominant Wrist Input: In both sessions, wrist responses were the most prevalent, with 66% (session 1) and 50% (session 2) of responding units showing wrist modulation.
  • Multijoint Responses: A majority of responding cells (63% in session 1, 68% in session 2) were modulated by more than one joint.
  • Decoding Accuracy: Passive movement decoding achieved high accuracy (85% and 79.8% in sessions 1 and 2, respectively). Attempted movement decoding was also successful (82.5% in session 1, 55% in session 2).
  • Poor Cross-Condition Decoding: Decoders trained on passive movement data could not significantly predict attempted movements (8% accuracy), and vice-versa (17% accuracy). This highlights the distinct neural representations.

Limitations and Future Work

  • Single Participant: The most significant limitation is that the study involved only one participant. While this participant provided a unique opportunity, the findings may not be generalizable to all stroke survivors or individuals with similar neurological conditions.
  • Specific Lesion Type: The participant had a brainstem stroke. The somatosensory responses might differ in individuals with other types of lesions (e.g., cortical strokes).
  • Limited Ipsilateral Exploration: Due to time constraints, the exploration of ipsilateral sensory responses was limited.
  • Spasticity: The participant exhibited spasticity, which could influence neural responses. Future studies with participants lacking significant motor tone (e.g., advanced ALS) are needed.

The authors suggest that future work should involve studying additional participants to confirm these findings and explore the organization of M1 in other patient populations. They also emphasize the importance of understanding the complex interplay between sensory and motor representations for designing more effective BCIs and rehabilitation strategies. The potential for incorporating kinesthetic feedback into BCIs is highlighted as a promising avenue for improving performance.

Summary updated

Abstract

Somatic sensory signals provide a major source of feedback to motor cortex. Changes in somatosensory systems after stroke or injury could profoundly influence brain computer interfaces (BCI) being developed to create new output signals from motor cortex activity patterns. We had the unique opportunity to study the responses of hand/arm area neurons in primary motor cortex to passive joint manipulation in a person with a long-standing brain stem stroke but intact sensory pathways. Neurons responded to passive manipulation of the contralateral shoulder, elbow, or wrist as predicted from prior studies of intact primates. Thus fundamental properties and organization were preserved despite arm/hand paralysis and damage to cortical outputs. The same neurons were engaged by attempted arm actions. These results indicate that intact sensory pathways retain the potential to influence primary motor cortex firing rates years after cortical outputs are interrupted and may contribute to online decoding of motor intentions for BCI applications.
BibTeX
@article{shaikhouni2013somatosensory,
  author = {Ammar Shaikhouni and John P. Donoghue and Leigh R. Hochberg},
  title = {{Somatosensory responses in a human motor cortex}},
  journal = {Journal of Neurophysiology},
  year = {2013},
  doi = {10.1152/jn.00368.2012},
}
Figure