Online decoding of rat self-paced locomotion speed from EEG using recurrent neural networks
arXiv:2602.18637v1 Announce Type: new Abstract: $\textit{Objective.}$ Accurate neural decoding of locomotion holds promise for advancing rehabilitation, prosthetic control, and understanding neural correlates of action. Recent studies have demonstrated decoding of locomotion kinematics across species on motorized treadmills. However, efforts to decode locomotion speed in more natural contexts$-$where pace is self-selected rather than externally imposed$-$are scarce, generally achieve only modest accuracy, and require intracranial implants. Here, we aim to decode self-paced locomotion speed non-invasively and continuously using cortex-wide EEG recordings from rats. $\textit{Approach.}$ We introduce an asynchronous brain$-$computer interface (BCI) that processes a stream of 32-electrode skull-surface EEG (0.01$-$45 Hz) to decode instantaneous speed from a non-motorized treadmill during self-paced locomotion in head-fixed rats. Using recurrent neural networks and a dataset of over 133 h
arXiv:2602.18637v1 Announce Type: new Abstract: $\textit{Objective.}$ Accurate neural decoding of locomotion holds promise for advancing rehabilitation, prosthetic control, and understanding neural correlates of action. Recent studies have demonstrated decoding of locomotion kinematics across species on motorized treadmills. However, efforts to decode locomotion speed in more natural contexts$-$where pace is self-selected rather than externally imposed$-$are scarce, generally achieve only modest accuracy, and require intracranial implants. Here, we aim to decode self-paced locomotion speed non-invasively and continuously using cortex-wide EEG recordings from rats. $\textit{Approach.}$ We introduce an asynchronous brain$-$computer interface (BCI) that processes a stream of 32-electrode skull-surface EEG (0.01$-$45 Hz) to decode instantaneous speed from a non-motorized treadmill during self-paced locomotion in head-fixed rats. Using recurrent neural networks and a dataset of over 133 h of recordings, we trained decoders to map ongoing EEG activity to treadmill speed. $\textit{Main results.}$ Our decoding achieves a correlation of 0.88 ($R^2$ = 0.78) for speed, primarily driven by visual cortex electrodes and low-frequency ($< 8$ Hz) oscillations. Moreover, pre-training on a single session permitted decoding on other sessions from the same rat, suggesting uniform neural signatures that generalize across sessions but fail to transfer across animals. Finally, we found that cortical states not only carry information about current speed, but also about future and past dynamics, extending up to 1000 ms. $\textit{Significance.}$ These findings demonstrate that self-paced locomotion speed can be decoded accurately and continuously from non-invasive, cortex-wide EEG. Our approach provides a framework for developing high-performing, non-invasive BCI systems and contributes to understanding distributed neural representations of action dynamics.
Executive Summary
This study presents a groundbreaking approach to non-invasively and continuously decoding self-paced locomotion speed from cortex-wide EEG recordings in rats. Using recurrent neural networks and a dataset of over 133 hours of recordings, the researchers achieved a correlation of 0.88 (R^2 = 0.78) for speed, primarily driven by visual cortex electrodes and low-frequency oscillations. The findings demonstrate the potential for developing high-performing, non-invasive brain-computer interface (BCI) systems and contribute to understanding distributed neural representations of action dynamics. The study's significance lies in its ability to decode self-paced locomotion speed, a complex neural activity, and its potential applications in rehabilitation, prosthetic control, and understanding neural correlates of action.
Key Points
- ▸ The study presents a novel approach to decoding self-paced locomotion speed from EEG recordings in rats.
- ▸ The researchers achieved high accuracy in decoding speed using recurrent neural networks and cortex-wide EEG recordings.
- ▸ The findings demonstrate the potential for developing non-invasive BCI systems and contribute to understanding neural representations of action dynamics.
Merits
Strength in Decoding Accuracy
The study achieved high decoding accuracy, with a correlation of 0.88 (R^2 = 0.78) for speed, demonstrating the effectiveness of the proposed approach.
Non-invasive BCI Development
The study provides a framework for developing high-performing, non-invasive BCI systems, which has significant implications for rehabilitation, prosthetic control, and neural research.
Demerits
Limited Generalizability
The study's findings were restricted to a single species (rats) and a specific experimental setup, limiting the generalizability of the results to other species and contexts.
Data Requirements
The study required a large dataset of over 133 hours of recordings, which may not be feasible or practical in all experimental settings.
Expert Commentary
This study represents a significant advancement in the field of neural decoding and brain-computer interface development. The researchers' ability to decode self-paced locomotion speed from cortex-wide EEG recordings in rats demonstrates the potential for non-invasive BCI systems. However, the study's limitations, including the restricted generalizability of the results and the data requirements, highlight the need for further research in this area. The implications of this study are significant, with potential applications in rehabilitation, prosthetic control, and neural research. As the field continues to evolve, it is essential to develop more sophisticated and widely applicable non-invasive BCI systems.
Recommendations
- ✓ Future studies should aim to replicate the findings in other species and experimental settings to increase the generalizability of the results.
- ✓ Researchers should explore the use of different neural decoding techniques and algorithms to improve the accuracy and efficiency of non-invasive BCI systems.
