Abstract:The technology of invasive Brain-Computer Interfaces (BCIs) has been developed in last decades, for its possibility to restore motor function of the disabilities. The main task of BCI system is to translate the cortical neural activities into commands of direct brain-controlled prosthetic devices. To study how cortical signals simultaneously recorded from primary motor cortex (M1) neurons were used for external devices control, invasive brain-machine interfaces in rat were investigated. In these experiments, rats were trained to control a relay to obtain water by pressing a lever over a pressure threshold. Microwire array was implanted in rat's motor cortex to record neural activities, and pressure was recorded by a pressure sensor. After spike detecting and sorting, totally 22-58 neurons were found in all 15 channels per rat (except the reference electrode in the array). To compute the firing rate of individual unit, the numbers of spikes in a time bin (Δt =100ms) were counted. Meanwhile, the pressure signal was also computed into bin size (Δt =100ms), by averaging pressure value in the bin period. After that some decoders were designed to make mapping between neural activities and pressure, such as Optimum Linear Estimation(OLE), Kalman filter (KF), and Kernel-Based Membership combined with Curve Fitting algorithm (KMCF). These decoders can be easily learned using a few minutes of training data and provides real-time estimates of hand position every 100ms given the firing rates of neurons in motor cortex. The performances of these decoders were evaluated by Correlation Coefficient (CC) between real and predictive pressure value. The results showed that the KMCF decoder performed best in these experiment, which had higher CC than others (CC=0.94). The higher CC indicated that the activity of motor cortex (M1) neurons can be used for detection of the corresponding movement states and estimation of continuous kinematic parameters. The rats could use - - their neural activities to directly control the relay and successfully get rewards after about one week of training. In further, with the development of the motor-related BCIs technology, BCIs will be possibly used to motor function restoration of paralytics and greatly rehabilitate their capacity of life. Furthermore, the experiment results provide insights into the nature of the neural coding of movement.

Motor Intention Decoding from the Upper Limb by Graph Convolutional Network Based on Functional Connectivity

Fast and Accurate Decoding of Finger Movements from ECoG Through Riemannian Features and Modern Machine Learning Techniques

GCNs-Net: A Graph Convolutional Neural Network Approach for Decoding Time-resolved EEG Motor Imagery Signals

Graph neural network based on brain inspired forward-forward mechanism for motor imagery classification in brain-computer interfaces

A coarse-to-fine adaptive spatial–temporal graph convolution network with residuals for motor imagery decoding from the same limb

Rapid Decoding of Hand Gestures in Electrocorticography Using Recurrent Neural Networks.

Motor Imagery Classification Based on Deep Convolutional Neural Network and Its Application in Exoskeleton Controlled by EEG

Real-time Decoding Algorithm in Brain Machine Interfaces

Design a Novel BCI for Neurorehabilitation Using Concurrent LFP and EEG Features: A Case Study

Decoding the Nonstationary Neural Activity in Motor Cortex for Brain Machine Interfaces

A brain topography graph embedded convolutional neural network for EEG-based motor imagery classification

An optimized EEGNet decoder for decoding motor image of four class fingers flexion

Deep Convolutional Neural Network for EEG-Based Motor Decoding

Decoding Multi-Class EEG Signals of Hand Movement Using Multivariate Empirical Mode Decomposition and Convolutional Neural Network

A Spatial Filter Temporal Graph Convolutional Network for decoding motor imagery EEG signals

Efficient Multi-View Graph Convolutional Network with Self-Attention for Multi-Class Motor Imagery Decoding

Coherence based graph convolution network for motor imagery-induced EEG after spinal cord injury

A Multifrequency Brain Network-Based Deep Learning Framework for Motor Imagery Decoding

A Novel Motor Imagery EEG Decoding Method Based on Feature Separation

Recognition of Upper Limb Motion Intention of EEG Signal Based on Convolutional Neural Network

Decoding Multi-Class Motor Imagery from Unilateral Limbs Using EEG Signals