Abstract:In the application of brain-computer interface (BCI), being able to accurately decode brain signals is a critical task. For the multi-class classification task of brain signal ECoG, how to improve the classification accuracy is one of the current research hotspots. ECoG acquisition uses a high-density electrode array and a high sampling frequency, which makes ECoG data have a certain high similarity and data redundancy in the temporal domain, and also unique spatial pattern in spatial domain. How to effectively extract features is both exciting and challenging. Previous work found that visual-related ECoG can carry visual information via frequency and spatial domain. Based on this finding, we focused on using deep learning to design frequency and spatial feature extraction modules, and proposed a Bi-Band ECoGNet model based on deep learning. The main contributions of this paper are: 1) The Bi-BCWT (Bi-Band Channel-Wise Transform) neural network module is designed to replace the time-consume method MST, this module greatly improves the model calculation and data storage efficiency, and effectively increases the training speed; 2) The Bi-BCWT module can effectively take into account the information both in low-frequency and high-frequency domain, which is more conducive to ECoG multi-classification tasks; 3) ECoG is acquired using 2D electrode array, the newly designed 2D Spatial-Temporal feature encoder can extract the 2D spatial feature better. Experiments have shown that the unique 2D spatial data structure can effectively improve classification accuracy; 3) Compared with previous work, the Bi-Band ECoGNet model is smaller and has higher performance, with an accuracy increase of 1.24%, and the model training speed is increased by 6 times, which is more suitable for BCI applications.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to improve the decoding accuracy of electrocorticogram (ECoG) signals in multi - classification tasks. Specifically, the paper conducts research on spatio - temporal feature extraction of ECoG signals, aiming to design a more efficient and accurate deep - learning model to handle the classification tasks of ECoG signals. ### Problem Background ECoG (Electrocorticography) is an invasive brain - computer interface technology that records electroencephalogram signals by directly placing a high - density electrode array on the cerebral cortex. Compared with traditional EEG, ECoG has a higher signal - to - noise ratio and spatial resolution, and thus is widely used in fields such as neurorehabilitation and epilepsy detection. However, since ECoG data has high similarity and redundancy in the time domain and a unique two - dimensional structure in the space domain, how to effectively extract these features has become a challenge. ### Limitations of Existing Methods 1. **Traditional Methods**: Traditional methods usually require manual design of feature extractors, relying on the understanding of the statistical information of electroencephalogram signals. Although these methods perform well on small - data sets, for large - data sets or cases with high similarity between categories, manually - designed features are often insufficient to provide sufficient classification capabilities. 2. **Deep - Learning Methods**: Existing deep - learning methods such as LSTM and CNN mainly focus on the extraction of time - domain features, ignoring the spatial information of electrodes. In addition, some methods that combine traditional methods (such as MST) and deep - learning, although improving performance, have problems such as complex parameter adjustment and slow calculation speed. ### Main Contributions of the Paper To solve the above problems, the paper proposes the Bi - Band ECoGNet model, and its main innovation points include: 1. **Bi - BCWT Module**: - A new neural network module Bi - BCWT (Bi - Band Channel - Wise Transform) is designed to replace the traditional MST method. This module effectively extracts low - frequency and high - frequency information through convolution kernels of different lengths (512 and 32), significantly improving the computational efficiency and training speed of the model. - The Bi - BCWT module can map ECoG signals to a three - dimensional feature space, thereby better capturing information in the frequency domain and the space domain. 2. **2D Spatio - Temporal Feature Encoder**: - In view of the two - dimensional electrode array structure of ECoG signals, a 2D spatio - temporal feature encoder is designed. This encoder can more effectively extract spatial features, further improving the classification accuracy. - Through experimental verification, the 2D spatio - temporal feature encoder is superior to the traditional 1D convolution kernel in terms of extracting spatial information. 3. **Model Lightweight and Efficiency**: - Compared with existing methods, the Bi - Band ECoGNet model is not only smaller in size (with a reduced number of parameters), but also has a significant improvement in performance. Experimental results show that the classification accuracy of the model is increased by 1.24% and the training speed is increased by 6 times. ### Summary By introducing the Bi - Band ECoGNet model, the paper solves the problem of insufficient feature extraction of ECoG signals in multi - classification tasks, especially achieving a breakthrough in the effective utilization of information in the frequency domain and the space domain. This makes the model more practical and efficient in brain - computer interface applications.

Bi-Band ECoGNet for ECoG Decoding on Classification Task

EEG Recognition Based on Multiple Types of Information by Using Wavelet Packet Transform and Neural Networks

Simultaneously Optimizing Spatial Spectral Features Based on Mutual Information for Eeg Classification

EEG-DBNet: A Dual-Branch Network for Temporal-Spectral Decoding in Motor-Imagery Brain-Computer Interfaces

A Simple Channel Compression Method for Brain Signal Decoding on Classification Task

EEGNet: A Compact Convolutional Network for EEG-based Brain-Computer Interfaces

Deep Learning for Micro-Electrocorticographic (µecog) Data

TACNet: Task-aware Electroencephalogram Classification for Brain-Computer Interface Through A Novel Temporal Attention Convolutional Network.

Advanced TSGL-EEGNet for Motor Imagery EEG-Based Brain-Computer Interfaces

EEGNet: a compact convolutional neural network for EEG-based brain–computer interfaces

A Method for the Spatial Interpolation of EEG Signals Based on the Bidirectional Long Short-Term Memory Network

IFBCLNet: Spatio-temporal frequency feature extraction-based MI-EEG classification convolutional network

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

Multi-Scale Neural Network for EEG Representation Learning in BCI

Explainable MST-ECoGNet Decode Visual Information from ECoG Signal

EEGGAN-Net: enhancing EEG signal classification through data augmentation

Two Heads are Better than One: A Bio-inspired Method for Improving Classification on EEG-ET Data

EEG-DCNet: A Fast and Accurate MI-EEG Dilated CNN Classification Method

Binary Convolutional Neural Network for Brain Computer Interfaces

High-Order Temporal Convolutional Network for Improving Classification Performance of SSVEP-EEG

An intertwined neural network model for EEG classification in brain-computer interfaces