Evaluating the effect of the cutoff frequencies during the pre-processing stage of motor imagery EEG signals classification

The Effect of Processing Techniques on the Classification Accuracy of Brain-Computer Interface Systems

Background/Objectives: Accurately classifying Electroencephalography (EEG) signals is essential for the effective operation of Brain-Computer Interfaces (BCI), which is needed for reliable neurorehabilitation applications. However, many factors in the processing pipeline can influence classification performance. The objective of this study is to assess the effects of different processing steps on classification accuracy in EEG-based BCI systems. Methods: This study explores the impact of various processing techniques and stages, including the FASTER algorithm for artifact rejection (AR), frequency filtering, transfer learning, and cropped training. The Physionet dataset, consisting of four motor imagery classes, was used as input due to its relatively large number of subjects. The raw EEG was tested with EEGNet and Shallow ConvNet. To examine the impact of adding a spatial dimension to the input data, we also used the Multi-branch Conv3D Net and developed two new models, Conv2D Net and Conv3D Net. Results: Our analysis showed that classification accuracy can be affected by many factors at every stage. Applying the AR method, for instance, can either enhance or degrade classification performance, depending on the subject and the specific network architecture. Transfer learning was effective in improving the performance of all networks for both raw and artifact-rejected data. However, the improvement in classification accuracy for artifact-rejected data was less pronounced compared to unfiltered data, resulting in reduced precision. For instance, the best classifier achieved 46.1% accuracy on unfiltered data, which increased to 63.5% with transfer learning. In the filtered case, accuracy rose from 45.5% to only 55.9% when transfer learning was applied. An unexpected outcome regarding frequency filtering was observed: networks demonstrated better classification performance when focusing on lower-frequency components. Higher frequency ranges were more discriminative for EEGNet and Shallow ConvNet, but only when cropped training was applied. Conclusions: The findings of this study highlight the complex interaction between processing techniques and neural network performance, emphasizing the necessity for customized processing approaches tailored to specific subjects and network architectures.

Characterization and classification of kinesthetic motor imagery levels

Objective.Kinesthetic Motor Imagery (KMI) represents a robust brain paradigm intended for electroencephalography (EEG)-based commands in brain-computer interfaces (BCIs). However, ensuring high accuracy in multi-command execution remains challenging, with data from C3 and C4 electrodes reaching up to 92% accuracy. This paper aims to characterize and classify EEG-based KMI of multilevel muscle contraction without relying on primary motor cortex signals.Approach.A new method based on Hurst exponents is introduced to characterize EEG signals of multilevel KMI of muscle contraction from electrodes placed on the premotor, dorsolateral prefrontal, and inferior parietal cortices. EEG signals were recorded during a hand-grip task at four levels of muscle contraction (0%, 10%, 40%, and 70% of the maximal isometric voluntary contraction). The task was executed under two conditions: first, physically, to train subjects in achieving muscle contraction at each level, followed by mental imagery under the KMI paradigm for each contraction level. EMG signals were recorded in both conditions to correlate muscle contraction execution, whether correct or null accurately. Independent component analysis (ICA) maps EEG signals from the sensor to the source space for preprocessing. For characterization, three algorithms based on Hurst exponents were used: the original (HO), using partitions (HRS), and applying semivariogram (HV). Finally, seven classifiers were used: Bayes network (BN), naive Bayes (NB), support vector machine (SVM), random forest (RF), random tree (RT), multilayer perceptron (MP), and k-nearest neighbors (kNN).Main results.A combination of the three Hurst characterization algorithms produced the highest average accuracy of 96.42% from kNN, followed by MP (92.85%), SVM (92.85%), NB (91.07%), RF (91.07%), BN (91.07%), and RT (80.35%). of 96.42% for kNN.Significance.Results show the feasibility of KMI multilevel muscle contraction detection and, thus, the viability of non-binary EEG-based BCI applications without using signals from the motor cortex.

EEG Complexity Analysis of Brain States, Tasks and ASD Risk

Autism spectrum disorder is an increasingly prevalent and debilitating neurodevelopmental condition and an electroencephalogram (EEG) diagnostic challenge. Despite large amounts of electrophysiological research over many decades, an EEG biomarker for autism spectrum disorder (ASD) has not been found. We hypothesized that reductions in complex dynamical system behaviour in the human central nervous system as part of the macroscale neuronal function during cognitive processes might be detectable in whole EEG for higher-risk ASD adults. In three studies, we compared the medians of correlation dimension, largest Lyapunov exponent, Higuchi's fractal dimension, multiscale entropy, multifractal detrended fluctuation analysis and Kolmogorov complexity during resting, cognitive and social skill tasks in 20 EEG channels of 39 adults over a range of ASD risk. We found heterogeneous complexity distribution with clusters of hierarchical sequences pointing to potential cognitive processing differences, but no clear distinction based on ASD risk. We suggest that there is indication of statistically significant differences between complexity measures of brain states and tasks. Though replication of our studies is needed with a larger sample, we believe that our electrophysiological and analytic approach has potential as a biomarker for earlier ASD diagnosis.

Multimodal Neurophysiological Transformer for Emotion Recognition

Understanding neural function often requires multiple modalities of data, including electrophysiogical data, imaging techniques, and demographic surveys. In this paper, we introduce a novel neurophysiological model to tackle major challenges in modeling multimodal data. First, we avoid non-alignment issues between raw signals and extracted, frequency-domain features by addressing the issue of variable sampling rates. Second, we encode modalities through "cross-attention" with other modalities. Lastly, we utilize properties of our parent transformer architecture to model long-range dependencies between segments across modalities and assess intermediary weights to better understand how source signals affect prediction. We apply our Multimodal Neurophysiological Transformer (MNT) to predict valence and arousal in an existing open-source dataset. Experiments on non-aligned multimodal time-series show that our model performs similarly and, in some cases, outperforms existing methods in classification tasks. In addition, qualitative analysis suggests that MNT is able to model neural influences on autonomic activity in predicting arousal. Our architecture has the potential to be fine-tuned to a variety of downstream tasks, including for BCI systems.

A Novel Time-Incremental End-to-End Shared Neural Network with Attention-Based Feature Fusion for Multiclass Motor Imagery Recognition

In the research of motor imagery brain-computer interface (MI-BCI), traditional electroencephalogram (EEG) signal recognition algorithms appear to be inefficient in extracting EEG signal features and improving classification accuracy. In this paper, we discuss a solution to this problem based on a novel step-by-step method of feature extraction and pattern classification for multiclass MI-EEG signals. First, the training data from all subjects is merged and enlarged through autoencoder to meet the need for massive amounts of data while reducing the bad effect on signal recognition because of randomness, instability, and individual variability of EEG data. Second, an end-to-end sharing structure with attention-based time-incremental shallow convolution neural network is proposed. Shallow convolution neural network (SCNN) and bidirectional long short-term memory (BiLSTM) network are used to extract frequency-spatial domain features and time-series features of EEG signals, respectively. Then, the attention model is introduced into the feature fusion layer to dynamically weight these extracted temporal-frequency-spatial domain features, which greatly contributes to the reduction of feature redundancy and the improvement of classification accuracy. At last, validation tests using BCI Competition IV 2a data sets show that classification accuracy and kappa coefficient have reached 82.7 ± 5.57% and 0.78 ± 0.074, which can strongly prove its advantages in improving classification accuracy and reducing individual difference among different subjects from the same network.