Abstract:Electroencephalography (EEG) serves as a practical auxiliary tool deployed to diagnose diverse brain-related disorders owing to its exceptional temporal resolution, non-invasive characteristics, and cost-effectiveness. In recent years, with the advancement of machine learning, automated EEG pathology diagnostics methods have flourished. However, most existing methods usually neglect the crucial spatial correlations in multi-channel EEG signals and the potential complementary information among different domain features, both of which are keys to improving discrimination performance. In addition, latent redundant and irrelevant features may cause overfitting, increased model complexity, and other issues. In response, we propose a novel feature-based framework designed to improve the diagnostic accuracy of multi-channel EEG pathology. This framework first applies a multi-resolution decomposition technique and a statistical feature extractor to construct a salient time-frequency feature space. Then, spatial distribution information is channel-wise extracted from this space to fuse with time-frequency features, thereby leveraging their complementarity to the fullest extent. Furthermore, to eliminate the redundancy and irrelevancy, a two-step dimension reduction strategy, including a lightweight multi-view time-frequency feature aggregation and a non-parametric statistical significance analysis, is devised to pick out the features with stronger discriminative ability. Comprehensive examinations of the Temple University Hospital Abnormal EEG Corpus V. 2.0.0 demonstrate that our proposal outperforms state-of-the-art methods, highlighting its significant potential in clinically automated EEG abnormality detection.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the accuracy of multi - channel electroencephalogram (EEG) pathological diagnosis. Specifically, the existing automatic EEG pathological diagnosis methods usually ignore the important spatial correlations in multi - channel EEG signals and the potential complementary information between features in different domains, which limits the improvement of discriminative performance. In addition, redundant and irrelevant features may lead to problems such as over - fitting and increased model complexity. To solve these problems, this paper proposes a new feature - based framework, aiming to improve the accuracy of multi - channel EEG pathological diagnosis by fusing multi - domain features. ### Summary of main problems: 1. **Ignoring spatial correlations and complementary information**: Existing methods usually overlook the spatial correlations in multi - channel EEG signals and the complementary information between features in different domains (such as the time domain, frequency domain and spatial domain). 2. **Impact of redundant and irrelevant features**: Redundant and irrelevant features may lead to problems such as over - fitting and increased model complexity, thus affecting the classification performance. ### Solutions: 1. **Multi - resolution decomposition and statistical feature extraction**: First, apply multi - resolution decomposition techniques (such as discrete wavelet transform DWT) and statistical feature extractors to construct a significant time - frequency feature space. 2. **Spatial distribution information extraction**: Extract the spatial distribution information channel by channel from the time - frequency feature space and fuse it with the time - frequency features to make full use of their complementarity. 3. **Two - step dimension reduction strategy**: In order to eliminate redundant and irrelevant features, a two - step dimension reduction strategy is designed, including lightweight multi - view time - frequency feature aggregation and non - parametric statistical significance analysis, to select features with stronger discrimination ability. 4. **Classifier application**: Input the optimized feature set into several different ensemble learning classifiers to classify EEG as normal or abnormal. Through these methods, the new framework proposed in the paper improves the accuracy and reliability of EEG pathological diagnosis in multiple aspects.

Automatic diagnostics of electroencephalography pathology based on multi-domain feature fusion

Automatic Epileptic Seizures Joint Detection Algorithm Based on Improved Multi-Domain Feature of cEEG and Spike Feature of aEEG.

Epileptic State Classification by Fusing Hand-Crafted and Deep Learning EEG Features

Electroencephalography (EEG) based epilepsy diagnosis via multiple feature space fusion using shared hidden space-driven multi-view learning

A multi-context learning approach for EEG epileptic seizure detection

A deep learning framework with multi-perspective fusion for interictal epileptiform discharges detection in scalp electroencephalogram

Multi-modal feature fusion with multi-head self-attention for epileptic EEG signals

A Multi-View Deep Learning Framework for EEG Seizure Detection

Multi-feature fusion learning for Alzheimer's disease prediction using EEG signals in resting state

Early seizure detection in childhood focal epilepsy with electroencephalogram feature fusion on deep autoencoder learning and channel correlations

A unified multi-level spectral–temporal feature learning framework for patient-specific seizure onset detection in EEG signals

Unified Convolutional Sparse Transformer for Disease Diagnosis, Monitoring, Drug Development, and Therapeutic Effect Prediction from EEG Raw Data

Transaldolase B: trapping of Schiff base intermediate between dihydroxyacetone and epsilon-amino group of active-site lysine residue by borohydride reduction.

Automatic Seizure Detection Using Multi-Input Deep Feature Learning Networks for EEG Signals

EPILEPTIC EEG IDENTIFICATION BASED ON HYBRID FEATURE EXTRACTION

Advanced framework for epilepsy detection through image-based EEG signal analysis

A scheme combining feature fusion and hybrid deep learning models for epileptic seizure detection and prediction

Investigating the Interpretability of Schizophrenia EEG Mechanism through a 3DCNN-based Hidden Layer Features Aggregation Framework

Automated Atrial Fibrillation and Ventricular Fibrillation Recognition Using a Multi-Angle Dual-Channel Fusion Network

Ensemble Fusion Models Using Various Strategies and Machine Learning for EEG Classification

Depression Detection Based on the Temporal-Spatial-frequency Feature Fusion of EEG