EEG and Brain-Computer Interfaces

This study applies explainable deep learning models to raw EEG data to diagnose schizophrenia, identifying key spectral and spatial features that align with prior research.

The paper develops and validates a new EEG source localization algorithm using BEM-FMM, reciprocity principle, and TES forward model matrix, achieving high-resolution source reconstruction with low computational cost.

This paper develops and validates a standardized EEG preprocessing and entropy estimation pipeline called APPLESEED to reliably estimate scale-wise entropy for developmental studies.

This review evaluates neurofeedback's impact on task performance in healthy participants, finding limited statistically significant effects and variability in study design.

This study compares the impact of using template versus individualized brain anatomies and EEG electrode positions on the localization accuracy of the primary auditory cortex using dSPM and sLORETA inverse solutions.

This paper introduces a rapid invisible frequency-tagging technique for BCIs, enabling real-time control with high accuracy and reduced visual fatigue.

The paper introduces a large-scale EEG database and competition to study EEG commonality and variability across subjects, sessions, and tasks, aiming to improve machine learning algorithms for EEG-based biometrics.

The paper analyzes sleep EEG spectra using the FOOOF method to separate aperiodic and oscillatory components, finding sleep stage-specific spectral features.

Open-source MATLAB pipeline (EEGLAB, fieldtrip) that synchronizes mobile EEG with motion and eye-tracking data, extracts gait, blink and heartbeat events, and clusters independent components per region of interest.

This paper identifies three critical steps to improve the interpretability and clinical adoption of EEG-SVM studies by addressing issues in technical reporting.

This paper develops a lightweight autoencoder model for real-time meditation state assessment using EEG data, showing that multivariate EEG features have substantial overlap and temporal dynamics not captured by simple one-class algorithms.

The paper evaluates EEG/MEG source-space network analysis techniques using a human brain computational model and identifies optimal methods for inverse modeling, connectivity measures, and dimensionality reduction.

This study tests if EEG-neurofeedback can help elderly people modulate brain activity, finding significant self-modulation in trained subjects but no cognitive improvement.

This paper introduces a real-time EEG sleep spindle detection algorithm that enables automated stimulation during sleep spindles with high sensitivity and precision.

This study improves EEG-based classification of imagined images and image reconstruction using a modified neural network with an attention module and GAN.

This paper collects and provides EEG data from students exposed to audio-visual content to analyze engagement and attention during learning.

The paper shows that motor representations remain stable after paralysis, enabling brain-computer interfaces to restore motor functions using preserved neural patterns.

The paper introduces DEED, a dataset of EEG recordings from sleep to assess dream emotions, and validates it using machine learning.

The paper evaluates the impact of movement intensity and automatic sample rejection on ICA decomposition quality across multiple EEG datasets, finding that moderate cleaning improves decomposition regardless of movement intensity.

Meggie provides an open-source, Python-based GUI for M/EEG analysis, enabling end-to-end processing with multi-subject support and plugin extendibility.

The paper introduces Bayesian Gaussian random walk models to estimate ERP voltage over time and electrode correlations using real P3b data.

This paper develops cross-modal, self-supervised neural decoders using multiple data streams to achieve high decoding accuracy without labeled data.

This paper benchmarks machine learning approaches for brain age prediction using M/EEG signals across four international cohorts, finding that spatially aware models achieve the highest performance.

The paper describes an immersive paradigm for studying natural social behavior and its neural correlates in epilepsy patients using synchronized behavioral and neurophysiological data.

This paper identifies a frequency doubling artifact in decoding accuracy metrics from neurophysiological data and proposes a filtering approach and alternative decoding method to mitigate it.

The paper proposes a multi-domain EEG analysis method using ResNet50 to detect acute cocaine and methamphetamine abuse in mice, achieving an F1-score of 95.10%.

The paper introduces HAPPE+ER, an automated ERP processing pipeline for EEG data, validated with adult and developmental data to improve standardization and efficiency.

This paper presents a transformer model to predict cognitive impairment from sleep EEG data, achieving 70.22% accuracy, outperforming traditional methods.

This study maps cortical engagement during motor imagery and cognitive tasks using MEG beta rhythm changes, showing task-specific patterns and high classification accuracy.

This paper introduces a model visualization approach to analyze CNN filters for EEG waveforms and spectral features, demonstrating its use in sleep stage classification.