EEG and Brain-Computer Interfaces

This paper introduces Hidden Semi-Markov Models (HSMM) to better model brain state durations in EEG and MEG data compared to traditional HMMs.

The paper evaluates whether optimizing wireless EEG sensor network topology can mitigate miniaturization effects in auditory attention detection.

This study decodes mental states from scalp EEG under natural conditions using SVM classifiers and shows that small sets of de-noised data perform similarly to large sets of noisy data.

The paper introduces a human-machine coupled system for efficient sleep spindle detection through iterative revision, achieving consistent results regardless of initial labeling.

This study evaluates how head volume conduction affects directional connectivity estimates from EEG source reconstruction using simulations and multiple connectivity measures.

This study demonstrates that residual sensory signals from a paralyzed hand can be decoded from the motor cortex and used to enhance sensorimotor function via a closed-loop brain-computer interface.

This study uses offline simulations to optimize parameters for decoded neurofeedback experiments, showing they align with real performance and explain individual differences.

This paper critiques two studies claiming to enable communication in complete locked-in state patients, showing methodological flaws and irreproducibility of their results.

This study uses real-time neurofeedback to adjust arousal levels and improve performance in a demanding sensory-motor task.

This study quantifies how fiducial location errors in EEG affect dipole source estimation and recommends accuracy thresholds to minimize uncertainty.

The paper develops a biophysical framework to interpret EEG data across multiple spatial scales, categorizing sources as genuine, equivalent, or representative.

This paper develops an SVM-based algorithm to automatically detect and score spike-wave discharges (SWDs) in mice, showing it correlates with human confidence and behavioral state changes.

This paper introduces an automated method to detect EEG electrodes in MR images during simultaneous EEG/fMRI, achieving 94% detection accuracy and 3.1 mm average position error.

This paper detects driver's movement intention in response to traffic lights using EEG-based decoding, showing anticipatory slow cortical potentials and single-trial classification performance.

This paper introduces the frequency-adaptive broadband (FAB) beamformer for EEG/MEG data, which outperforms existing methods in reconstructing broadband source signals.

This paper develops an automatic EEG spike detection method for epilepsy studies, showing it can detect more spikes, identify additional brain regions, and speed up analysis compared to manual methods.

The paper compares two EEG filtering methods for extracting the μ-rhythm and assesses their impact on corticospinal excitability measured via TMS.

This paper develops a computationally efficient seizure detection algorithm for epilepsy management using EEG data from a mouse model, achieving 98.51% accuracy with reduced data requirements.

This paper uses EEG network analysis to detect cognitive impairment in preschool children with epilepsy, achieving 85% sensitivity and 84% specificity.

This paper introduces a coherence-based spectro-spatial filter to predict stimulus features from electrocorticographic recordings, showing improved performance over traditional methods.

This study examines if P300 brain dynamics observed in lab settings are consistent in real-world driving scenarios.

This paper introduces an automated method to classify and quantify interictal spikes in EEG data from mice, using probabilistic modeling to predict epileptogenic treatment.

This paper introduces a method to model brain responses to arbitrary visual stimuli using EEG, enabling flexible and high-speed brain-computer interfaces with increased information transfer rates.

This paper introduces a geometric classification method for brain network dynamics that achieves high temporal resolution and anatomical inference using conic derivative discriminants.

This paper proposes an algorithm to reduce neurofeedback latency in EEG by optimizing FIR filter weights and analytic signal reconstruction, enabling predictive feedback with maintained accuracy.

This paper shows that reward modulates neural activity in M1 during BMI control and demonstrates using this modulation to improve BMI decoder performance.

The paper uses EEG data and Schulte tables to classify human personality into three groups based on spatio-temporal and time-frequency brain activity patterns.

This paper develops a noninvasive EEG-based BCI system for robotic control, showing high accuracy in motor-imagery classification and demonstrating feasibility for assistive robotic applications.

This study uses EEG and machine learning to predict consumer preferences and commercial success better than traditional questionnaires.

SEREEGA generates simulated EEG data with known ground truth to evaluate signal processing and machine learning methods.