Search Results (1,774)

The temporal course of neuronal electric activity within brain networks, or neurodynamics, reflects the structural and functional properties of the neuronal populations that generate it. Using intracranial stereo-electroencephalography (sEEG) recordings from the public Montreal Neurological Institute (MNI) atlas, we investigated neurodynamics in the primary motor (M1), somatosensory (S1), and auditory (A1) cortices. We tested whether modifying the Higuchi fractal dimension (HFD) by replacing the Euclidean distance with the Fréchet distance could improve sensitivity to local neurodynamics by incorporating trajectory-based similarities in signal evolution. Using a conservative within-subject approach established in the previous literature, we compared signals recorded from different cortical areas within the same individuals (M1 vs. S1: # of people = 16; M1 vs. A1: # = 9; S1 vs. A1: # = 6). To delve deeper into the new measure’s meaning, it was tested on sequences with known fractal properties, the Brownian motion and the Weierstrass function. Results showed that the newly introduced Fréchet-based HFD (HFDf), similarly to standard HFD, consistently discriminated cortical areas at the intra-subject level, confirming the robustness of fractal dimension as a descriptor of region-specific neurodynamics. Contrary to our hypothesis, HFDf did not provide additional sensitivity across areas and notably, it displayed less evident reduction of values in sleep than awake. While cortical regions may share common governing principles across spatiotemporal scales, these do not necessarily translate into strict similarity in temporal signal morphology. We suggest that these findings support that the free-scale nature of neurodynamics is not a self-similar one. This refinement of quantitative tools for cortical neurodynamic mapping paves the way towards novel tools for neuroimaging-informed neuromodulation strategies. Full article

Decoding mental states from non-invasive neural recordings is central to brain-computer interface research. Multimodal acquisition that combines electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) couples the high temporal resolution of EEG with the spatial specificity of fNIRS, compensating for the individual limitations of each modality. While such hybrid systems achieve strong intra-subject performance, cross-subject generalization remains constrained by inter-individual variability in neural responses. This study introduces the Eigenvalue-Guided Positional Encoding Classification Transformer (EPECT), an architecture that integrates eigenvalue-aware multi-head self-attention with sinusoidal positional encoding to capture both the spectral structure of the learned feature representations and the temporal ordering of multimodal sequences. Stacked one-dimensional convolutions extract local patterns prior to transformer encoding, and global average pooling aggregates the final representation for classification. EPECT was evaluated on two publicly available EEG-fNIRS datasets covering motor imagery (MI), n-back, discrimination/selection response (DSR), and word generation (WG) paradigms under a cross-subject protocol. The model achieved classification accuracies of 97.3%, 96.3%, 98.1%, and 97.9% on the MI, n-back, DSR, and WG tasks, respectively. Ablation studies quantified the contribution of each architectural component, and integrated gradients analysis revealed structured modality-specific attribution patterns aligned with task-relevant cortical regions. Additional experiments with synthetic cortical perturbations demonstrate the sensitivity of EPECT to subtle activity changes, indicating potential utility for tracking neurorehabilitation outcomes in future clinical applications. Full article

Background: Autism spectrum disorder (ASD) is characterized by widespread aberrations in brain scalp-level synchronization. Phase-amplitude coupling (PAC), which reflects cross-frequency neuronal oscillatory interactions, serves as a crucial metric for assessing functional brain integration. However, the specific patterns of PAC at both intra-region and inter-region scalp levels in young children with ASD, as well as their precise associations with clinical symptoms, remain unclear. Methods: This study enrolled 237 children with ASD aged 3–9 years and 201 age-matched typically developing (TD) children. Resting-state electroencephalography (EEG) data were acquired from all participants. The analysis systematically examined low-frequency oscillation phase (δ, θ, α) modulation of high-frequency oscillation amplitude (β and low γ) from both intra-region and inter-region dimensions. The PAC strength was quantified using the modulation index (MI). Multiple comparisons were corrected using the Bonferroni method. Finally, correlations between PAC metrics and Autism Behavior Checklist (ABC) scores were analyzed. Results: Compared to the control group, children with ASD exhibited significant frequency-specific PAC abnormalities: (1) Multi-regional γ hyper-coupling: There was a significant enhancement in the modulation of γ amplitude by δ/θ/α phase across the measured scalp regions, suggesting abnormal high-frequency synchronization. (2) Dissociated β modulation patterns: The ASD group showed increased δ–β coupling (predominantly in frontal, temporal, and occipital lobes) alongside significantly reduced α–β coupling (localized to frontal and central regions). This reflects both an abnormal locking of slow-wave activity to the β band and a diminished regulatory role of α oscillations. (3) Clinical correlation: Notably, abnormally elevated PAC strength (particularly in the δ/θ/α–γ bands) showed a negative correlation with clinical symptom severity—that is, stronger coupling was associated with lower scores on the ABC. Conclusions: Leveraging a large-sample dataset, this study characterizes the landscape of aberrant cross-frequency interactions in young children with ASD. Our findings indicate that the neuroelectrical activity in ASD goes beyond mere connectivity anomalies by demonstrating altered PAC strength at both the intra-region and inter-region levels. Notably, the strength of this aberrant intra-region PAC is correlated with clinical symptoms. Full article

Background/Objectives: Electroencephalography (EEG) is widely applied in emotion recognition. Integrating diverse frequency and spatial features to improve performance remains a major challenge. Methods: This paper proposes two preprocessing methods to map EEG signals into image-style representations. These methods preserve the spatial topology and enable effective feature extraction using convolutional neural networks. The first method is a spatial concatenation method (SCM). It projects three feature types onto color channels, providing a structural prior that encourages the network to learn the three feature types within local spatial windows. It differs from traditional spectral mixing, which maps frequency bands to color channels. The second method is a band-wise stacking method (BSM). It treats frequency bands as independent depth frames to form a three-dimensional tensor. This structure is designed to facilitate the learning of inter-band relationships while preserving band-specific information. Dedicated convolutional neural network architectures are designed for these tensor structures, aligned with the spatial and spectral organization of the proposed SCM and BSM. Results: Experiments on the DEAP and DREAMER datasets for binary Arousal and Valence classification show that both representations achieve competitive results. The BSM achieves higher accuracy than the SCM on the DREAMER dataset, while both methods perform comparably on the DEAP dataset. Conclusions: The proposed strategies offer efficient convolutional neural network approaches for EEG emotion recognition systems. Full article

Interpretability is essential for deploying deep learning (DL) models in electroencephalography (EEG)-based neurotechnology, particularly in brain–computer interfaces and clinical decision-support settings. Existing post hoc explainable artificial intelligence (XAI) methods often yield single-domain attribution maps, limiting their capacity to characterize the joint spatial, temporal, and spectral structure of EEG dynamics. In addition, perturbation-based strategies may disrupt physiological signal organization, whereas gradient-based methods require access to model internals and are therefore tied to specific classifier architectures. Here, we introduce space–time–frequency KernelSHAP (STF-KernelSHAP), a model-agnostic Shapley framework for physiologically coherent EEG explainability. The method comprises three stages. First, EEG trials are decomposed into structured channel–time–frequency cells using segment-wise spectral analysis. Second, coalitions are formed over complete channel–time–frequency cells and reconstructed in the signal domain to support physiologically informed perturbations. Third, class-conditional relevance is estimated with a KernelSHAP-based weighted surrogate model that uses only model outputs, enabling architecture-independent Shapley estimation. We evaluate STF-KernelSHAP on two prerecorded public datasets: the GIGA motor imagery/movement execution (MI-ME) dataset for motor imagery (MI) decoding and the IEEE DataPort EEG Data for Attention-Deficit/Hyperactivity Disorder (ADHD)/Control Children dataset for ADHD detection. For ADHD detection, the T-GARNet base classifier interpreted with STF-KernelSHAP achieved 73.33% accuracy and 79.86% area under the curve (AUC); these values characterize classifier performance rather than the explainer itself. We compare the framework against KernelSHAP, local interpretable model-agnostic explanations (LIME), Occlusion, Integrated Gradients, and gradient-weighted class activation mapping++ (Grad-CAM++). Fidelity is assessed with Deletion and remove and debias (ROAD), while qualitative analyses examine topographic and frequency-band attribution maps. Results show that STF-KernelSHAP remains functionally competitive with established XAI methods while providing window-dependent and frequency-specific explanations. Overall, STF-KernelSHAP offers a physiologically informed and model-agnostic alternative for multidomain EEG interpretability. Full article

Cognitive neuroscience has addressed the understanding of human brain processes through numerous techniques and psychological paradigms. In general, different types of tasks have been used depending on the specific cognitive operation under study. Since these tasks are usually designed to register responses at the single-trial level, the most common methodological approach to electroencephalography (EEG) is to obtain event-related potentials (ERPs). Crucially, the linear analysis methods associated with ERPs often overlook the intrinsic non-linear and multiscale dynamics of brain activity. Hence, to better characterize brain activity, there is increasing interest in the study of the non-linearity and complexity of EEGs. Given that experiments relating cognitive processing and EEG complexity are still scarce, this work is a narrative review of studies in which non-clinical cognitive processing, such as memory, perception, or attention, is addressed using complexity measures. Here, we focus on EEG metrics derived from the concepts of fractality, information, and randomness across different temporal and spatial scales. We discuss how these measures complement more classical analyses, try to integrate the findings using a predictability–regularity framework, and finally, we point out possible future directions with which to advance current knowledge about the relationship between cognition and EEG complexity. Full article

To address cross-subject domain shift and insufficient complementary fusion of time-frequency information in EEG-based emotion recognition, this paper proposes a multi-task adversarial network: Path Adversarial Dual-Branch Network for EEG Emotion Recognition (PADB-Net). The model adopts a dual-branch parallel architecture for time and frequency domains, processing raw EEG waveforms and differential entropy features respectively, and extracts discriminative features using lightweight depthwise separable convolutions and channel attention. A path adversarial module is introduced for the first time in emotion recognition to align time-domain and frequency-domain feature distributions, solving the single-branch dominance problem in dual-branch fusion. Together with a domain adversarial module, the overall distributions of source and target domains as well as the internal distributions of the two modality branches are aligned within a unified framework. Experiments on a dataset containing healthy subjects and patients with major depressive disorder show that the full model significantly outperforms single-adversarial and non-adversarial baselines in accuracy, AUC, F1-score, sensitivity, and specificity, verifying the synergistic gain of the dual-adversarial mechanism. On the HybridBCI dataset, PADB-Net achieves 77.80% accuracy, 84.50% AUC, and 79.40% F1-score with only 6.45 K trainable parameters. When transferred to the public SEED dataset for three-class emotion recognition, the model attains F1-scores of 71.83% (negative), 68.99% (neutral), and 73.37% (positive), demonstrating strong cross-dataset generalizability. Full article

Electroencephalography (EEG) is a noninvasive tool used by healthcare professionals to measure brain electrical activity. EEG analysis can indicate various anomalies linked to different brain pathologies, including seizures. Traditionally, the analysis is confined to two-dimensional displays and relies exclusively on the visual modality, limiting a comprehensive overview. EEG analysis through visualisation is challenging and time-consuming, and artificial intelligence (AI) is increasingly used to aid the process of seizure detection. However, the educational value of AI-assisted seizure detection models depends on the explainability of the underlying models. Explainable AI can help learners understand the features and patterns associated with seizure detection and also support informed use of AI-based decision support systems. M2EEG-VR leverages the focus and immersive capabilities of virtual reality (VR) with the aim of developing a multi-modal platform for EEG seizure detection analysis with a human-in-the-loop. The ability to understand EEG and seizure patterns is key to addressing and effectively treating many neurological conditions. Neonatal seizure detection is particularly challenging where seizure patterns are subtle and context dependent. This study advances toward multi-modal analysis by encoding EEG signals into auditory representations using AI that aids in the acoustic detection of the presence of neonatal seizures in EEG. The platform also introduces a 3D brain model with a spatial mapping of seizure regions. In a user study (N = 20, 4 prior EEG experience, 16 no prior EEG experience), participants achieved higher seizure detection accuracy in the combined visual and auditory condition (mean = 7.6 ± 1.2) than in visual-only or audio-only modes. These preliminary findings suggest that a multi-modal environment may improve the accuracy of detection. However, further controlled studies are needed to ascertain the performance benefits. Usability was rated excellent (SUS = 83 ± 11), and task load remained moderate (NASA-TLX = 36.6). The findings suggest that VR multi-modal interaction can reduce cognitive load and enhance the explainability of complex EEG data in a focused virtual environment. The analysis of the diagnostic accuracy showed that participants without prior EEG knowledge performed similarly across all modalities to those with prior EEG knowledge. This implies that the accessibility barrier is reduced for novice users using the tool for the EEG review/detection task. This, together with high usability and moderate task load scores, indicates that the tool may be suitable for medical training applications. A multi-modal EEG in VR may prove useful in education and also be used as a test bench to further explore AI with human-in-the-loop paradigms for seizure detection. Full article

Transcranial magnetic stimulation-electroencephalography (TMS–EEG) biomarkers have recently become available as a means to obtain new understanding of the causal chains of neuronal signaling in the brain. This is a key piece in the puzzle of how the brain is organized and how it works. Using dMRI tractography, we can map the circuit beneath a chosen cortical target; TMS can then stimulate it, and EEG records responses that reflect—and may even be caused by—activity in that structural circuit. The chain of events after stimulus delivery can be observed and quantified using current neuroimaging and TMS–EEG technology, a matter of tremendous relevance on how to approach novel therapeutic approaches in clinical conditions. Herein, we elaborate upon a perspective of how groundbreaking multi-locus TMS (mTMS) technology associated with EEG and multimodal neuroimaging can be applied to modulate the flow dynamics of the glymphatic system (GS). The enhancement of the GS waste clearance functionality has been shown to improve significantly symptom severity in neurodegenerative disorders such as Alzheimer’s (AD) and Parkinson’s disease (PD) or long COVID. In this perspective paper, we consider that next-generation therapeutics using versatile technologies such as noninvasive neuromodulation and neuroimaging will provide important benefits in public health and in how society can address the management of these difficult-to-deal-with ailments more effectively. Full article

Deception detection is a multifaceted challenge that has gained attention in domains such as forensics, security, and human–computer interaction. However, most EEG-based studies focus on binary classification between truthful and deceptive responses, overlooking the complexity of cognitive processes underlying different deceptive strategies. To address this limitation, we present a multi-class EEG dataset designed to investigate distinct behavioral roles in deception, including honest, bluffer, liar, and deceiver, collected from 51 participants using a controlled mock-crime scenario. In this setup, subjects were assigned predefined roles and interrogated under a standardized protocol with carefully designed questions and responses. EEG signals were recorded using a 16-channel Biosemi ActiveTwo system at a sampling rate of 2048 Hz, with event markers enabling precise temporal segmentation of experimental phases. The dataset captures neural activity associated with varying cognitive load and decision-making across deception types. To the best of our knowledge, this is the first EEG dataset that explicitly incorporates and differentiates four distinct deception-related behavioral roles within a unified experimental framework. Full article

Automatic analysis of electroencephalography (EEG) recordings relies on large, high-quality labeled datasets. Manual segmentation by medical experts is resource-intensive and time-consuming. Moreover, to overcome potential subjectivity in labeling, independent annotation by at least two experts is required. Therefore, reliable automatic data labeling is essential for obtaining the large datasets needed to train robust AI models. In this paper, we show that a properly trained state-of-the-art deep neural network (DNN) achieves labeling performance comparable to inter-expert agreement in the task of segmenting epileptiform activity patterns. To this end, we first compiled a custom database of EEG recordings containing such patterns. Second, five experts based on part of these recordings independently assessed spike-wave index (SWI), which is a key diagnostic criterion that indicates the percentage of the EEG recording during which epileptic discharges are observed. Third, we compared the expert assessments with SWI calculated based on automatic segmentation by the trained DNN. Our results demonstrate that the 1D U-Net architecture achieves competitive overall performance and aligns well with both expert assessments and expert-derived SWI values. Thus, automated segmentation and analysis of EEG recordings holds great promise for accelerating diagnosis and developing targeted therapeutic strategies for epilepsy. Full article

We present an interpretable evolutionary-fuzzy feature extraction framework for high-dimensional electroencephalography (EEG) classification. The proposed method combines an evolution strategy (ES) optimizer with fuzzy membership encoding to automatically discover compact, nonlinear feature representations from raw EEG signals. Applied to a chemosensory experiment distinguishing nasal breathing conditions during taste perception (N = 10 between-subjects participants, 1600 trials, 612 raw features), the framework achieves 89.50% cross-validated accuracy, equivalent to or exceeding all 25-feature baselines, while reducing dimensionality by 95.9% (from 612 to 25 features). The method produces fully interpretable fuzzy rules, enabling neuroscientists to inspect the decision logic rather than relying on nontransparent classifiers. A comprehensive validation including noise robustness analysis (0–30% Gaussian noise) and between-subjects generalization assessment is provided. Due to the between-subjects design, this study focuses on demonstrating the within-dataset discriminative capacity and the interpretability of the feature extraction pipeline, rather than claiming true subject-independent generalization. Full article

Autism spectrum disorder (ASD) is a neurodevelopmental condition where early detection is crucial for improving outcomes. Electroencephalography (EEG) offers a non-invasive approach for identifying potential biomarkers. In this study, we investigate hemispheric asymmetry using the Lateralisation Index (LI) derived from EEG signals and evaluate its effectiveness for ASD classification. Using a small dataset of children with and without ASD, we applied several machine learning models, including Logistic Regression, Support Vector Machines and Random Forest. Particular attention was given to evaluation strategies to avoid overfitting and data leakage. While initial results suggested moderate classification performance, repeated validation indicated unstable generalisation. Our findings highlight both the potential and limitations of LI-based features in small-sample settings and emphasise the importance of robust evaluation in EEG-based machine learning studies. Full article

Background/Objectives: Minimally invasive surgery benefits patients but places physical and cognitive demands on surgeons. While robot-assisted laparoscopic surgery (RALS) reduces musculoskeletal strain, it may increase cognitive load. This study examined whether physical and psychological preparatory protocols (warm-ups) influence surgeon strain during RALS. Methods: Ten consultant surgeons from East Lancashire Hospitals NHS Trust (UK) participated in a preregistered, randomized study. Each performed RALS under three conditions: control, physical warm-up (10 min simulation tasks on the Da Vinci system), and psychological warm-up (10 min PETTLEP-based mental imagery). Electromyography (EMG) and electroencephalography (EEG) were recorded during key surgical phases. EMG data were normalized to maximal voluntary contractions. Results: The physical warm-up significantly increased EMG activity in the right deltoid and right trapezius (p < 0.05) compared to control, with no differences observed in other muscle groups. EEG alpha power data did not significantly differ between conditions. Conclusions: These findings suggest that brief physical warm-up can enhance muscle activation in key regions involved in RALS, potentially improving motor control and reducing fatigue. Incorporating such strategies may support surgeon performance and well-being. Full article

Background: Assessing cortical and muscle activity simultaneously during task performance may inform motor function post-stroke. This study evaluated brain–muscle functional connectivity (corticomuscular coherence, CMC) in early stroke recovery. Methods: Individuals with stroke in an inpatient rehabilitation facility (IRF) completed motor assessments and simultaneous electroencephalography (EEG) and electromyography (EMG) recordings during a grip task at IRF admission and discharge. Beta (20–30 Hz) CMC was measured between EEG electrodes overlying the primary motor cortex (M1) and supplementary motor area (SMA) and EMG leads overlying the first dorsal interosseous (FDI). Neurotypical controls completed identical EEG/EMG recordings. Correlational analyses were performed to ascertain CMC and motor assessment associations. CMC differences by Group (Stroke vs. Controls), Time (Admission vs. Discharge), and Extremity (Affected/Dominant vs. Less Affected/Non-Dominant) were estimated using mixed-effects linear models. Results: Thirty individuals with stroke (14 females, mean age 67.0 ± 9.8 years, 10.4 ± 3.5 days post-stroke) and 17 controls (8 females, mean age 75.3 ± 13 years) participated. Individuals with stroke exhibited reduced beta CMC between SMA and affected FDI (F_(1,36.1) = 5.73, p = 0.02, Cohen’s f = 0.40) compared to controls, with lower CMC involving the affected vs. less affected extremity (F_(1,73.0) = 5.72, p = 0.01, Cohen’s f = 0.28). Greater beta SMA–FDI CMC at admission related to poorer motor recovery (ρ = −0.59, p = 0.01). Group and Extremity CMC differences were not observed over time, nor were there changes in affected extremity CMC from admission to discharge. Conclusions: Beta SMA–FDI CMC is a marker of neural injury, exhibiting extremity-specific differences early post-stroke. While beta SMA–FDI CMC correlated with motor recovery, the absence of change over time during hospitalization necessitates longitudinal assessments to clarify its trajectory alongside recovery. Full article