Cognitive Neurodynamics最新文献

Brain-controlled wheelchair (BCW) technology enables direct wheelchair control via brain-computer interfaces (BCIs), eliminating the need for physical limb interaction. Motor imagery-based BCIs (MI-BCIs) are widely used in non-invasive BCIs due to their ability to provide intuitive neural control without external stimuli. However, developing a BCW system based on MI-BCIs remains challenging, particularly in achieving reliable multi-class classification accuracy.To address this challenge, this study proposes an advanced feature extraction algorithm to enhance MI-BCI performance using a custom-built five-class MI-EEG dataset. The proposed method, EHT-CSP, integrates Ensemble Empirical Mode Decomposition Hilbert-Huang Transform (EEMD-HHT) with Time-Frequency Common Spatial Pattern (TFCSP). Specifically, it extracts marginal spectrum entropy and energy spectrum entropy via EEMD-HHT. It then combines these features with TFCSP-derived feature vectors to improve feature discrimination. The Light Gradient Boosting Machine is then employed for classification. The proposed MI-BCI system is evaluated through both offline analysis and real-world BCW obstacle avoidance experiments. Results demonstrate that the algorithm achieves an average classification accuracy of 78.45%, with all participants successfully completing BCW navigation tasks. In this study, LightGBM and EHT-CSP are compared with other algorithms respectively, and it is verified that the proposed model is superior to the existing models.

One of the areas where artificial intelligence (AI) technologies are used is the detection and diagnosis of mental disorders. AI approaches, including machine learning and deep learning models, can identify early signs of bipolar disorder, schizophrenia, autism spectrum disorder, depression, suicidality, and dementia by analyzing speech patterns, behaviors, and physiological data. These approaches increase diagnostic accuracy and enable timely intervention, which is crucial for effective treatment. This paper presents a comprehensive literature review of AI approaches applied to mental disorder detection using various data sources, such as survey, Electroencephalography (EEG) signal, text and image. Applications include predicting anxiety and depression levels in online games, detecting schizophrenia from EEG signals, detecting autism spectrum disorder, analyzing text-based indicators of suicidality and depression, and diagnosing dementia from magnetic resonance imaging images. eXtreme Gradient Boosting (XGBoost), light gradient-boosting machine (LightGBM), random forest (RF), support vector machine (SVM), K-nearest neighbor were designed as machine learning models, and convolutional neural networks (CNN), long short-term memory (LSTM) and gated recurrent unit (GRU) models suitable for the dataset were designed as deep learning models. Data preprocessing techniques such as wavelet transforms, normalization, clustering were used to optimize model performances, and hyperparameter optimization and feature extraction were performed. While the LightGBM technique had the highest performance with 96% accuracy for anxiety and depression prediction, the optimized SVM stood out with 97% accuracy. Autism spectrum disorder classification reached 98% accuracy with XGBoost, RF and LightGBM. The LSTM model achieved a high accuracy of 83% in schizophrenia diagnosis. The GRU model showed the best performance with 93% accuracy in text-based suicide and depression detection. In the detection of dementia, LSTM and GRU models have demonstrated their effectiveness in data analysis by reaching 99% accuracy. The findings of the study highlight the effectiveness of LSTM and GRU for sequential data analysis and their applicability in medical imaging or natural language processing. XGBoost and LightGBM are noted to be highly accurate ML tools for clinical diagnoses. In addition, hyperparameter optimization and advanced data pre-processing approaches are confirmed to significantly improve model performance. The results obtained with this study have revealed the potential to improve clinical decision support systems for mental disorders with AI, facilitating early diagnosis and personalized treatment strategies.

The main aim of this study is to propose a novel convolutional neural network, named BrainNeXt, for the automated brain disorders detection using magnetic resonance images (MRI) images. Furthermore, we aim to investigate the performance of our proposed network on various medical applications. To achieve high/robust image classification performance, we gathered a new MRI dataset belonging to four classes: (1) Alzheimer's disease, (2) chronic ischemia, (3) multiple sclerosis, and (4) control. Inspired by ConvNeXt, we designed BrainNeXt as a lightweight classification model by incorporating the structural elements of the Swin Transformers Tiny model. By training our model on the collected dataset, a pretrained BrainNeXt model was obtained. Additionally, we have suggested a feature engineering (FE) approach based on the pretrained BrainNeXt, which extracted features from fixed-sized patches. To select the most discriminative/informative features, we employed the neighborhood component analysis selector in the feature selection phase. As the classifier for our patch-based FE approach, we utilized the support vector machine classifier. Our recommended BrainNeXt approach achieved an accuracy of 100% and 91.35% for training and validation. The recommended model obtained the test classification accuracy of 94.21%. To further improve the classification performance, we suggested a patch-based DFE approach, which achieved a test accuracy of 99.73%. The obtained results, surpassing 90% accuracy on the test dataset, demonstrate the effectiveness and high classification performance of the proposed models.

This paper investigates the finite-time synchronization of complex-valued memristive neural networks (CVMNNs) with time-varying delays using an event-triggered control approach. The analysis is conducted in a holistic manner, utilizing the one-norm and sign functions of complex numbers, thereby eliminating the need for decomposition. To alleviate communication pressure, an event-triggered controller is introduced, accompanied by specific conditions and criteria to guarantee synchronization within a finite time frame. Additionally, a direct estimate of the synchronization time is provided, and a positive lower bound on the minimum event interval is derived to prevent Zeno behavior. Building on this event-triggered strategy, a self-triggered mechanism is designed to eliminate the necessity for continuous monitoring. The proposed method is straightforward and easily implementable, with its effectiveness demonstrated through illustrative examples and simulation results.

The top-down regulation of prior content facilitates the efficiency of following speech perception through the theta-band synchronization between higher-level cognitive regions and lower-level phonetic processing areas. However, how this regulation affects tone processing and its corresponding functional pathway remains unknown. In this study, we conducted three different auditory oddball paradigms which differed in prior constraints among Mandarin Chinese speakers. We calculated the amplitude of P3 differences caused by tone variations to evaluate the efficiency of tone processing within each paradigm. Theta-band functional connectivity (FC) related to lower-level phonetic processing areas was also analyzed at the source level to identify the specific top-down regulation loop. Our results showed that top-down regulation effects modulated responses to upcoming tonal processing reflected by smaller P3 amplitude differences with the occurrence of semantic priming. Results of FC analysis revealed different corresponding cortical contributions depending on priming content. Semantic-driven top-down regulation enhances FC between the the left caudal middle frontal gyrus and lower-level phonetic processing area. Moreover, when the prior constraint is semantically violated, enhanced FC between the left pars triangularis and the left supramarginal gyrus with lower-level phonetic processing regions were seen. Our study provides neurophysiological insights into the effects of top-down regulation on tone perception.

Supplementary information: The online version contains supplementary material available at 10.1007/s11571-025-10314-1.

Recent studies combining electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) have shown promising results linking neural and vascular responses. This study analyzes the topographical effect of auditory stimulus intensity on cortical activation and explores neurovascular coupling between fNIRS hemodynamic signals and auditory-evoked potentials (AEPs), extracted from EEG. Forty healthy volunteers (13 males, 27 females; mean age = 22.27 ± 3.96 years) listened to complex tones of varying intensities (50-, 70-, and 90-dB SPL) across seven frequencies (range of 400-2750 Hz) in blocks of five, while EEG and fNIRS were recorded. PERMANOVA analysis revealed that increasing intensity modulated hemodynamic activity, leading to amplitude changes and enhanced recruitment of auditory and prefrontal cortices. To isolate stimulus-specific activity, Spearman correlations were computed on residuals-components of AEPs and fNIRS responses with individual trends removed. The N1 amplitude increase was correlated with higher superior temporal gyrus (STG) and superior frontal gyrus (SFG) activity, and reduced activity in inferior frontal gyrus (IFG) for the oxygenated hemoglobin (HbO), while the deoxygenated hemoglobin (HbR) was associated with increased activity in one channel near the Supramarginal Gyrus (SMG). P2 amplitude increase was associated with higher activation in SFG and IFG for HbO, while for HbR with the activity in SMG, angular gyrus (AnG), SFG, and IFG. Additionally, internal correlations between fNIRS channels revealed strong associations within auditory and frontal regions. These findings provide insights into existing models of neurovascular coupling by showing how stimulus properties, such as intensity, modulate the relationship between neural activity and vascular responses.

Supplementary information: The online version contains supplementary material available at 10.1007/s11571-025-10281-7.

The metacognition of one's planning strategy constitutes a "second-level" of metacognition that goes beyond the knowledge and monitoring of one's cognition and refers to the ability to use awareness mechanisms to regulate execution of present or future actions effectively. This study investigated the relation between metacognition of one's planning strategy and the behavioral and electrophysiological (EEG) correlates that support strategic planning abilities during performance in a complex decision-making task. Moreover, a possible link between task execution, metacognition, and individual differences (i.e., personality profiles and decision-making styles) was explored. A modified version of the Tower of Hanoi task was proposed to a sample of healthy participants, while their behavioral and EEG neurofunctional correlates of strategic planning were collected throughout the task with decisional valence. After the task, a metacognitive scale, the 10-item Big Five Inventory, the General Decision-Making Style inventory, and the Maximization Scale were administered. Results showed that the metacognitive scale enables to differentiate between the specific dimensions and levels of metacognition that are related to strategic planning behavioral performance and decision. Higher EEG delta power over left frontal cortex (AF7) during task execution positively correlates with the metacognition of one's planning strategy for the whole sample. While increased beta activity over the left frontal cortex (AF7) during task execution, higher metacognitive beliefs of efficacy and less willingness to change their strategy a posteriori were correlated with specific personality profiles and decision-making styles. These findings allow researchers to delve deeper into the multiple facets of metacognition of one's planning strategy in decision-making.

Insomnia, as a common sleep disorder, is the most common complaints in medical practice affecting a large proportion of the population on a situational, recurrent or chronic basis. It has been demonstrated that, during wakefulness, patients with insomnia exhibit increased EEG power in theta, beta, and gamma band. However, the relevant mechanisms underlying such power changes are still lack of understanding. In this paper, by combining the neural computational model with the real EEG data, we focus on exploring what's behind the EEG power changes for insomniac. We first develop a modified Liley model, named FSR-Liley, by respectively considering the fast and slow synaptic responses in inhibitory neurons along with the one-way projection between them. Then we introduce a parameter selection and evaluation method based on Markov chain Monte Carlo algorithm and Wasserstein distance, by which the sensitive parameters are selected automatically, and meanwhile, the optimal values of selected parameters are evaluated. Finally, through combining with EEG data, we determine the sensitive parameters in FSR-Liley and accordingly provide the mechanistic hypotheses: (1) decrease in $P_{e{i}_f}$ , corresponding to the input from the thalamus to cortical inhibitory population with fast synaptic response, leads to the increased theta and beta power; (2) decrease in $N_{e{i}_f}$ , corresponding to the projection from cortical excitatory population to inhibitory population with fast synaptic response, causes the increased gamma power. The results in this paper provide insights into the mechanisms of EEG power changes in insomnia and establish a theoretical foundation to support further experimental research.

Granule cells (GCs) are mainly responsible for receiving and integrating information from the entorhinal cortex and transferring it to the hippocampus to accomplish memory-related functions such as pattern separation. Owing to the heterogeneity of GCs, there are also two other subtypes, namely semilunar granule cells (SGCs) and hilar ectopic granule cells (HEGCs). In order to investigate their differences, here we examine the disparities in dendritic integration among the different subtypes of GCs. By utilizing biological experimental data, we developed detailed multi-compartment models for each type of GC. Our findings reveal that under the excitatory synaptic inputs (mediated by AMPA receptors), the dendritic integration of GCs, SGCs and HEGCs are linear, sublinear, and supralinear respectively. Furthermore, we propose that the sublinear integration observed in SGCs may be attributed to a high density of V-type potassium channels (K $_{\text{V}}^{}$ ) distributed in dendrites with smaller volume and higher input resistance; while the supralinear integration seen in HEGCs may be due to a high density of T-type calcium channels (Ca $_{\text{T}}^{}$ ) distributed in dendrites with larger volume and lower input resistance. Additionally, sodium channels, six types of potassium channels (K $_{\text{A}}^{}$ , K $_{\text{M}}^{}$ , sK $_{\text{DR}}^{}$ , fK $_{\text{DR}}^{}$ , BK, SK), and two types of calcium channels (Ca $_{\text{N}}^{}$ , Ca $_{\text{L}}^{}$ ) have minimal influence on their respective integration modes. We also found different integration modes exhibit varied somatic firing rates when subjected to different spatial synaptic activation sets, the HEGCs with the supralinear integration demonstrate higher somatic firing rates than the SGCs with the sublinear integration. These results provide theoretical insights into understanding the distinct roles played by these three subtypes of granule cells in memory-related functions within the dentate gyrus.

Supplementary information: The online version contains supplementary material available at 10.1007/s11571-025-10226-0.