Cognitive Neurodynamics最新文献

This research scrutinizes the simultaneous evolution of each layer within a multilayered complex neural network and elucidates the effect of synaptic plasticity on inter-layer dynamics. In the absence of synaptic plasticity, a predominant feedforward effect is observed, resulting in the manifestation of complete synchrony in deep networks, with each layer assuming a chimera state. A significant increase in the number of synchronized neurons is observed as the layers augment, culminating in complete synchronization in the deeper sections. The study categorizes the layers into three distinct parts: the initial layers (1–4) demonstrate the emergence of non-uniformity in the random firing of neurons; the middle layers (5–7) exhibit an amplification of this non-uniformity, forming a higher degree of synchronization; and the final layers (8–10) display a completely synchronized process. The introduction of synaptic plasticity disrupts this synchrony, inducing periodic oscillation characteristics across layers. The specificity of these oscillations is notably accentuated with increasing network depth. These insights shed light on the interplay between neural network complexity and synaptic plasticity in influencing synchronization dynamics, presenting avenues for enhanced neural network architectures and refined neuroscientific models. The findings underscore the imperative to delve deeper into the implications of synaptic plasticity on the structure and function of intricate multi-layer neural networks.

Critical states present scale-free dynamics, optimizing neuronal complexity and serving as a potential biomarker in cognitively impaired patients. We explored electroencephalogram (EEG) criticality in amnesic Mild Cognitive Impairment patients with clinical improvement in working memory, verbal memory, verbal fluency and overall executive functions after the completion of a 6-month prospective memory training. We compared “before” and “after” stationary resting-state EEG records of right-handed MCI patients (n = 17; 11 females), using the method of critical fluctuations and Haar wavelet analysis. Improvement of criticality indices was observed in most electrodes, with mean values being higher after prospective memory training. Significant criticality enhancement was found in the subgroup analysis of frontotemporal electrodes [mean dif: 0.10; Z = 7, p = 0.019]. In the isolated electrode signal analysis, significant post-intervention improvement was noted in pooled criticality indices of electrodes T6 [mean dif: 0.204; t(10) = −2.3, p = 0.044] and F4 [mean dif: 0.0194; t(10) = −2.82; p = 0.018]. EEG criticality agreed with clinical improvement, consisting a possible quantifiable and easy-to-obtain biomarker in MCI and Alzheimer’s disease (AD), especially in patients under cognitive training/rehabilitation. We highlight the role of EEG in prognostication, monitoring and potentially early treatment optimization in MCI or AD patients. Further standardization of the methodology in larger patient cohorts could be valuable for AD theragnostics in patients receiving disease-modifying treatments by providing insights regarding synaptic brain plasticity.

Graphical Abstract

Critical states’ scale-free dynamics optimize neuronal complexity, emerging as biomarkers in cognitive neuroscience. Applying the method of critical fluctuations and Haar wavelet analysis in stationary EEG time-series, we demonstrate criticality enhancement in the frontotemporal electroencephalographic (EEG) recordings of mild cognitive impairment (MCI) patients after a 6-month prospective memory training, suggesting EEG criticality as a possible monitoring biomarker in MCI and Alzheimer’s disease.

Alzheimer’s disease (AD) and frontotemporal dementia (FTD) are two main types of dementia. These diseases have similar symptoms, and they both may be considered as AD. Early detection of dementia and differential diagnosis between AD and FTD can lead to more effective management of the disease and contributes to the advancement of knowledge and potential treatments. In this approach, several features were extracted from electroencephalogram (EEG) signals of 36 subjects diagnosed with AD, 23 FTD subjects, and 29 healthy controls (HC). Mann–Whitney U-test and t-test methods were employed for the selection of the best discriminative features. The Fp1 channel for FTD patients exhibited the most significant differences compared to AD. In addition, connectivity features in the delta and alpha subbands indicated promising discrimination among these two groups. Moreover, for dementia diagnosis (AD + FTD vs. HC), central brain regions including Cz and Pz channels proved to be determining for the extracted features. Finally, four machine learning (ML) algorithms were utilized for the classification purpose. For differentiating between AD and FTD, and dementia diagnosis, an accuracy of 87.8% and 93.5% were achieved respectively, using the tenfold cross-validation technique and employing support vector machines (SVM) as the classifier.

Brain-computer interface (BCI) based on the motor imagery paradigm typically utilizes multi-channel electroencephalogram (EEG) to ensure accurate capture of physiological phenomena. However, excessive channels often contain redundant information and noise, which can significantly degrade BCI performance. Although there have been numerous studies on EEG channel selection, most of them require manual feature extraction, and the extracted features are difficult to fully represent the effective information of EEG signals. In this paper, we propose a spatio-temporal-graph attention network for channel selection (STGAT-CS) of EEG signals. We consider the EEG channels and their inter-channel connectivity as a graph and treat the channel selection problem as a node classification problem on the graph. We leverage the multi-head attention mechanism of graph attention network to dynamically capture topological relationships between nodes and update node features accordingly. Additionally, we introduce one-dimensional convolution to automatically extract temporal features from each channel in the original EEG signal, thereby obtaining more comprehensive spatiotemporal characteristics. In the classification tasks of the BCI Competition III Dataset IVa and BCI Competition IV Dataset I, STGAT-CS achieved average accuracies of 91.5% and 85.4% respectively, demonstrating the effectiveness of the proposed method.

Dementia is a neuro-degenerative disorder with a high death rate, mainly due to high human error, time, and cost of the current clinical diagnostic techniques. The existing dementia detection methods using hand-crafted electroencephalogram (EEG) signal features are unreliable. A convolution neural network using spatiotemporal EEG signals (STEADYNet) is presented to improve the dementia detection. The STEADYNet uses a multichannel temporal EEG signal as input. The network is grouped into feature extraction and classification components. The feature extraction comprises two convolution layers to generate complex features, a max-pooling layer to reduce the EEG signal’s spatiotemporal redundancy, and a dropout layer to improve the network’s generalization. The classification processes the feature extraction output nonlinearly using two fully-connected layers to generate salient features and a softmax layer to generate disease probabilities. Two publicly available multiclass datasets of dementia are used for evaluation. The STEADYNet outperforms existing automatic dementia detection methods with accuracies of (99.29%), (99.65%), and (92.25%) for Alzheimer's disease, mild cognitive impairment, and frontotemporal dementia, respectively. The STEADYNet has a low inference time and floating point operations, suitable for real-time applications. It may aid neurologists in efficient detection and treatment. A Python implementation of the STEADYNet is available at https://github.com/SandeepSangle12/STEADYNet.git

Fatigue driving significantly contributes to global vehicle accidents and fatalities, making driver fatigue level estimation crucial. Electroencephalography (EEG) is a proven reliable predictor of brain states. With Deep Learning (DL) advancements, brain state estimation algorithms have improved significantly. Nonetheless, EEG’s multi-domain nature and the intricate spatial-temporal-frequency correlations among EEG channels present challenges in developing precise DL models. In this work, we introduce an innovative Attention-based Cross-Frequency Graph Convolutional Network (ACF-GCN) for estimating drivers’ reaction times using EEG signals from theta, alpha, and beta bands. This method utilizes a multi-head attention mechanism to detect long-range dependencies between EEG channels across frequencies. Concurrently, the transformer’s encoder module learns node-level feature maps from the attention-score matrix. Subsequently, the Graph Convolutional Network (GCN) integrates this matrix with feature maps to estimate driver reaction time. Our validation on a publicly available dataset shows that ACF-GCN outperforms several state-of-the-art methods. We also explore the brain dynamics within the cross-frequency attention-score matrix, identifying theta and alpha bands as key influencers in fatigue estimating performance. The ACF-GCN method advances brain state estimation and provides insights into the brain dynamics underlying multi-channel EEG signals.

This work aims to explore the control effect of DBS on Alzheimer's disease (AD) from a neurocomputational perspective. Firstly, a data-driven cortical network model is constructed using the Diffusion Tensor Imaging data. Then, a typical electrophysiological feature of EEG slowing in AD is reproduced by reducing the synaptic connectivity parameters. The corresponding changes in kinetic behavior mainly include an oscillation decrease in the amplitude and frequency of the pyramidal neuron population. Subsequently, DBS current with specific parameters is introduced into three potential targets of the hippocampus, the nucleus accumbens and the olfactory tubercle, respectively. The results indicate that applying DBS to simulated mild AD patients induces an increase in relative alpha power, a decrease in relative theta power, and a significant rightward shift of the dominant frequency. This is consistent with the EEG reversal in pharmacological treatments for AD. Further, the optimal stimulation strategy of DBS is investigated through spectral and statistical analyses. Specifically, the pathological symptoms of AD could be alleviated by adjusting the critical parameters of DBS, and the control effect of DBS on various targets is that the hippocampus is superior to the olfactory tubercle and nucleus accumbens. Finally, using correlation analysis between the power increments and the nodal degrees, it is concluded that the control effect of DBS is related to the importance of the nodes in the brain network. This study provides a theoretical guidance for determining DBS targets and parameters, which may have a substantial impact on the development of DBS treatment for AD.

Transcranial magnetic stimulation (TMS) is a non-invasive neuromodulation technique to activate or inhibit the activity of neurons, and thereby regulate their excitability. This technique has demonstrated potential in the treatment of neuropsychiatric disorders, such as depression. However, the effect of TMS on neurons with different severity of depression is still unclear, limiting the development of efficient and personalized clinical application parameters. In this study, a multi-scale computational model was developed to investigate and quantify the differences in neuronal responses to TMS with different degrees of depression. The microscale neuronal models we constructed represent the hippocampal CA1 region in rats under normal conditions and with varying severities of depression (mild, moderate, and major depressive disorder). These models were then coupled to a macroscopic TMS-induced E-Fields model of a rat head comprising multiple types of tissue. Our results demonstrate alterations in neuronal membrane potential and calcium concentration across varying levels of depression severity. As depression severity increases, the peak membrane potential and polarization degree of neuronal soma and dendrites gradually decline, while the peak calcium concentration decreases and the peak arrival time prolongs. Concurrently, the electric fields thresholds and amplification coefficient gradually rise, indicating an increasing difficulty in activating neurons with depression. This study offers novel insights into the mechanisms of magnetic stimulation in depression treatment using multi-scale computational models. It underscores the importance of considering depression severity in treatment strategies, promising to optimize TMS therapeutic approaches.

Perception of voice means acoustic electric conversion in the auditory system, and changes of external magnetic field can affect the neural activities by taming the channel current via some field components including memristor and Josephson junction. Combination of two capacitors via an electric component is effective to describe the physical property of artificial cell membrane, which is often used to reproduce the characteristic of electric activities in cell membrane. Involvement of two capacitive variables for two capacitors in the neural circuit can discern the effect of field diversity in the media in two sides of the cell membrane in theoretical way. A Josephson junction is used to couple a piezoelectric neural circuit composed of two capacitors, one inductor and one nonlinear resistor. Field energy is mainly kept in the capacitive and inductive components, and it is obtained and converted into dimensionless energy function. The Hamilton energy function in an equivalent auditory neuron is verified by using the Helmholtz theorem. Noisy excitation on the neural circuit can be detected via the Josephson junction channel and similar stochastic resonance is detected by regulating the noise intensity, as a result, the average energy reaches a peak value under stochastic resonance. An adaptive law controls the bifurcation parameter, which is relative to the membrane property, and energy shift controls the mode selection during continuous growth of the bifurcation parameter. That is, external energy injection derived from acoustic wave or magnetic field will control the energy level, and then suitable firing patterns are controlled effectively.

Ischemic stroke (IS) is characterized by high mortality, disability rates, and a high risk of recurrence. Motor dysfunction, such as limb hemiparesis, dysphagia, auditory disorders, and speech disorders, usually persists after stroke, which imposes a heavy burden on society and the health care system. Traditional rehabilitation therapies may be ineffective in promoting functional recovery after stroke, and alternative strategies are urgently needed. The Food and Drug Administration (FDA) has approved invasive vagus nerve stimulation (iVNS) for the improvement of refractory epilepsy, treatment-resistant depression, obesity, and moderate to severe upper limb motor impairment following chronic ischemic stroke. Additionally, the FDA has approved transcutaneous vagus nerve stimulation (tVNS) for the improvement of cluster headaches and acute migraines. Recent studies have demonstrated that vagus nerve stimulation (VNS) has neuroprotective effects in both transient and permanent cerebral ischemia animal models, significantly improving upper limb motor impairments, auditory deficits, and swallowing difficulties. Firstly, this article reviews two potential neuronal death pathways following IS, including autophagy and inflammatory responses. Then delves into the current status of preclinical and clinical research on the functional recovery following IS with VNS, as well as the potential mechanisms mediating its neuroprotective effects. Finally, the optimal parameters and timing of VNS application are summarized, and the future challenges and directions of VNS in the treatment of IS are discussed. The application of VNS in stroke rehabilitation research has reached a critical stage, and determining how to safely and effectively translate this technology into clinical practice is of utmost importance. Further preclinical and clinical studies are needed to elucidate the therapeutic mechanisms of VNS.