A Review on Machine Learning for EEG Signal Processing in Bioengineering

Global trends of delayed graft function in kidney transplantation from 2013 to 2023: a bibliometric analysis

Delayed graft function (DGF) is an early complication after kidney transplantation. The literature on DGF has experienced substantial growth. However, there is a lack of bibliometric analysis of DGF. This study aimed to analyze the scientific outputs of DGF and explore its hotspots from 2013 to 2023 by using CiteSpace and VOSviewer. The 2058 pieces of literature collected in the Web of Science Core Collection (WOSCC) from 1 January 2013 to 31 December 2023 were visually analyzed in terms of the annual number of publications, authors, countries, journals, literature co-citations, and keyword clustering by using CiteSpace and VOSviewer. We found that the number of papers published in the past ten years showed a trend of first increasing and then decreasing; 2021 was the year with the most posts. The largest number of papers was published by the University of California System, and the largest number of papers was published by the United States. The top five keyword frequency rankings are: 'delayed graft function', 'kidney transplantation', 'renal transplantation', 'survival', and 'recipients'. These emerging trends include 'brain death donors', 'blood absence re-injection injuries', 'tacrolimus', 'older donors and recipients', and 'artificial intelligence and DGF'. In summary, this study reveals the authors and institutions that could be cooperated with and discusses the research hotspots in the past ten years. It provides a reference and direction for future research and application of DGF.

Cascaded Thinning in Upscale and Downscale Representation for EEG Signal Processing

Smoothing filters are widely used in EEG signal processing for noise removal while preserving signals' features. Inspired by our recent work on Upscale and Downscale Representation (UDR), this paper proposes a cascade arrangement of some effective image-processing techniques for signal filtering in the image domain. The UDR concept is to visualize EEG signals at an appropriate line width and convert it to a binary image. The smoothing process is then conducted by skeletonizing the signal object to a unit width and projecting it back to the time domain. Two successive UDRs could result in a better-smoothing performance, but their binary image conversion should be restricted. The process is computationally ineffective, especially at higher line width values. Cascaded Thinning UDR (CTUDR) is proposed, exploiting morphological operations to perform a two-stage upscale and downscale within one binary image representation. CTUDR is verified on a signal smoothing and classification task and compared with conventional techniques, such as the Moving Average, the Binomial, the Median, and the Savitzky Golay filters. Simulated EEG data with added white Gaussian noise is employed in the former, while cognitive conflict data obtained from a 3D object selection task is utilized in the latter. CTUDR outperforms its counterparts, scoring the best fitting error and correlation coefficient in signal smoothing while achieving the highest gain in Accuracy (0.7640%) and F-measure (0.7607%) when used as a smoothing filter for training data of EEGNet.

Detection of virtual reality motion sickness based on EEG using asymmetry of entropy and cross-frequency coupling

The existence of Virtual Reality Motion Sickness (VRMS) is a key factor restricting the further development of the VR industry, and the premise to solve this problem is to be able to accurately and effectively detect its occurrence. In view of the current lack of high-accuracy and effective detection methods, this paper proposes a VRMS detection method based on entropy asymmetry and cross-frequency coupling value asymmetry of EEG. First of all, the EEG of the four selected pairs of electrodes on the bilateral brain are subjected to Multivariate Variational Mode Decomposition (MVMD) respectively, and three types of entropy values on the low-frequency and high-frequency components are calculated, namely approximate entropy, fuzzy entropy and permutation entropy, as well as three types of phase-amplitude coupling features between the low-frequency and high-frequency components, namely the mean value, standard deviation and correlation coefficient; Secondly, the difference of the entropies and the cross-frequency coupling features between the left electrodes and the right electrodes are calculated; Finally, the final feature set are selected via t-test and fed into the SVM for classification, thus realizing the automatic detection of VRMS. The results show that the three classification indexes under this method, i.e., accuracy, sensitivity and specificity, reach 99.5 %, 99.3 % and 99.7 %, respectively, and the value of the area under the ROC curve reached 1, which proves that this method can be an effective indicator for detecting the occurrence of VRMS.

Brain-computer Interaction in the Smart Era

The brain-computer interface (BCI) system serves as a critical link between external output devices and the human brain. A monitored object's mental state, sensory cognition, and even higher cognition are reflected in its electroencephalography (EEG) signal. Nevertheless, unprocessed EEG signals are frequently contaminated with a variety of artifacts, rendering the analysis and elimination of impurities from the collected EEG data exceedingly challenging, not to mention the manual adjustment thereof. Over the last few decades, the rapid advancement of artificial intelligence (AI) technology has contributed to the development of BCI technology. Algorithms derived from AI and machine learning have significantly enhanced the ability to analyze and process EEG electrical signals, thereby expanding the range of potential interactions between the human brain and computers. As a result, the present BCI technology with the help of AI can assist physicians in gaining a more comprehensive understanding of their patients' physical and psychological status, thereby contributing to improvements in their health and quality of life.

Enhancing EEG-Based MI-BCIs with Class-Specific and Subject-Specific Features Detected by Neural Manifold Analysis

This paper presents an innovative approach leveraging Neuronal Manifold Analysis of EEG data to identify specific time intervals for feature extraction, effectively capturing both class-specific and subject-specific characteristics. Different pipelines were constructed and employed to extract distinctive features within these intervals, specifically for motor imagery (MI) tasks. The methodology was validated using the Graz Competition IV datasets 2A (four-class) and 2B (two-class) motor imagery classification, demonstrating an improvement in classification accuracy that surpasses state-of-the-art algorithms designed for MI tasks. A multi-dimensional feature space, constructed using NMA, was built to detect intervals that capture these critical characteristics, which led to significantly enhanced classification accuracy, especially for individuals with initially poor classification performance. These findings highlight the robustness of this method and its potential to improve classification performance in EEG-based MI-BCI systems.

Efficient Multi-View Graph Convolutional Network with Self-Attention for Multi-Class Motor Imagery Decoding

Research on electroencephalogram-based motor imagery (MI-EEG) can identify the limbs of subjects that generate motor imagination by decoding EEG signals, which is an important issue in the field of brain-computer interface (BCI). Existing deep-learning-based classification methods have not been able to entirely employ the topological information among brain regions, and thus, the classification performance needs further improving. In this paper, we propose a multi-view graph convolutional attention network (MGCANet) with residual learning structure for multi-class MI decoding. Specifically, we design a multi-view graph convolution spatial feature extraction method based on the topological relationship of brain regions to achieve more comprehensive information aggregation. During the modeling, we build an adaptive weight fusion (Awf) module to adaptively merge feature from different brain views to improve classification accuracy. In addition, the self-attention mechanism is introduced for feature selection to expand the receptive field of EEG signals to global dependence and enhance the expression of important features. The proposed model is experimentally evaluated on two public MI datasets and achieved a mean accuracy of 78.26% (BCIC IV 2a dataset) and 73.68% (OpenBMI dataset), which significantly outperforms representative comparative methods in classification accuracy. Comprehensive experiment results verify the effectiveness of our proposed method, which can provide novel perspectives for MI decoding.

EEG-based Signatures of Schizophrenia, Depression, and Aberrant Aging: A Supervised Machine Learning Investigation

BACKGROUND

Electroencephalography (EEG) is a noninvasive, cost-effective, and robust tool, which directly measures in vivo neuronal mass activity with high temporal resolution. Combined with state-of-the-art machine learning (ML) techniques, EEG recordings could potentially yield in silico biomarkers of severe mental disorders.

HYPOTHESIS

Pathological and physiological aging processes influence the electrophysiological signatures of schizophrenia (SCZ) and major depressive disorder (MDD).

STUDY DESIGN

From a single-center cohort (N = 735, 51.6% male) comprising healthy control individuals (HC, N = 245) and inpatients suffering from SCZ (N = 250) or MDD (N = 240), we acquired resting-state 19 channel-EEG recordings. Using repeated nested cross-validation, support vector machine models were trained to (1) classify patients with SCZ or MDD and HC individuals and (2) predict age in HC individuals. The age model was applied to patient groups to calculate Electrophysiological Age Gap Estimation (EphysAGE) as the difference between predicted and chronological age. The links between EphysAGE, diagnosis, and medication were then further explored.

STUDY RESULTS

The classification models robustly discriminated SCZ from HC (balanced accuracy, BAC = 72.7%, P < .001), MDD from HC (BAC = 67.0%, P < .001), and SCZ from MDD individuals (BAC = 63.2%, P < .001). Notably, central alpha (8-11 Hz) power decrease was the most consistently predictive feature for SCZ and MDD. Higher EphysAGE was associated with an increased likelihood of being misclassified as SCZ in HC and MDD (ρHC = 0.23, P < .001; ρMDD = 0.17, P = .01).

CONCLUSIONS

ML models can extract electrophysiological signatures of MDD and SCZ for potential clinical use. However, the impact of aging processes on diagnostic separability calls for timely application of such models, possibly in early recognition settings.

Quantifying instability in neurological disorders EEG based on phase space DTM function

Classifying individuals with neurological disorders and healthy subjects using EEG is a crucial area of research. The current feature extraction approach focuses on the frequency domain features in each of the EEG frequency bands and functional brain networks. In recent years, researchers have discovered and extensively studied stability differences in the electroencephalograms (EEG) of patients with neurological disorders. Based on this, this paper proposes a feature descriptor to characterize EEG instability. The proposed method starts by forming a signal point cloud through Phase Space Reconstruction (PSR). Subsequently, a pseudo-metric space is constructed, and pseudo-distances are calculated based on the consistent measure of the point cloud. Finally, Distance to Measure (DTM) Function are generated to replace the distance function in the original metric space. We calculated the relative distances in the point cloud by measuring signal similarity and, based on this, summarized the point cloud structures formed by EEG with different stabilities after PSR. This process demonstrated that Multivariate Kernel Density Estimation (MKDE) based on a Gaussian kernel can effectively separate the mappings of different stable components within the signal in the phase space. The two average DTM values are then proposed as feature descriptors for EEG instability.In the validation phase, the proposed feature descriptor is tested on three typical neurological disorders: epilepsy, Alzheimer's disease, and Parkinson's disease, using the Bonn dataset, CHB-MIT, the Florida State University dataset, and the Iowa State University dataset. DTM values are used as feature inputs for four different machine learning classifiers, and The results show that the best classification accuracy of the proposed method reaches 98.00 %, 96.25 %, 96.71 % and 95.34 % respectively, outperforming commonly used nonlinear descriptors. Finally, the proposed method is tested and analyzed using noisy signals, demonstrating its robustness compared to other methods.

SHAP value-based ERP analysis (SHERPA): Increasing the sensitivity of EEG signals with explainable AI methods

Conventionally, event-related potential (ERP) analysis relies on the researcher to identify the sensors and time points where an effect is expected. However, this approach is prone to bias and may limit the ability to detect unexpected effects or to investigate the full range of the electroencephalography (EEG) signal. Data-driven approaches circumvent this limitation, however, the multiple comparison problem and the statistical correction thereof affect both the sensitivity and specificity of the analysis. In this study, we present SHERPA - a novel approach based on explainable artificial intelligence (XAI) designed to provide the researcher with a straightforward and objective method to find relevant latency ranges and electrodes. SHERPA is comprised of a convolutional neural network (CNN) for classifying the conditions of the experiment and SHapley Additive exPlanations (SHAP) as a post hoc explainer to identify the important temporal and spatial features. A classical EEG face perception experiment is employed to validate the approach by comparing it to the established researcher- and data-driven approaches. Likewise, SHERPA identified an occipital cluster close to the temporal coordinates for the N170 effect expected. Most importantly, SHERPA allows quantifying the relevance of an ERP for a psychological mechanism by calculating an "importance score". Hence, SHERPA suggests the presence of a negative selection process at the early and later stages of processing. In conclusion, our new method not only offers an analysis approach suitable in situations with limited prior knowledge of the effect in question but also an increased sensitivity capable of distinguishing neural processes with high precision.

Simultaneous EEG-fNIRS Data Classification Through Selective Channel Representation and Spectrogram Imaging

The integration of electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) can facilitate the advancement of brain-computer interfaces (BCIs). However, existing research in this domain has grappled with the challenge of the efficient selection of features, resulting in the underutilization of the temporal richness of EEG and the spatial specificity of fNIRS data.To effectively address this challenge, this study proposed a deep learning architecture called the multimodal DenseNet fusion (MDNF) model that was trained on two-dimensional (2D) EEG data images, leveraging advanced feature extraction techniques. The model transformed EEG data into 2D images using a short-time Fourier transform, applied transfer learning to extract discriminative features, and consequently integrated them with fNIRS-derived spectral entropy features. This approach aimed to bridge existing gaps in EEG-fNIRS-based BCI research by enhancing classification accuracy and versatility across various cognitive and motor imagery tasks.Experimental results on two public datasets demonstrated the superiority of our model over existing state-of-the-art methods.Thus, the high accuracy and precise feature utilization of the MDNF model demonstrates the potential in clinical applications for neurodiagnostics and rehabilitation, thereby paving the method for patient-specific therapeutic strategies.

Machine-learning-based models assist the prediction of pulmonary embolism in autoimmune diseases: A retrospective, multicenter study

BACKGROUND

Pulmonary embolism (PE) is a severe and acute cardiovascular syndrome with high mortality among patients with autoimmune inflammatory rheumatic diseases (AIIRDs). Accurate prediction and timely intervention play a pivotal role in enhancing survival rates. However, there is a notable scarcity of practical early prediction and risk assessment systems of PE in patients with AIIRD.

METHODS

In the training cohort, 60 AIIRD with PE cases and 180 age-, gender-, and disease-matched AIIRD non-PE cases were identified from 7254 AIIRD cases in Tongji Hospital from 2014 to 2022. Univariable logistic regression (LR) and least absolute shrinkage and selection operator (LASSO) were used to select the clinical features for further training with machine learning (ML) methods, including random forest (RF), support vector machines (SVM), neural network (NN), logistic regression (LR), gradient boosted decision tree (GBDT), classification and regression trees (CART), and C5.0 models. The performances of these models were subsequently validated using a multicenter validation cohort.

RESULTS

In the training cohort, 24 and 13 clinical features were selected by univariable LR and LASSO strategies, respectively. The five ML models (RF, SVM, NN, LR, and GBDT) showed promising performances, with an area under the receiver operating characteristic (ROC) curve (AUC) of 0.962-1.000 in the training cohort and 0.969-0.999 in the validation cohort. CART and C5.0 models achieved AUCs of 0.850 and 0.932, respectively, in the training cohort. Using D-dimer as a pre-screening index, the refined C5.0 model achieved an AUC exceeding 0.948 in the training cohort and an AUC above 0.925 in the validation cohort. These results markedly outperformed the use of D-dimer levels alone.

CONCLUSION

ML-based models are proven to be precise for predicting the onset of PE in patients with AIIRD exhibiting clinical suspicion of PE.

TRIAL REGISTRATION

Chictr.org.cn : ChiCTR2200059599.

GRU-powered sleep stage classification with permutation-based EEG channel selection

We present a new approach to classifying the sleep stage that incorporates a computationally inexpensive method based on permutations for channel selection and takes advantage of deep learning power, specifically the gated recurrent unit (GRU) model, along with other deep learning methods. By systematically permuting the electroencephalographic (EEG) channels, different combinations of EEG channels are evaluated to identify the most informative subset for the classification of the 5-class sleep stage. For analysis, we used an EEG dataset that was collected at the International Institute for Integrative Sleep Medicine (WPI-IIIS) at the University of Tsukuba in Japan. The results of these explorations provide many new insights such as the (1) drastic decrease in performance when channels are fewer than 3, (2) 3-random channels selected by permutation provide the same or better prediction than the 3 channels recommended by the American Academy of Sleep Medicine (AASM), (3) N1 class suffers the most in prediction accuracy as the channels drop from 128 to 3 random or 3 AASM, and (4) no single channel provides acceptable levels of accuracy in the prediction of 5 classes. The results obtained show the GRU's ability to retain essential temporal information from EEG data, which allows capturing the underlying patterns associated with each sleep stage effectively. Using permutation-based channel selection, we enhance or at least maintain as high model efficiency as when using high-density EEG, incorporating only the most informative EEG channels.

HiRENet: Novel convolutional neural network architecture using Hilbert-transformed and raw electroencephalogram (EEG) for subject-independent emotion classification

BACKGROUND AND OBJECTIVES

Convolutional neural networks (CNNs) are the most widely used deep-learning framework for decoding electroencephalograms (EEGs) due to their exceptional ability to extract hierarchical features from high-dimensional EEG data. Traditionally, CNNs have primarily utilized multi-channel raw EEG data as the input tensor; however, the performance of CNN-based EEG decoding may be enhanced by incorporating phase information alongside amplitude information.

METHODS

This study introduces a novel CNN architecture called the Hilbert-transformed (HT) and raw EEG network (HiRENet), which incorporates both raw and HT EEG as inputs. This concurrent use of HT and raw EEG aims to integrate phase information with existing amplitude information, potentially offering a more comprehensive reflection of functional connectivity across various brain regions. The HiRENet model was developed using two CNN frameworks: ShallowFBCSPNet and a CNN with a residual block (ResCNN). The performance of the HiRENet model was assessed using a lab-made EEG database to classify human emotions, comparing three input modalities: raw EEG, HT EEG, and a combination of both signals. Additionally, the computational complexity was evaluated to validate the computational efficiency of the ResCNN design.

RESULTS

The HiRENet model based on ResCNN achieved the highest classification accuracy, with 86.03% for valence and 84.01% for arousal classifications, surpassing traditional CNN methodologies. Considering computational efficiency, ResCNN demonstrated superiority over ShallowFBCSPNet in terms of speed and inference time, despite having a higher parameter count.

CONCLUSION

Our experimental results showed that the proposed HiRENet can be potentially used as a new option to improve the overall performance for deep learning-based EEG decoding problems.

Complexity measures reveal age-dependent changes in electroencephalogram during working memory task

The alterations in electroencephalogram (EEG) signals are the complex outputs of functional factors, such as normal physiological aging, pathological process, which results in further cognitive decline. It is not clear that when brain aging initiates, but elderly people are vulnerable to be incipient of neurodegenerative diseases such as Alzheimer's disease. The EEG signals were recorded from 20 healthy middle age and 20 healthy elderly subjects while performing a working memory task. Higuchi's fractal dimension (HFD), Katz's fractal dimension (KFD), sample entropy and three Hjorth parameters were extracted to analyse the complexity of EEG signals. Four machine learning classifiers, multilayer perceptron (MLP), support vector machine (SVM), K-nearest neighbour (KNN), and logistic model tree (LMT) were employed to distinguish the EEG signals of middle age and elderly age groups. HFD, KFD and Hjorth complexity were found significantly correlated with age. MLP achieved the highest overall accuracy of 93.75%. For posterior region, the maximum accuracy of 92.50% was achieved using MLP. Since fractal dimension associated with the complexity of EEG signals, HFD, KFD and Hjorth complexity demonstrated the decreased complexity from middle age to elderly groups. The complexity features appear to be more appropriate indicators of monitoring EEG signal complexity in healthy aging.

Decoding electroencephalographic responses to visual stimuli compatible with electrical stimulation

Electrical stimulation of the visual nervous system could improve the quality of life of patients affected by acquired blindness by restoring some visual sensations, but requires careful optimization of stimulation parameters to produce useful perceptions. Neural correlates of elicited perceptions could be used for fast automatic optimization, with electroencephalography as a natural choice as it can be acquired non-invasively. Nonetheless, its low signal-to-noise ratio may hinder discrimination of similar visual patterns, preventing its use in the optimization of electrical stimulation. Our work investigates for the first time the discriminability of the electroencephalographic responses to visual stimuli compatible with electrical stimulation, employing a newly acquired dataset whose stimuli encompass the concurrent variation of several features, while neuroscience research tends to study the neural correlates of single visual features. We then performed above-chance single-trial decoding of multiple features of our newly crafted visual stimuli using relatively simple machine learning algorithms. A decoding scheme employing the information from multiple stimulus presentations was implemented, substantially improving our decoding performance, suggesting that such methods should be used systematically in future applications. The significance of the present work relies in the determination of which visual features can be decoded from electroencephalographic responses to electrical stimulation-compatible stimuli and at which granularity they can be discriminated. Our methods pave the way to using electroencephalographic correlates to optimize electrical stimulation parameters, thus increasing the effectiveness of current visual neuroprostheses.

Development and validation of a preoperative CT‑based radiomics nomogram to differentiate tuberculosis granulomas from lung adenocarcinomas: an external validation study

BACKGROUND

An accurate and non-invasive approach is urgently needed to distinguish tuberculosis granulomas from lung adenocarcinomas. This study aimed to develop and validate a nomogram based on contrast enhanced-compute tomography (CE-CT) to preoperatively differentiate tuberculosis granuloma from lung adenocarcinoma appearing as solitary pulmonary solid nodules (SPSN).

METHODS

This retrospective study analyzed 143 patients with lung adenocarcinoma (mean age: 62.4 ± 6.5 years; 54.5% female) and 137 patients with tuberculosis granulomas (mean age: 54.7 ± 8.2 years; 29.2% female) from two centers between March 2015 and June 2020. The training and internal validation cohorts included 161 and 69 patients (7:3 ratio) from center No.1, respectively. The external testing cohort included 50 patients from center No.2. Clinical factors and conventional radiological characteristics were analyzed to build independent predictors. Radiomics features were extracted from each CT-volume of interest (VOI). Feature selection was performed using univariate and multivariate logistic regression analysis, as well as the least absolute shrinkage and selection operator (LASSO) method. A clinical model was constructed with clinical factors and radiological findings. Individualized radiomics nomograms incorporating clinical data and radiomics signature were established to validate the clinical usefulness. The diagnostic performance was assessed using the receiver operating characteristic (ROC) curve analysis with the area under the receiver operating characteristic curve (AUC).

RESULTS

One clinical factor (CA125), one radiological characteristic (enhanced-CT value) and nine radiomics features were found to be independent predictors, which were used to establish the radiomics nomogram. The nomogram demonstrated better diagnostic efficacy than any single model, with respective AUC, accuracy, sensitivity, and specificity of 0.903, 0.857, 0.901, and 0.807 in the training cohort; 0.933, 0.884, 0.893, and 0.892 in the internal validation cohort; 0.914, 0.800, 0.937, and 0.735 in the external test cohort. The calibration curve showed a good agreement between prediction probability and actual clinical findings.

CONCLUSION

The nomogram incorporating clinical factors, radiological characteristics and radiomics signature provides additional value in distinguishing tuberculosis granuloma from lung adenocarcinoma in patients with a SPSN, potentially serving as a robust diagnostic strategy in clinical practice.

EEG-based emotion recognition systems; comprehensive study

Emotion recognition technology through EEG signal analysis is currently a fundamental concept in artificial intelligence. This recognition has major practical implications in emotional health care, human-computer interaction, and so on. This paper provides a comprehensive study of different methods for extracting electroencephalography (EEG) features for emotion recognition from four different perspectives, including time domain features, frequency domain features, time-frequency features, and nonlinear features. We summarize the current pattern recognition methods adopted in most related works, and with the rapid development of deep learning (DL) attracting the attention of researchers in this field, we pay more attention to deep learning-based studies and analyse the characteristics, advantages, disadvantages, and applicable scenarios. Finally, the current challenges and future development directions in this field were summarized. This paper can help novice researchers in this field gain a systematic understanding of the current status of emotion recognition research based on EEG signals and provide ideas for subsequent related research.

Advances of machine learning-assisted small extracellular vesicles detection strategy

Detection of extracellular vesicles (EVs), particularly small EVs (sEVs), is of great significance in exploring their physiological characteristics and clinical applications. The heterogeneity of sEVs plays a crucial role in distinguishing different types of cells and diseases. Machine learning, with its exceptional data processing capabilities, offers a solution to overcome the limitations of conventional detection methods for accurately classifying sEV subtypes and sources. Principal component analysis, linear discriminant analysis, partial least squares discriminant analysis, XGBoost, support vector machine, k-nearest neighbor, and deep learning, along with some combined methods such as principal component-linear discriminant analysis, have been successfully applied in the detection and identification of sEVs. This review focuses on machine learning-assisted detection strategies for cell identification and disease prediction via sEVs, and summarizes the integration of these strategies with surface-enhanced Raman scattering, electrochemistry, inductively coupled plasma mass spectrometry and fluorescence. The performance of different machine learning-based detection strategies is compared, and the advantages and limitations of various machine learning models are also evaluated. Finally, we discuss the merits and limitations of the current approaches and briefly outline the perspective of potential research directions in the field of sEV analysis based on machine learning.

BCI-controlled wheelchairs: end-users' perceptions, needs, and expectations, an interview-based study

PURPOSE

Brain-computer interface (BCI)-controlled wheelchairs have the potential to improve the independence of people with mobility impairments. The low uptake of BCI devices has been linked to a lack of knowledge among researchers of the needs of end-users that should influence BCI development.

MATERIALS AND METHODS

This study used semi-structured interviews to learn about the perceptions, needs, and expectations of spinal cord injury (SCI) patients with regards to a BCI-controlled wheelchair. Topics discussed in the interview include: paradigms, shared control, safety, robustness, channel selection, hardware, and experimental design. The interviews were recorded and then transcribed. Analysis was carried out using coding based on grounded theory principles.

RESULTS

The majority of participants had a positive view of BCI-controlled wheelchair technology and were willing to use the technology. Core issues were raised regarding safety, cost and aesthetics. Interview discussions were linked to state-of-the-art BCI technology. The results challenge the current reliance of researchers on the motor-imagery paradigm by suggesting end-users expect highly intuitive paradigms. There also needs to be a stronger focus on obstacle avoidance and safety features in BCI wheelchairs. Finally, the development of control approaches that can be personalized for individual users may be instrumental for widespread adoption of these devices.

CONCLUSIONS

This study, based on interviews with SCI patients, indicates that BCI-controlled wheelchairs are a promising assistive technology that would be well received by end-users. Recommendations for a more person-centered design of BCI controlled wheelchairs are made and clear avenues for future research are identified.

Single-channel qEEG characteristics distinguish delirium from no delirium, but not postoperative from non-postoperative delirium

OBJECTIVE

This exploratory study examined quantitative electroencephalography (qEEG) changes in delirium and the use of qEEG features to distinguish postoperative from non-postoperative delirium.

METHODS

This project was part of the DeltaStudy, a cross-sectional,multicenterstudy in Intensive Care Units (ICUs) and non-ICU wards. Single-channel (Fp2-Pz) four-minutes resting-state EEG was analyzed in 456 patients. After calculating 98 qEEG features per epoch, random forest (RF) classification was used to analyze qEEG changes in delirium and to test whether postoperative and non-postoperative delirium could be distinguished.

RESULTS

An area under the receiver operatingcharacteristic curve (AUC) of 0.76 (95% Confidence Interval (CI) 0.71-0.80) was found when classifying delirium with a sensitivity of 0.77 and a specificity of 0.63 at the optimal operating point. The classification of postoperative versus non-postoperative delirium resulted in an AUC of 0.50 (95%CI 0.38-0.61).

CONCLUSIONS

RF classification was able to discriminate delirium from no delirium with reasonable accuracy, while also identifying new delirium qEEG markers like autocorrelation and theta peak frequency. RF classification could not distinguish postoperative from non-postoperative delirium.

SIGNIFICANCE

Single-channel EEG differentiates between delirium and no delirium with reasonable accuracy. We found no distinct EEG profile for postoperative delirium, which may suggest that delirium is one entity, whether it develops postoperatively or not.

Editorial for the Special Issue "Machine Learning in Computer Vision and Image Sensing: Theory and Applications"

Machine learning (ML) models have experienced remarkable growth in their application for multimodal data analysis over the past decade [...].

Differentiating Epileptic and Psychogenic Non-Epileptic Seizures Using Machine Learning Analysis of EEG Plot Images

The treatment of epilepsy, the second most common chronic neurological disorder, is often complicated by the failure of patients to respond to medication. Treatment failure with anti-seizure medications is often due to the presence of non-epileptic seizures. Distinguishing non-epileptic from epileptic seizures requires an expensive and time-consuming analysis of electroencephalograms (EEGs) recorded in an epilepsy monitoring unit. Machine learning algorithms have been used to detect seizures from EEG, typically using EEG waveform analysis. We employed an alternative approach, using a convolutional neural network (CNN) with transfer learning using MobileNetV2 to emulate the real-world visual analysis of EEG images by epileptologists. A total of 5359 EEG waveform plot images from 107 adult subjects across two epilepsy monitoring units in separate medical facilities were divided into epileptic and non-epileptic groups for training and cross-validation of the CNN. The model achieved an accuracy of 86.9% (Area Under the Curve, AUC 0.92) at the site where training data were extracted and an accuracy of 87.3% (AUC 0.94) at the other site whose data were only used for validation. This investigation demonstrates the high accuracy achievable with CNN analysis of EEG plot images and the robustness of this approach across EEG visualization software, laying the groundwork for further subclassification of seizures using similar approaches in a clinical setting.

The path from task-specific to general purpose artificial intelligence for medical diagnostics: A bibliometric analysis

Artificial intelligence (AI) has revolutionized many fields, and its potential in healthcare has been increasingly recognized. Based on diverse data sources such as imaging, laboratory tests, medical records, and electrophysiological data, diagnostic AI has witnessed rapid development in recent years. A comprehensive understanding of the development status, contributing factors, and their relationships in the application of AI to medical diagnostics is essential to further promote its use in clinical practice. In this study, we conducted a bibliometric analysis to explore the evolution of task-specific to general-purpose AI for medical diagnostics. We used the Web of Science database to search for relevant articles published between 2010 and 2023, and applied VOSviewer, the R package Bibliometrix, and CiteSpace to analyze collaborative networks and keywords. Our analysis revealed that the field of AI in medical diagnostics has experienced rapid growth in recent years, with a focus on tasks such as image analysis, disease prediction, and decision support. Collaborative networks were observed among researchers and institutions, indicating a trend of global cooperation in this field. Additionally, we identified several key factors contributing to the development of AI in medical diagnostics, including data quality, algorithm design, and computational power. Challenges to progress in the field include model explainability, robustness, and equality, which will require multi-stakeholder, interdisciplinary collaboration to tackle. Our study provides a holistic understanding of the path from task-specific, mono-modal AI toward general-purpose, multimodal AI for medical diagnostics. With the continuous improvement of AI technology and the accumulation of medical data, we believe that AI will play a greater role in medical diagnostics in the future.

Intelligent assessment of atrial fibrillation gradation based on sinus rhythm electrocardiogram and baseline information

BACKGROUND

Atrial fibrillation (AF) is a progressive arrhythmia that significantly affects a patient's quality of life. The 4S-AF scheme is clinically recommended for AF management; however, the evaluation process is complex and time-consuming. This renders its promotion in primary medical institutions challenging. This retrospective study aimed to simplify the evaluation process and present an objective assessment model for AF gradation.

METHODS

In total, 189 12-lead electrocardiogram (ECG) recordings from 64 patients were included in this study. The data were annotated into two groups (mild and severe) according to the 4S-AF scheme. Using a preprocessed ECG during the sinus rhythm (SR), we obtained a synthesized vectorcardiogram (VCG). Subsequently, various features were calculated from both signals, and age, sex, and medical history were included as baseline characteristics. Different machine learning models, including support vector machines, random forests (RF), and logistic regression, were finally tested with a combination of feature selection techniques.

RESULTS

The proposed method demonstrated excellent performance in the classification of AF gradation. With an optimized feature set of VCG and baseline features, the RF model achieved accuracy, sensitivity, and specificity of 83.02 %, 80.56 %, and 88.24 %, respectively, under the inter-patient paradigm.

CONCLUSION

Our results demonstrate the value of physiological signals in AF gradation evaluation, and VCG signals were effective in identifying mild and severe AF. Considering its low computational complexity and high assessment performance, the proposed model is expected to serve as a useful prognostic tool for clinical AF management.

Support-vector classification of low-dose nitrous oxide administration with multi-channel EEG power spectra

Support-vector machines (SVMs) can potentially improve patient monitoring during nitrous oxide anaesthesia. By elucidating the effects of low-dose nitrous oxide on the power spectra of multi-channel EEG recordings, we quantified the degree to which these effects generalise across participants. In this single-blind, cross-over study, 32-channel EEG was recorded from 12 healthy participants exposed to 0, 20, 30 and 40% end-tidal nitrous oxide. Features of the delta-, theta-, alpha- and beta-band power were used within a 12-fold, participant-wise cross-validation framework to train and test two SVMs: (1) binary SVM classifying EEG during 0 or 40% exposure (chance = 50%); (2) multi-class SVM classifying EEG during 0, 20, 30 or 40% exposure (chance = 25%). Both the binary (accuracy 92%) and the multi-class (accuracy 52%) SVMs classified EEG recordings at rates significantly better than chance (p < 0.001 and p = 0.01, respectively). To determine the relative importance of frequency band features for classification accuracy, we systematically removed features before re-training and re-testing the SVMs. This showed the relative importance of decreased delta power and the frontal region. SVM classification identified that the most important effects of nitrous oxide were found in the delta band in the frontal electrodes that was consistent between participants. Furthermore, support-vector classification of nitrous oxide dosage is a promising method that might be used to improve patient monitoring during nitrous oxide anaesthesia.

Classification and automatic scoring of arousal intensity during sleep stages using machine learning

Arousal during sleep can result in sleep fragmentation and various physiological effects, impairing cognitive function and raising blood pressure and heart rate. However, the current definition of arousal has limitations in assessing both amplitude and duration, making it challenging to measure sleep fragmentation accurately. Moreover, there is inconsistency among inter-raters in arousal scoring, which renders it susceptible to subjective variability. Therefore, this study aims to identify a highly accurate classifier for each sleep stage by employing optimized feature selection and machine learning models. According to electroencephalography (EEG) signals during the arousal phase, the intensity level was categorized into four levels. For control, the non-arousal cases were used as level 0 and referred as sham arousal, resulting in five arousal intensity levels. Wavelet transform was applied to analyze sleep arousal to extract features from EEG. Based on these features, we classified arousal intensity levels through machine learning algorithms. Due to the different characteristics of EEG in each sleep stage, the classification model was optimized for the four sleep stages. Excluding sham arousals, a total of 13,532 arousal events were used. The lowest intensity in the entire data, level 1, was computed to be 3107, level 2 was 3384, level 3 was 3472, and the highest intensity of level 4 was 3,569. The optimized classification model for each sleep stage achieved an average sensitivity of 82.68%, specificity of 95.68%, and AUROC of 96.30%. The sensitivity of the control, arousal intensity level 0, was 83.07%, a 1.25% increase over the unoptimized model and a 14.22% increase over previous research. This study used machine learning techniques to develop classifiers for each sleep stage, improving the accuracy of arousal intensity classification. The classifiers showed high sensitivity and specificity and revealed the unique characteristics of arousal intensity during different sleep stages. These findings represent a novel approach to arousal research and have implications for developing more accurate predictive models in sleep research.

Predicting effluent quality parameters for wastewater treatment plant: A machine learning-based methodology

Wastewater Treatment Plants (WWTPs) present complex biochemical processes of high variability and difficult prediction. This study presents an innovative approach using Machine Learning (ML) models to predict wastewater quality parameters. In particular, the models are applied to datasets from both a simulated wastewater treatment plant (WWTP), using DHI WEST software (WEST WWTP), and a real-world WWTP database from Santa Catarina Brewery AMBEV, located in Lages/SC - Brazil (AMBEV WWTP). A distinctive aspect is the evaluation of predictive performance in continuous data scenarios and the impact of changes in WWTP operations on predictive model performance, including changes in plant layout. For both plants, three different scenarios were addressed, and the quality of predictions by random forest (RF), support vector machine (SVM), and multilayer perceptron (MLP) models were evaluated. The prediction quality by the MLP model reached an R2 of 0.72 for TN prediction in the WEST WWTP output, and the RF model better adapted to the real data of the AMBEV WWTP, despite the significant discrepancy observed between the real and the predicted data. Techniques such as Partial Dependence Plots (PDP) and Permutation Importance (PI) were used to assess the importance of features, particularly in the simulated WEST tool scenario, showing a strong correlation of prediction results with influent parameters related to nitrogen content. The results of this study highlight the importance of collecting and storing high-quality data and the need for information on changes in WWTP operation for predictive model performance. These contributions advance the understanding of predictive modeling for wastewater quality and provide valuable insights for future practice in wastewater treatment.

A low-cost, wireless, 4-channel EEG measurement system used in virtual reality environments

The combination of Virtual Reality (VR) technology and Electroencephalography (EEG) measurements has shown tremendous potential in the fields of psychology and neuroscience research. However, the majority of EEG measurement devices currently available are expensive, bulky, uncomfortable to wear, and difficult to integrate with VR headsets. These limitations have hindered the development of related research fields. This study describes a low-cost (60.07 USD), small-sized, wireless, high-precision, low-power consumption 4-channel EEG measurement system (NeuroVista) for frontal area EEG measurements, which can be used with a VR headset, enabling EEG measurements in VR environments. The system has an input-referred noise of less than 0.9480 μ V r m s , a common mode rejection ratio of over 96 dB, a measurement resolution of less than 0.1 μ V , a bandwidth of 0.5 ∼ 45 Hz, and works at a sampling rate of 250 Hz. It also supports metal dry electrodes and includes a built-in analog bandpass filter, right-leg drive circuit, and built-in digital lowpass filter and notch filter, which can reduce noise during measurement. Researchers can reconstruct the electrode system to measure regions of interest according to their needs.

Research on the classification of early-stage brain edema by combining intrinsic optical signal imaging and laser speckle contrast imaging

The early detection and pathological classification of brain edema are very important for symptomatic treatment. The dual-optical imaging system (DOIS) consists of intrinsic optical signal imaging (IOSI) and laser speckle contrast imaging (LSCI), which can acquire cerebral hemodynamic parameters of mice in real-time, including changes of oxygenated hemoglobin concentration ( Δ C HbO 2 ), deoxyhemoglobin concentration (ΔCHbR) and relative cerebral blood flow (rCBF) within the field of view. The slope sum of Δ C HbO 2 , ΔCHbR and rCBF was proposed to classify vasogenic edema (VE) and cytotoxic edema (CE). The slope sum values in the VE and CE group remain statistically different and the classification results provide higher accuracy of more than 93% for early brain edema detection. In conclusion, the differences of hemodynamic parameters between VE and CE in the early stage were revealed and the method helps in the classification of early brain edema.

Inter-hospital moderate and advanced Alzheimer's disease detection through convolutional neural networks

Electroencephalography (EEG) has been a fundamental technique in the identification of health conditions since its discovery. This analysis specifically centers on machine learning (ML) and deep learning (DL) methodologies designed for the analysis of electroencephalogram (EEG) data to categorize individuals with Alzheimer's Disease (AD) into two groups: Moderate or Advanced Alzheimer's dementia. Our study is based on a comprehensive database comprising 668 volunteers from 5 different hospitals, collected over a decade. This diverse dataset enables better training and validation of our results. Among the methods evaluated, the CNN (deep learning) approach outperformed others, achieving a remarkable classification accuracy of 97.45% for patients with Moderate Alzheimer's Dementia (ADM) and 97.03% for patients with Advanced Alzheimer's Dementia (ADA). Importantly, all the compared methods were rigorously assessed under identical conditions. The proposed DL model, specifically CNN, effectively extracts time domain features from EEG data in time, resulting in a significant reduction in learnable parameters and data redundancy.

Predicting rock-paper-scissors choices based on single-trial EEG signals

Decision prediction based on neurophysiological signals is of great application value in many real-life situations, especially in human-AI collaboration or counteraction. Single-trial analysis of electroencephalogram (EEG) signals is a very valuable step in the development of an online decision-prediction system. However, previous EEG-based decision-prediction methods focused mainly on averaged EEG signals of all decision-making trials to predict an individual's general decision tendency (e.g., risk seeking or aversion) over a period rather than on a specific decision response in a single trial. In the present study, we used a rock-paper-scissors game, which is a common multichoice decision-making task, to explore how to predict participants' single-trial choice with EEG signals. Forty participants, comprising 20 females and 20 males, played the game with a computer player for 330 trials. Considering that the decision-making process of this game involves multiple brain regions and neural networks, we proposed a new algorithm named common spatial pattern-attractor metagene (CSP-AM) to extract CSP features from different frequency bands of EEG signals that occurred during decision making. The results showed that a multilayer perceptron classifier achieved an accuracy significantly exceeding the chance level among 88.57% (31 of 35) of participants, verifying the classification ability of CSP features in multichoice decision-making prediction. We believe that the CSP-AM algorithm could be used in the development of proactive AI systems.

Exploring Convolutional Neural Network Architectures for EEG Feature Extraction

The main purpose of this paper is to provide information on how to create a convolutional neural network (CNN) for extracting features from EEG signals. Our task was to understand the primary aspects of creating and fine-tuning CNNs for various application scenarios. We considered the characteristics of EEG signals, coupled with an exploration of various signal processing and data preparation techniques. These techniques include noise reduction, filtering, encoding, decoding, and dimension reduction, among others. In addition, we conduct an in-depth analysis of well-known CNN architectures, categorizing them into four distinct groups: standard implementation, recurrent convolutional, decoder architecture, and combined architecture. This paper further offers a comprehensive evaluation of these architectures, covering accuracy metrics, hyperparameters, and an appendix that contains a table outlining the parameters of commonly used CNN architectures for feature extraction from EEG signals.

Dyslexia, the Amsterdam Way

The current aim is to illustrate our research on dyslexia conducted at the Developmental Psychology section of the Department of Psychology, University of Amsterdam, in collaboration with the nationwide IWAL institute for learning disabilities (now RID). The collaborative efforts are institutionalized in the Rudolf Berlin Center. The first series of studies aimed at furthering the understanding of dyslexia using a gamified tool based on an artificial script. Behavioral measures were augmented with diffusion modeling in one study, and indices derived from the electroencephalogram were used in others. Next, we illustrated a series of studies aiming to assess individuals who struggle with reading and spelling using similar research strategies. In one study, we used methodology derived from the machine learning literature. The third series of studies involved intervention targeting the phonics of language. These studies included a network analysis that is now rapidly gaining prominence in the psychopathology literature. Collectively, the studies demonstrate the importance of letter-speech sound mapping and word decoding in the acquisition of reading. It was demonstrated that focusing on these abilities may inform the prediction, classification, and intervention of reading difficulties and their neural underpinnings. A final section examined dyslexia, conceived as a neurobiological disorder. This analysis converged on the conclusion that recent developments in the psychopathology literature inspired by the focus on research domain criteria and network analysis might further the field by staying away from longstanding debates in the dyslexia literature (single vs. a multiple deficit, category vs. dimension, disorder vs. lack of skill).

Neurophysiological measures and correlates of cognitive load in attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD) and dyslexia: A scoping review and research recommendations

Working memory is integral to a range of critical cognitive functions such as reasoning and decision-making. Although alterations in working memory have been observed in neurodivergent populations, there has been no review mapping how cognitive load is measured in common neurodevelopmental conditions such as attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD) and dyslexia. This scoping review explores the neurophysiological measures used to study cognitive load in these specific populations. Our findings highlight that electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) are the most frequently used methods, with a limited number of studies employing functional near-infrared spectroscopy (fNIRs), magnetoencephalography (MEG) or eye-tracking. Notably, eye-related measures are less commonly used, despite their prominence in cognitive load research among neurotypical individuals. The review also highlights potential correlates of cognitive load, such as neural oscillations in the theta and alpha ranges for EEG studies, blood oxygenation level-dependent (BOLD) responses in lateral and medial frontal brain regions for fMRI and fNIRS studies and eye-related measures such as pupil dilation and blink rate. Finally, critical issues for future studies are discussed, including the technical challenges associated with multimodal approaches, the possible impact of atypical features on cognitive load measures and balancing data richness with participant well-being. These insights contribute to a more nuanced understanding of cognitive load measurement in neurodivergent populations and point to important methodological considerations for future neuroscientific research in this area.

Classification of First-Episode Psychosis with EEG Signals: ciSSA and Machine Learning Approach

First-episode psychosis (FEP) typically marks the onset of severe psychiatric disorders and represents a critical period in the field of mental health. The early diagnosis of this condition is essential for timely intervention and improved clinical outcomes. In this study, the classification of FEP was investigated using the analysis of electroencephalography (EEG) signals and circulant spectrum analysis (ciSSA) sub-band signals. FEP poses a significant diagnostic challenge in the realm of mental health, and it is aimed at introducing a novel and effective approach for early diagnosis. To achieve this, the LASSO method was utilized to select the most significant features derived from entropy, frequency, and statistical-based characteristics obtained from ciSSA sub-band signals, as well as their hybrid combinations. Subsequently, a high-performance classification model has been developed using machine learning techniques, including ensemble, support vector machine (SVM), and artificial neural network (ANN) methods. The results of this study demonstrated that the hybrid features extracted from EEG signals' ciSSA sub-bands, in combination with the SVM method, achieved a high level of performance, with an area under curve (AUC) of 0.9893, an accuracy of 96.23%, a sensitivity of 0.966, a specificity of 0.956, a precision of 0.9667, and an F1 score of 0.9666. This has revealed the effectiveness of the ciSSA-based method for classifying FEP from EEG signals.

Efficacy of novel attention-based gated recurrent units transformer for depression detection using electroencephalogram signals

Purpose

Depression is a global challenge causing psychological and intellectual problems that require efficient diagnosis. Electroencephalogram (EEG) signals represent the functional state of the human brain and can help build an accurate and viable technique for the early prediction and treatment of depression.

Methods

An attention-based gated recurrent units transformer (AttGRUT) time-series model is proposed to efficiently identify EEG perturbations in depressive patients. Statistical, spectral and wavelet features were first extracted from the 60-channel EEG signal data. Then, two feature selection techniques, recursive feature elimination and the Boruta algorithm, both with Shapley additive explanations, were utilised for selecting essential features.

Results

The proposed model outperformed the two baseline and two hybrid time-series models-long short-term memory (LSTM), gated recurrent units (GRU), convolutional neural network-LSTM (CNN-LSTM), and CNN-GRU-achieving an accuracy of up to 98.67%. Feature selection considerably increased the performance across all time-series models.

Conclusion

Based on the obtained results, novel feature selection greatly affected the results of the baseline and hybrid time-series models. The proposed AttGRUT can be implemented and tested in other domains by using different modalities for prediction.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13755-022-00205-8.

Machine learning-based inverse design methods considering data characteristics and design space size in materials design and manufacturing: a review

In the last few decades, the influence of machine learning has permeated many areas of science and technology, including the field of materials science. This toolkit of data driven methods accelerated the discovery and production of new materials by accurately predicting the complicated physical processes and mechanisms that are not fully described by existing materials theories. However, the availability of a growing number of increasingly complex machine learning models confronts us with the question of "which machine learning algorithm to employ". In this review, we provide a comprehensive review of common machine learning algorithms used for materials design, as well as a guideline for selecting the most appropriate model considering the nature of the design problem. To this end, we classify the material design problems into four categories of: (i) the training data set being sufficiently large to capture the trend of design space (interpolation problem), (ii) a vast design space that cannot be explored thoroughly with the initial training data set alone (extrapolation problem), (iii) multi-fidelity datasets (small accurate dataset and large approximate dataset), and (iv) only a small dataset available. The most successful machine learning-based surrogate models and design approaches will be discussed for each case along with pertinent literature. This review focuses mostly on the use of ML algorithms for the inverse design of complicated composite structures, a topic that has received a lot of attention recently with the rise of additive manufacturing.

Classification of attention deficit/hyperactivity disorder based on EEG signals using a EEG-Transformer model∗

Objective. Attention-deficit/hyperactivity disorder (ADHD) is the most common neurodevelopmental disorder in adolescents that can seriously impair a person's attention function, cognitive processes, and learning ability. Currently, clinicians primarily diagnose patients based on the subjective assessments of the Diagnostic and Statistical Manual of Mental Disorders-5, which can lead to delayed diagnosis of ADHD and even misdiagnosis due to low diagnostic efficiency and lack of well-trained diagnostic experts. Deep learning of electroencephalogram (EEG) signals recorded from ADHD patients could provide an objective and accurate method to assist physicians in clinical diagnosis.Approach. This paper proposes the EEG-Transformer deep learning model, which is based on the attention mechanism in the traditional Transformer model, and can perform feature extraction and signal classification processing for the characteristics of EEG signals. A comprehensive comparison was made between the proposed transformer model and three existing convolutional neural network models.Main results. The results showed that the proposed EEG-Transformer model achieved an average accuracy of 95.85% and an average AUC value of 0.9926 with the fastest convergence speed, outperforming the other three models. The function and relationship of each module of the model are studied by ablation experiments. The model with optimal performance was identified by the optimization experiment.Significance. The EEG-Transformer model proposed in this paper can be used as an auxiliary tool for clinical diagnosis of ADHD, and at the same time provides a basic model for transferable learning in the field of EEG signal classification.

Overview of methods and available tools used in complex brain disorders

Complex brain disorders, including Alzheimer's dementia, sleep disorders, and epilepsy, are chronic conditions that have high prevalence individually and in combination, increasing mortality risk, and contributing to the socioeconomic burden of patients, their families and, their communities at large. Although some literature reviews have been conducted mentioning the available methods and tools used for supporting the diagnosis of complex brain disorders and processing different files, there are still limitations. Specifically, these research works have focused primarily on one single brain disorder, i.e., sleep disorders or dementia or epilepsy. Additionally, existing research initiatives mentioning some tools, focus mainly on one single type of data, i.e., electroencephalography (EEG) signals or actigraphies or Magnetic Resonance Imaging, and so on. To tackle the aforementioned limitations, this is the first study conducting a comprehensive literature review of the available methods used for supporting the diagnosis of multiple complex brain disorders, i.e., Alzheimer's dementia, sleep disorders, epilepsy. Also, to the best of our knowledge, we present the first study conducting a comprehensive literature review of all the available tools, which can be exploited for processing multiple types of data, including EEG, actigraphies, and MRIs, and receiving valuable forms of information which can be used for differentiating people in a healthy control group and patients suffering from complex brain disorders. Additionally, the present study highlights both the benefits and limitations of the existing available tools.

Transformers for autonomous recognition of psychiatric dysfunction via raw and imbalanced EEG signals

Early identification of mental disorders, based on subjective interviews, is extremely challenging in the clinical setting. There is a growing interest in developing automated screening tools for potential mental health problems based on biological markers. Here, we demonstrate the feasibility of an AI-powered diagnosis of different mental disorders using EEG data. Specifically, this work aims to classify different mental disorders in the following ecological context accurately: (1) using raw EEG data, (2) collected during rest, (3) during both eye open, and eye closed conditions, (4) at short 2-min duration, (5) on participants with different psychiatric conditions, (6) with some overlapping symptoms, and (7) with strongly imbalanced classes. To tackle this challenge, we designed and optimized a transformer-based architecture, where class imbalance is addressed through focal loss and class weight balancing. Using the recently released TDBRAIN dataset (n= 1274 participants), our method classifies each participant as either a neurotypical or suffering from major depressive disorder (MDD), attention deficit hyperactivity disorder (ADHD), subjective memory complaints (SMC), or obsessive-compulsive disorder (OCD). We evaluate the performance of the proposed architecture on both the window-level and the patient-level. The classification of the 2-min raw EEG data into five classes achieved a window-level accuracy of 63.2% and 65.8% for open and closed eye conditions, respectively. When the classification is limited to three main classes (MDD, ADHD, SMC), window level accuracy improved to 75.1% and 69.9% for eye open and eye closed conditions, respectively. Our work paves the way for developing novel AI-based methods for accurately diagnosing mental disorders using raw resting-state EEG data.

Enhancing Performance of Convolutional Neural Network-Based Epileptic Electroencephalogram Diagnosis by Asymmetric Stochastic Resonance

Epilepsy is a chronic disorder that leads to transient neurological dysfunction and is clinically diagnosed primarily by electroencephalography. Several intelligent systems have been proposed to automatically detect seizures, among which deep convolutional neural networks (CNNs) have shown better performance than traditional machine-learning algorithms. Owing to artifacts and noise, the raw electroencephalogram (EEG) must be preprocessed to improve the signal-to-noise ratio prior to being fed into the CNN classifier. However, because of the spectrum overlapping of uncontrollable noise with EEG, traditional filters cause information loss in EEG; thus, the potential of classifiers cannot be fully exploited. In this study, we propose a stochastic resonance-effect-based EEG preprocessing module composed of three asymmetrical overdamped bistable systems in parallel. By setting different asymmetries for the three parallel units, the inherent noise can be transferred to the different spectral components of the EEG through the asymmetric stochastic resonance effect. In this process, the proposed preprocessing module not only avoids the loss of information of EEG but also provides a CNN with high-quality EEG of diversified frequency information to enhance its performance. By combining the proposed preprocessing module with a residual neural network, we developed an intelligent diagnostic system for predicting seizure onset. The developed system achieved an average sensitivity of 98.96% on the CHB-MIT dataset and 95.45% on the Siena dataset, with a false prediction rate of 0.048/h and 0.033/h, respectively. In addition, a comparative analysis demonstrated the superiority of the developed diagnostic system with the proposed preprocessing module over other existing methods.

A deep domain adaptation framework with correlation alignment for EEG-based motor imagery classification

It is impractical to collect sufficient and well-labeled EEG data in Brain-computer interface because of the time-consuming data acquisition and costly annotation. Conventional classification methods reusing EEG data from different subjects and time periods (across domains) significantly decrease the classification accuracy of motor imagery. In this paper, we propose a deep domain adaptation framework with correlation alignment (DDAF-CORAL) to solve the problem of distribution divergence for motor imagery classification across domains. Specifically, a two-stage framework is adopted to extract deep features for raw EEG data. The distribution divergence caused by subjected-related and time-related variations is further minimized by aligning the covariance of the source and target EEG feature distributions. Finally, the classification loss and adaptation loss are optimized simultaneously to achieve sufficient discriminative classification performance and low feature distribution divergence. Extensive experiments on three EEG datasets demonstrate that our proposed method can effectively reduce the distribution divergence between the source and target EEG data. The results show that our proposed method delivers outperformance (an average classification accuracy of 92.9% for within-session, an average kappa value of 0.761 for cross-session, and an average classification accuracy of 83.3% for cross-subject) in two-class classification tasks compared to other state-of-the-art methods.

Electroencephalography Signal Processing: A Comprehensive Review and Analysis of Methods and Techniques

The electroencephalography (EEG) signal is a noninvasive and complex signal that has numerous applications in biomedical fields, including sleep and the brain-computer interface. Given its complexity, researchers have proposed several advanced preprocessing and feature extraction methods to analyze EEG signals. In this study, we analyze a comprehensive review of numerous articles related to EEG signal processing. We searched the major scientific and engineering databases and summarized the results of our findings. Our survey encompassed the entire process of EEG signal processing, from acquisition and pretreatment (denoising) to feature extraction, classification, and application. We present a detailed discussion and comparison of various methods and techniques used for EEG signal processing. Additionally, we identify the current limitations of these techniques and analyze their future development trends. We conclude by offering some suggestions for future research in the field of EEG signal processing.

A Light-Weight Artificial Neural Network for Recognition of Activities of Daily Living

Human activity recognition (HAR) is essential for the development of robots to assist humans in daily activities. HAR is required to be accurate, fast and suitable for low-cost wearable devices to ensure portable and safe assistance. Current computational methods can achieve accurate recognition results but tend to be computationally expensive, making them unsuitable for the development of wearable robots in terms of speed and processing power. This paper proposes a light-weight architecture for recognition of activities using five inertial measurement units and four goniometers attached to the lower limb. First, a systematic extraction of time-domain features from wearable sensor data is performed. Second, a small high-speed artificial neural network and line search method for cost function optimization are used for activity recognition. The proposed method is systematically validated using a large dataset composed of wearable sensor data from seven activities (sitting, standing, walking, stair ascent/descent, ramp ascent/descent) associated with eight healthy subjects. The accuracy and speed results are compared against methods commonly used for activity recognition including deep neural networks, convolutional neural networks, long short-term memory and convolutional-long short-term memory hybrid networks. The experiments demonstrate that the light-weight architecture can achieve a high recognition accuracy of 98.60%, 93.10% and 84.77% for seen data from seen subjects, unseen data from seen subjects and unseen data from unseen subjects, respectively, and an inference time of 85 μs. The results show that the proposed approach can perform accurate and fast activity recognition with a reduced computational complexity suitable for the development of portable assistive devices.

A comparative study of machine learning methods for classifying ERP scalp distribution

Objective. Machine learning (ML) methods are used in different fields for classification and regression purposes with different applications. These methods are also used with various non-invasive brain signals, including Electroencephalography (EEG) signals to detect some patterns in the brain signals. ML methods are considered critical tools for EEG analysis since could overcome some of the limitations in the traditional methods of EEG analysis such as Event-related potentials (ERPs) analysis. The goal of this paper was to apply ML classification methods on ERP scalp distribution to investigate the performance of these methods in identifying numerical information carried in different finger-numeral configurations (FNCs). FNCs in their three forms of montring, counting, and non-canonical counting are used for communication, counting, and doing arithmetic across the world between children and even adults. Studies have shown the relationship between perceptual and semantic processing of FNCs, and neural differences in visually identifying different types of FNCs.Approach.A publicly available 32-channel EEG dataset recorded for 38 participants while they were shown a picture of an FNC (i.e., three categories and four numbers of 1,2,3, and 4) was used. EEG data were pre-processed and ERP scalp distribution of different FNCs was classified across time by six ML methods, including support vector machine, linear discriminant analysis, naïve Bayes, decision tree, K-nearest neighbor, and neural network. The classification was conducted in two conditions: classifying all FNCs together (i.e., 12 classes) and classifying FNCs of each category separately (i.e., 4 classes).Results.The support vector machine had the highest classification accuracy for both conditions. For classifying all FNCs together, the K-nearest neighbor was the next in line; however, the neural network could retrieve numerical information from the FNCs for category-specific classification.Significance.The significance of this study is in exploring the application of multiple ML methods in recognizing numerical information contained in ERP scalp distribution of different finger-numeral configurations.

In search of a composite biomarker for chronic pain by way of EEG and machine learning: where do we currently stand?

Machine learning is becoming an increasingly common component of routine data analyses in clinical research. The past decade in pain research has witnessed great advances in human neuroimaging and machine learning. With each finding, the pain research community takes one step closer to uncovering fundamental mechanisms underlying chronic pain and at the same time proposing neurophysiological biomarkers. However, it remains challenging to fully understand chronic pain due to its multidimensional representations within the brain. By utilizing cost-effective and non-invasive imaging techniques such as electroencephalography (EEG) and analyzing the resulting data with advanced analytic methods, we have the opportunity to better understand and identify specific neural mechanisms associated with the processing and perception of chronic pain. This narrative literature review summarizes studies from the last decade describing the utility of EEG as a potential biomarker for chronic pain by synergizing clinical and computational perspectives.

Hybrid Deep Learning Approach for Stress Detection Using Decomposed EEG Signals

Stress has an impact, not only on a person's physical health, but also on the ability to perform at the workplace in daily life. The well-established relation between psychological stress and its pathogeneses highlights the need for detecting psychological stress early, in order to prevent disease advancement and to save human lives. Electroencephalography (EEG) signal recording tools are widely used to collect these psychological signals/brain rhythms in the form of electric waves. The aim of the current research was to apply automatic feature extraction to decomposed multichannel EEG recordings, in order to efficiently detect psychological stress. The traditional deep learning techniques, namely the convolution neural network (CNN), long short-term memory (LSTM), bidirectional long short-term memory (BiLSTM), gated recurrent unit (GRU) and recurrent neural network (RNN) models, have been frequently used for stress detection. A hybrid combination of these techniques may provide improved performance, and can handle long-term dependencies in non-linear brain signals. Therefore, this study proposed an integration of deep learning models, called DWT-based CNN, BiLSTM, and two layers of a GRU network, to extract features and classify stress levels. Discrete wavelet transform (DWT) analysis was used to remove the non-linearity and non-stationarity from multi-channel (14 channel) EEG recordings, and to decompose them into different frequency bands. The decomposed signals were utilized for automatic feature extraction using the CNN, and the stress levels were classified using BiLSTM and two layers of GRU. This study compared five combinations of the CNN, LSTM, BiLSTM, GRU and RNN models with the proposed model. The proposed hybrid model performed better in classification accuracy compared to the other models. Therefore, hybrid combinations are appropriate for the clinical intervention and prevention of mental and physical problems.

A low-cost robotic hand prosthesis with apparent haptic sense controlled by electroencephalographic signals

One of the biggest problems faced by amputees is obtaining a suitable low-cost prosthesis. To address this problem, the design and implementation of a transradial prosthesis controlled by electroencephalographic (EEG) signals was carried out. This prosthesis is an alternative to prostheses using electromyographic (EMG) signals, which are very complex and exhausting for the patient to execute. We collected EEG signal data using the Emotiv Insight Headset, which were then processed to control the movement of the prosthesis, known as the Zero Arm. Additionally, we incorporated Machine Learning algorithms to classify different types of objects and shapes. The prosthesis also features a haptic feedback system, which simulates the function of mechanoreceptors in the skin, providing the user with a sense of touch when using the prosthesis. Our research has yielded a viable and cost-effective prosthetic limb. We utilized 3D printing and easily obtainable servomotors and controllers, making the prosthesis affordable and accessible. Performance tests of the Zero Arm prosthesis have yielded promising results. The prosthesis demonstrated an average success rate of 86.67% across various tasks, indicating its reliability and effectiveness. Additionally, the prosthesis has an average recognition rate of 70% for different types of objects, a noteworthy accomplishment.

Autonomous Self-Adaptive and Self-Aware Optical Wireless Communication Systems

The future age of optical networks demands autonomous functions to optimize available resources. With autonomy, the communication network should be able to learn and adapt to the dynamic environment. Among the different autonomous tasks, this work considers building self-adaptive and self-awareness-free space optic (FSO) networks by exploiting advances in artificial intelligence. In this regard, we study the use of machine learning (ML) techniques to build self-adaptive and self-awareness FSO systems capable of classifying the modulation format/baud rate and predicting the number of channel impairments. The study considers four modulation formats and four baud rates applicable in current commercial FSO systems. Moreover, two main channel impairments are considered. The results show that the proposed ML algorithm is capable of achieving 100% classification accuracy for the considered modulation formats/baud rates even under harsh channel conditions. Moreover, the prediction accuracy of the channel impairments ranges between 71% and 100% depending on the predicted parameter type and channel conditions.

High performance clean versus artifact dry electrode EEG data classification using Convolutional Neural Network transfer learning

Objective

Convolutional Neural Networks (CNNs) are promising for artifact detection in electroencephalography (EEG) data, but require large amounts of data. Despite increasing use of dry electrodes for EEG data acquisition, dry electrode EEG datasets are sparse. We aim to develop an algorithm for clean versus artifact dry electrode EEG data classification using transfer learning.

Methods

Dry electrode EEG data were acquired in 13 subjects while physiological and technical artifacts were induced. Data were per 2-second segment labeled as clean or artifact and split in an 80% train and 20% test set. With the train set, we fine-tuned a pre-trained CNN for clean versus artifact wet electrode EEG data classification using 3-fold cross validation. The three fine-tuned CNNs were combined in one final clean versus artifact classification algorithm, in which the majority vote was used for classification. We calculated accuracy, F1-score, precision, and recall of the pre-trained CNN and fine-tuned algorithm when applied to unseen test data.

Results

The algorithm was trained on 0.40 million and tested on 0.17 million overlapping EEG segments. The pre-trained CNN had a test accuracy of 65.6%. The fine-tuned clean versus artifact classification algorithm had an improved test accuracy of 90.7%, F1-score of 90.2%, precision of 89.1% and recall of 91.2%.

Conclusions

Despite a relatively small dry electrode EEG dataset, transfer learning enabled development of a high performing CNN-based algorithm for clean versus artifact classification.

Significance

Development of CNNs for classification of dry electrode EEG data is challenging as dry electrode EEG datasets are sparse. Here, we show that transfer learning can be used to overcome this problem.