Automatic diagnosis of alcohol use disorder using EEG features

Explaining electroencephalogram channel and subband sensitivity for alcoholism detection

Alcoholism, a progressive loss of control over alcohol consumption, deteriorates mental and physical health over time. Automatic alcoholism detection can aid in early interventions and timely corrective actions. For this purpose, electroencephalogram (EEG) signals are investigated using explainable artificial intelligence (XAI) techniques to obtain biomarkers. EEG signals are decomposed into five frequency bands to generate vectors of band powers of all channels. These vectors are input to different machine learning models and their ensembles to detect alcoholism. The reliability and generalization of these models are investigated using three oversampling techniques: Synthetic minority over-sampling technique, Adaptive Synthetic Sampling data augmentation, and Gaussian. Comparative analysis using an open-source EEG dataset showed superior performance for artificial neural network (ANN), with an accuracy of 97.36% and an F1-score of 97.88%. The oversampling techniques further improved performance across various ML models, and SMOTE-based ANN outperformed with an accuracy of 97.93% and an F1-score of 97.99%. The best ANN model is investigated using three XAI techniques, Local Interpretable Model-agnostic Explanations (LIME), Submodular Pick LIME, and Morris sensitivity analysis, for explaining the sensitivity of EEG channels and bands in alcoholism detection. The frequency band explanations showed relatively higher performance using beta and gamma bands with 95.45% and 97.36% accuracy, respectively. The EEG channel explanations showed higher performance using biomarkers from parietal and central regions, providing 95.41% and 91.50% accuracy, respectively. A combined explanation of all three XAI techniques indicated that three from the parietal and three from the central regions are the most important for improving detection performance. This study validates the effectiveness of the ANN in detecting alcoholism using EEG signals and emphasizes the significance of frequency bands and EEG channels. These explanations could be applied for early detection and monitoring.

Predicting the Risk of Driving Under the Influence of Alcohol Using EEG-Based Machine Learning

Driving under the influence of alcohol (DUIA) is closely associated with alcohol use disorder (AUD). Our previous study on machine learning (ML) algorithms revealed a very high accuracy of decision trees with neuropsychological features in predicting the risk of DUIA despite limited data availability. Thus, this study aimed at comparing six well-known ML algorithms based on electroencephalographic (EEG) signals to differentiate adults with AUD and DUIA (AUD-DD) from those with AUD without DUIA (AUD-NDD) and controls. Fifteen AUD-DD and 10 AUD-NDD participants were recruited from a single tertiary referral center. Fourteen social drinkers without DUIA served as controls. Their EEG signals related to driving conditions were gathered using a VR headset with eight electrodes (F3, F4, Fz, C3, C4, Cz, P3, and P4). Based on the labeled features of EEG asymmetry and theta/beta ratio (TBR), comparisons between different algorithms were conducted. Fz and Cz electrodes exhibited differences in TBR across the three groups (all p < 0.02), while there were no significant differences between AUD-DD individuals and social drinkers. In contrast, asymmetries of between-group differences were not observed (all p > 0.09). K-nearest neighbors (KNN) with TBR showed the highest accuracy (83 %) in distinguishing AUD-DD individuals from controls, while logistic regression (LR), support vector machines (SVM), and naive Bayes (NB) with EEG asymmetric features demonstrated high accuracy in identifying DUIA (all 80 %) in AUD adults. LR, SVM, and NB with asymmetry may be employed in predicting DUIA among AUD adults, while KNN with TBR may be used for identifying DUIA in the general population.

Predicting Treatment-seeking Status for Alcohol Use Disorder Using Polygenic Scores and Machine Learning in a Deeply-Phenotyped Sample

Background

Few individuals with alcohol use disorder (AUD) receive treatment. Previous studies have shown drinking behavior, psychological problems, and substance dependence to predict treatment seeking. However, to date, no studies have incorporated polygenic scores (PGS), a measure of genetic risk for AUD.

Methods

Using the Yale-Penn sample, we identified 9,103 individuals diagnosed with DSM-IV AUD and indicated treatment-seeking status. We implemented a random forest (RF) model to predict treatment-seeking based on 91 clinically relevant phenotypes. We calculated AUD PGS for those with genetic data (African ancestry [AFR] n=3,192, European ancestry [EUR] n=3,553) and generated RF models for each ancestry group, first without and then with PGS. Lastly, we developed models stratified by age (< and ≥40 years old).

Results

66.6% reported treatment seeking (M age =40.0, 62.4% male). Across models, top predictors included years of alcohol use and related psychological problems, psychiatric diagnoses, and heart disease. In the models without PGS, we found 79.8% accuracy and 0.85 AUC for EUR and 75% and 0.78 for AFR; the addition of PGS did not substantially change these metrics. PGS was the 10 th most important predictor for EUR and 23 rd for AFR. In the age-stratified analysis, PGS ranked 8 th for <40 and 48 th for ≥40 in EUR ancestry, and it ranked 14 th for <40 and 24 th for ≥40 in the AFR sample.

Conclusion

Alcohol use, psychiatric issues, and comorbid medical disorders were predictors of treatment seeking. Incorporating PGS did not substantially alter performance, but was a more important predictor in younger individuals with AUD.

Highlights

While alcohol use problems are common, few individuals seek treatmentWe used machine learning in a deeply-phenotyped sample to predict treatment-seekingWe, for the first time, incorporated polygenic risk for alcohol use as a predictorAlcohol use variables, psychiatric issues, and medical problems were key predictors.

Machine learning and clinical neurophysiology

Clinical neurophysiology constructs a wealth of dynamic information pertaining to the integrity and function of both central and peripheral nervous systems. As with many technological fields, there has been an explosion of data in neurophysiology over recent years, and this requires considerable analysis by experts. Computational algorithms and especially advances in machine learning (ML) have the ability to assist with this task and potentially reveal hidden insights. In this update article, we will provide a brief overview where such technology is being applied in clinical neurophysiology and possible future directions.

Resting-state EEG, Substance use and Abstinence After Chronic use: A Systematic Review

Resting-state EEG reflects intrinsic brain activity and its alteration represents changes in cognition that are related to neuropathology. Thereby, it provides a way of revealing the neurocognitive mechanisms underpinning chronic substance use. In addition, it is documented that some neurocognitive functions can recover following sustained abstinence. We present a systematic review to synthesize how chronic substance use is associated with resting-state EEG alterations and whether these spontaneously recover from abstinence. A literature search in Medline, PsycINFO, Embase, CINAHL, Web of Science, and Scopus resulted in 4088 articles, of which 57 were included for evaluation. It covered the substance of alcohol (18), tobacco (14), cannabis (8), cocaine (6), opioids (4), methamphetamine (4), and ecstasy (4). EEG analysis methods included spectral power, functional connectivity, and network analyses. It was found that long-term substance use with or without substance use disorder diagnosis was associated with broad intrinsic neural activity alterations, which were usually expressed as neural hyperactivation and decreased neural communication between brain regions. Some studies found the use of alcohol, tobacco, cocaine, cannabis, and methamphetamine was positively correlated with these changes. These alterations can partly recover from abstinence, which differed between drugs and may reflect their neurotoxic degree. Moderating factors that may explain results inconsistency are discussed. In sum, resting-state EEG may act as a potential biomarker of neurotoxic effects of chronic substance use. Recovery effects awaits replication in larger samples with prolonged abstinence. Balanced sex ratio, enlarged sample size, advanced EEG analysis methods, and transparent reporting are recommended for future studies.

Hybrid classification model for eye state detection using electroencephalogram signals

The electroencephalography (EEG) signal is an essential source of Brain-Computer Interface (BCI) technology implementation. The BCI is nothing but a non-muscle communication medium among the external devices and the brain. The basic concept of BCI is to enable the interaction among the neurological ill patients to others with the help of brain signals. EEG signal classification is an essential requirement for various applications such as motor imagery classification, drug effects diagnosis, emotion classification, seizure prediction/detection, eye state prediction/detection, and so on. Thus, there is a need for an efficient classification model that can deal with the EEG datasets more adequately with better classification accuracy, which will further help in developing the automatic solution for the medical domain. In this paper, we have introduced a hybrid classification model for eye state detection using electroencephalogram (EEG) signals. This hybrid classification model has been evaluated with the other traditional machine learning models, eight classification models (Prepossessed + Hypertuned) and six state-of-the-art methods to assess its appropriateness and correctness. This proposed classification model establishes a machine learning-based hybrid model for the classification of eye state using EEG signals with greater exactness. It is also capable of solving the issue of outlier detection and removal to address the class imbalance problem, which will offer the solution toward building the robotic or smart machine-based solution for social well-being.

Effective Connectivity in Default Mode Network for Alcoholism Diagnosis

Alcohol Use Disorder (AUD) is a chronic relapsing brain disease characterized by excessive alcohol use, loss of control over alcohol intake, and negative emotional states under no alcohol consumption. The key factor in successful treatment of AUD is the accurate diagnosis for better medical and therapy management. Conventionally, for individuals to be diagnosed with AUD, certain criteria as outlined in the Diagnostic and Statistical Manual of Mental Disorders (DSM) should be met. However, this process is subjective in nature and could be misleading due to memory problems and dishonesty of some AUD patients. In this paper, an assessment scheme for objective diagnosis of AUD is proposed. For this purpose, EEG recording of 31 healthy controls and 31 AUD patients are used for the calculation of effective connectivity (EC) between the various regions of the brain Default Mode Network (DMN). The EC is estimated using partial directed coherence (PDC) which are then used as input to a 3D Convolutional Neural Network (CNN) for binary classification of AUD cases. Using 5-fold cross validation, the classification of AUD vs. HC effective connectivity matrices using the proposed 3D-CNN gives an accuracy of 87.85 ± 4.64 %. For further validation, 32 and 30 subjects are randomly selected for training and testing, respectively, giving 100% correct classification of all the testing subjects.

Predicting risk for Alcohol Use Disorder using longitudinal data with multimodal biomarkers and family history: a machine learning study

Predictive models have succeeded in distinguishing between individuals with Alcohol use Disorder (AUD) and controls. However, predictive models identifying who is prone to develop AUD and the biomarkers indicating a predisposition to AUD are still unclear. Our sample (n = 656) included offspring and non-offspring of European American (EA) and African American (AA) ancestry from the Collaborative Study of the Genetics of Alcoholism (COGA) who were recruited as early as age 12 and were unaffected at first assessment and reassessed years later as AUD (DSM-5) (n = 328) or unaffected (n = 328). Machine learning analysis was performed for 220 EEG measures, 149 alcohol-related single nucleotide polymorphisms (SNPs) from a recent large Genome-wide Association Study (GWAS) of alcohol use/misuse and two family history (mother DSM-5 AUD and father DSM-5 AUD) features using supervised, Linear Support Vector Machine (SVM) classifier to test which features assessed before developing AUD predict those who go on to develop AUD. Age, gender, and ancestry stratified analyses were performed. Results indicate significant and higher accuracy rates for the AA compared with the EA prediction models and a higher model accuracy trend among females compared with males for both ancestries. Combined EEG and SNP features model outperformed models based on only EEG features or only SNP features for both EA and AA samples. This multidimensional superiority was confirmed in a follow-up analysis in the AA age groups (12-15, 16-19, 20-30) and EA age group (16-19). In both ancestry samples, the youngest age group achieved higher accuracy score than the two other older age groups. Maternal AUD increased the model's accuracy in both ancestries' samples. Several discriminative EEG measures and SNPs features were identified, including lower posterior gamma, higher slow wave connectivity (delta, theta, alpha), higher frontal gamma ratio, higher beta correlation in the parietal area, and 5 SNPs: rs4780836, rs2605140, rs11690265, rs692854, and rs13380649. Results highlight the significance of sampling uniformity followed by stratified (e.g., ancestry, gender, developmental period) analysis, and wider selection of features, to generate better prediction scores allowing a more accurate estimation of AUD development.

Alterations of Functional Connectivity During the Resting State and Their Associations With Visual Memory in College Students Who Binge Drink

This study investigated the characteristics of neural oscillation and functional connectivity (FC) in college students engaging in binge drinking (BD) using resting-state electroencephalography (EEG). Also, the associations of visual memory, evaluated by the Rey-Osterrieth Complex Figure Test (RCFT), and neural oscillation with FC during the resting state were investigated. The BD (n = 35) and non-BD (n = 35) groups were selected based on scores of the Korean version of the Alcohol use disorders (AUDs) Identification Test and the Alcohol Use Questionnaire. EEG was performed for 6 min while the participants rested with eyes closed. The theta, lower-alpha, and upper alpha powers did not differ between the BD and non-BD groups. Concerning FC, the BD group exhibited stronger theta coherence than that of the non-BD group, and in the lower and upper alpha bands, the BD group showed stronger coherence in some areas but weaker coherence in others compared with the non-BD group. However, these significant results were not observed after Bonferroni correction. The BD group showed significantly lower delayed recall scores on the RCFT than did the non-BD group. A positive correlation between the left prefrontal-parietal-occipital midline connection and performance on the delayed recall of the RCFT was observed in the BD group. The present results could suggest that binge drinkers have alterations in brain FC, which may be related to their visual memory deficits.

Bi-Dimensional Approach Based on Transfer Learning for Alcoholism Pre-disposition Classification via EEG Signals

Recent statistics have shown that the main difficulty in detecting alcoholism is the unreliability of the information presented by patients with alcoholism; this factor confusing the early diagnosis and it can reduce the effectiveness of treatment. However, electroencephalogram (EEG) exams can provide more reliable data for analysis of this behavior. This paper proposes a new approach for the automatic diagnosis of patients with alcoholism and introduces an analysis of the EEG signals from a two-dimensional perspective according to changes in the neural activity, highlighting the influence of high and low-frequency signals. This approach uses a two-dimensional feature extraction method, as well as the application of recent Computer Vision (CV) techniques, such as Transfer Learning with Convolutional Neural Networks (CNN). The methodology to evaluate our proposal used 21 combinations of the traditional classification methods and 84 combinations of recent CNN architectures used as feature extractors combined with the following classical classifiers: Gaussian Naive Bayes, K-Nearest Neighbor (k-NN), Multilayer Perceptron (MLP), Random Forest (RF) and Support Vector Machine (SVM). CNN MobileNet combined with SVM achieved the best results in Accuracy (95.33%), Precision (95.68%), F1-Score (95.24%), and Recall (95.00%). This combination outperformed the traditional methods by up to 8%. Thus, this approach is applicable as a classification stage for computer-aided diagnoses, useful for the triage of patients, and clinical support for the early diagnosis of this disease.

EEG sleep stages identification based on weighted undirected complex networks

BACKGROUND AND OBJECTIVE

Sleep scoring is important in sleep research because any errors in the scoring of the patient's sleep electroencephalography (EEG) recordings can cause serious problems such as incorrect diagnosis, medication errors, and misinterpretations of patient's EEG recordings. The aim of this research is to develop a new automatic method for EEG sleep stages classification based on a statistical model and weighted brain networks.

METHODS

Each EEG segment is partitioned into a number of blocks using a sliding window technique. A set of statistical features are extracted from each block. As a result, a vector of features is obtained to represent each EEG segment. Then, the vector of features is mapped into a weighted undirected network. Different structural and spectral attributes of the networks are extracted and forwarded to a least square support vector machine (LS-SVM) classifier. At the same time the network's attributes are also thoroughly investigated. It is found that the network's characteristics vary with their sleep stages. Each sleep stage is best represented using the key features of their networks.

RESULTS

In this paper, the proposed method is evaluated using two datasets acquired from different channels of EEG (Pz-Oz and C3-A2) according to the R&K and the AASM without pre-processing the original EEG data. The obtained results by the LS-SVM are compared with those by Naïve, k-nearest and a multi-class-SVM. The proposed method is also compared with other benchmark sleep stages classification methods. The comparison results demonstrate that the proposed method has an advantage in scoring sleep stages based on single channel EEG signals.

CONCLUSIONS

An average accuracy of 96.74% is obtained with the C3-A2 channel according to the AASM standard, and 96% with the Pz-Oz channel based on the R&K standard.

Combining Inter-Subject Modeling with a Subject-Based Data Transformation to Improve Affect Recognition from EEG Signals

Existing correlations between features extracted from Electroencephalography (EEG) signals and emotional aspects have motivated the development of a diversity of EEG-based affect detection methods. Both intra-subject and inter-subject approaches have been used in this context. Intra-subject approaches generally suffer from the small sample problem, and require the collection of exhaustive data for each new user before the detection system is usable. On the contrary, inter-subject models do not account for the personality and physiological influence of how the individual is feeling and expressing emotions. In this paper, we analyze both modeling approaches, using three public repositories. The results show that the subject's influence on the EEG signals is substantially higher than that of the emotion and hence it is necessary to account for the subject's influence on the EEG signals. To do this, we propose a data transformation that seamlessly integrates individual traits into an inter-subject approach, improving classification results.

A review on EEG-based methods for screening and diagnosing alcohol use disorder

The screening test for alcohol use disorder (AUD) patients has been of subjective nature and could be misleading in particular cases such as a misreporting the actual quantity of alcohol intake. Although the neuroimaging modality such as electroencephalography (EEG) has shown promising research results in achieving objectivity during the screening and diagnosis of AUD patients. However, the translation of these findings for clinical applications has been largely understudied and hence less clear. This study advocates the use of EEG as a diagnostic and screening tool for AUD patients that may help the clinicians during clinical decision making. In this context, a comprehensive review on EEG-based methods is provided including related electrophysiological techniques reported in the literature. More specifically, the EEG abnormalities associated with the conditions of AUD patients are summarized. The aim is to explore the potentials of objective techniques involving quantities/features derived from resting EEG, event-related potentials or event-related oscillations data.

An EEG-based machine learning method to screen alcohol use disorder

Screening alcohol use disorder (AUD) patients has been challenging due to the subjectivity involved in the process. Hence, robust and objective methods are needed to automate the screening of AUD patients. In this paper, a machine learning method is proposed that utilized resting-state electroencephalography (EEG)-derived features as input data to classify the AUD patients and healthy controls and to perform automatic screening of AUD patients. In this context, the EEG data were recorded during 5 min of eyes closed and 5 min of eyes open conditions. For this purpose, 30 AUD patients and 15 aged-matched healthy controls were recruited. After preprocessing the EEG data, EEG features such as inter-hemispheric coherences and spectral power for EEG delta, theta, alpha, beta and gamma bands were computed involving 19 scalp locations. The selection of most discriminant features was performed with a rank-based feature selection method assigning a weight value to each feature according to a criterion, i.e., receiver operating characteristics curve. For example, a feature with large weight was considered more relevant to the target labels than a feature with less weight. Therefore, a reduced set of most discriminant features was identified and further be utilized during classification of AUD patients and healthy controls. As results, the inter-hemispheric coherences between the brain regions were found significantly different between the study groups and provided high classification efficiency (Accuracy = 80.8, sensitivity = 82.5, and specificity = 80, F-Measure = 0.78). In addition, the power computed in different EEG bands were found significant and provided an overall classification efficiency as (Accuracy = 86.6, sensitivity = 95, specificity = 82.5, and F-Measure = 0.88). Further, the integration of these EEG feature resulted into even higher results (Accuracy = 89.3 %, sensitivity = 88.5 %, specificity = 91 %, and F-Measure = 0.90). Based on the results, it is concluded that the EEG data (integration of the theta, beta, and gamma power and inter-hemispheric coherence) could be utilized as objective markers to screen the AUD patients and healthy controls.