Deep learning applied to electroencephalogram data in mental disorders: A systematic review

Data leakage in deep learning studies of translational EEG

A growing number of studies apply deep neural networks (DNNs) to recordings of human electroencephalography (EEG) to identify a range of disorders. In many studies, EEG recordings are split into segments, and each segment is randomly assigned to the training or test set. As a consequence, data from individual subjects appears in both the training and the test set. Could high test-set accuracy reflect data leakage from subject-specific patterns in the data, rather than patterns that identify a disease? We address this question by testing the performance of DNN classifiers using segment-based holdout (in which segments from one subject can appear in both the training and test set), and comparing this to their performance using subject-based holdout (where all segments from one subject appear exclusively in either the training set or the test set). In two datasets (one classifying Alzheimer's disease, and the other classifying epileptic seizures), we find that performance on previously-unseen subjects is strongly overestimated when models are trained using segment-based holdout. Finally, we survey the literature and find that the majority of translational DNN-EEG studies use segment-based holdout. Most published DNN-EEG studies may dramatically overestimate their classification performance on new subjects.

Deep insights into MCI diagnosis: A comparative deep learning analysis of EEG time series

BACKGROUND

Individuals in the early stages of Alzheimer's Disease (AD) are typically diagnosed with Mild Cognitive Impairment (MCI). MCI represents a transitional phase between normal cognitive function and AD. Electroencephalography (EEG) records carry valuable insights into cerebral cortex brain activities to analyze neuronal degeneration. To enhance the precision of dementia diagnosis, automatic and intelligent methods are required for the analysis and processing of EEG signals.

NEW METHODS

This paper aims to address the challenges associated with MCI diagnosis by leveraging EEG signals and deep learning techniques. The analysis in this study focuses on processing the information embedded within the sequence of raw EEG time series data. EEG recordings are collected from 10 Healthy Controls (HC) and 10 MCI participants using 19 electrodes during a 30 min eyes-closed session. EEG time series are transformed into 2 separate formats of input tensors and applied to deep neural network architectures. Convolutional Neural Network (CNN) and ResNet from scratch are performed with 2D time series with different segment lengths. Furthermore, EEGNet and DeepConvNet architectures are utilized for 1D time series.

RESULTS

ResNet demonstrates superior effectiveness in detecting MCI when compared to CNN architecture. Complete discrimination is achieved using EEGNet and DeepConvNet for noisy segments.

COMPARISON WITH EXISTING METHODS

ResNet has yielded a 3 % higher accuracy rate compared to CNN. None of the architectures in the literature have achieved 100 % accuracy except proposed EEGNet and DeepConvnet.

CONCLUSION

Deep learning architectures hold great promise in enhancing the accuracy of early MCI detection.

Finding neural correlates of depersonalisation/derealisation disorder via explainable CNN-based analysis guided by clinical assessment scores

Mental health disorders are typically diagnosed based on subjective reports (e.g., through questionnaires) followed by clinical interviews to evaluate the self-reported symptoms. Therefore, considering the interconnected nature of psychiatric disorders, their accurate diagnosis is a real challenge without indicators of underlying physiological dysfunction. Depersonalisation/derealisation disorder (DPD) is an example of dissociative disorder affecting 1-2 % of the population. DPD is characterised mainly by persistent disembodiment, detachment from surroundings, and feelings of emotional numbness, which can significantly impact patients' quality of life. The underlying neural correlates of DPD have been investigated for years to understand and help with a more accurate and in-time diagnosis of the disorder. However, in terms of EEG studies, which hold great importance due to their convenient and inexpensive nature, the literature has often been based on hypotheses proposed by experts in the field, which require prior knowledge of the disorder. In addition, participants' labelling in research experiments is often derived from the outcome of the Cambridge Depersonalisation Scale (CDS), a subjective assessment to quantify the level of depersonalisation/derealisation, the threshold and reliability of which might be challenged. As a result, we aimed to propose a novel end-to-end EEG processing pipeline based on deep neural networks for DPD biomarker discovery, which requires no prior handcrafted labelled data. Alternatively, it can assimilate knowledge from clinical outcomes like CDS as well as data-driven patterns that differentiate individual brain responses. In addition, the structure of the proposed model targets the uncertainty in CDS scores by using them as prior information only to guide the unsupervised learning task in a multi-task learning scenario. A comprehensive evaluation has been done to confirm the significance of the proposed deep structure, including new ways of network visualisation to investigate spectral, spatial, and temporal information derived in the learning process. We argued that the proposed EEG analytics could also be applied to investigate other psychological and mental disorders currently indicated on the basis of clinical assessment scores. The code to reproduce the results presented in this paper is openly accessible at https://github.com/AbbasSalami/DPD_Analysis.

AS, Hatlestad-Hall C, Yazidi A, Hammer HL, Hebold Haraldsen IRJ. Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models

Introduction

A challenge when applying an artificial intelligence (AI) deep learning (DL) approach to novel electroencephalography (EEG) data, is the DL architecture's lack of adaptability to changing numbers of EEG channels. That is, the number of channels cannot vary neither in the training data, nor upon deployment. Such highly specific hardware constraints put major limitations on the clinical usability and scalability of the DL models.

Methods

In this work, we propose a technique for handling such varied numbers of EEG channels by splitting the EEG montages into distinct regions and merge the channels within the same region to a region representation. The solution is termed Region Based Pooling (RBP). The procedure of splitting the montage into regions is performed repeatedly with different region configurations, to minimize potential loss of information. As RBP maps a varied number of EEG channels to a fixed number of region representations, both current and future DL architectures may apply RBP with ease. To demonstrate and evaluate the adequacy of RBP to handle a varied number of EEG channels, sex classification based solely on EEG was used as a test example. The DL models were trained on 129 channels, and tested on 32, 65, and 129-channels versions of the data using the same channel positions scheme. The baselines for comparison were zero-filling the missing channels and applying spherical spline interpolation. The performances were estimated using 5-fold cross validation.

Results

For the 32-channel system version, the mean AUC values across the folds were: RBP (93.34%), spherical spline interpolation (93.36%), and zero-filling (76.82%). Similarly, on the 65-channel system version, the performances were: RBP (93.66%), spherical spline interpolation (93.50%), and zero-filling (85.58%). Finally, the 129-channel system version produced the following results: RBP (94.68%), spherical spline interpolation (93.86%), and zero-filling (91.92%).

Conclusion

In conclusion, RBP obtained similar results to spherical spline interpolation, and superior results to zero-filling. We encourage further research and development of DL models in the cross-dataset setting, including the use of methods such as RBP and spherical spline interpolation to handle a varied number of EEG channels.

Lysophosphatidylcholine acyltransferase level predicts the severity and prognosis of patients with community-acquired pneumonia: a prospective multicenter study

Background

Identifying the diagnosis as well as prognosis for patients presented with community-acquired pneumonia (CAP) remains challenging. We aimed to identify the role of lysophosphatidylcholine acyl-transferase (LPCAT) for CAP along with assessing this protein's effectiveness as a biomarker for severity of disease and mortality.

Methods

Prospective multicenter research study was carried out among hospitalized patients. A total of 299 CAP patients (including 97 severe CAP patients [SCAP]) and 20 healthy controls (HC) were included. A quantitative enzyme-linked immunosorbent test kit was employed for detecting the LPCAT level in plasma. We developed a deep-learning-based binary classification (SCAP or non-severe CAP [NSCAP]) model to process LPCAT levels and other laboratory test results.

Results

The level of LPCAT in patients with SCAP and death outcome was significantly higher than that in other patients. LPCAT showed the highest predictive value for SCAP. LPCAT was able to predict 30-day mortality among CAP patients, combining LPCAT values with PSI scores or CURB-65 further enhance mortality prediction accuracy.

Conclusion

The on admission level of LPCAT found significantly raised among SCAP patients and strongly predicted SCAP patients but with no correlation to etiology. Combining the LPCAT value with CURB-65 or PSI improved the 30-day mortality forecast significantly.

Trial registration

NCT03093220 Registered on March 28th, 2017.

Point of Care Testing (POCT) in Psychopathology Using Fractal Analysis and Hilbert Huang Transform of Electroencephalogram (EEG)

Research has shown that relying only on self-reports for diagnosing psychiatric disorders does not yield accurate results at all times. The advances of technology as well as artificial intelligence and other machine learning algorithms have allowed the introduction of point of care testing (POCT) including EEG characterization and correlations with possible psychopathology. Nonlinear methods of EEG analysis have significant advantages over linear methods. Empirical mode decomposition (EMD) is a reliable nonlinear method of EEG pre-processing. In this chapter, we compare two existing EEG complexity measures - Higuchi fractal dimension (HFD) and sample entropy (SE), with our newly proposed method using Higuchi fractal dimension from the Hilbert Huang transform (HFD-HHT). We present an example using the three complexity measures on a 2-minute EEG recorded from a healthy 20-year-old male after signal pre-processing. Furthermore, we showed the usefulness of these complexity measures in the classification of major depressive disorder (MDD) with healthy controls. Our study is in line with previous research and has shown an increase in HFD and SE values in the full, alpha and beta frequency bands suggestive of an increase in EEG irregularity. Moreover, the HFD-HHT values decreased in those three bands for majority of electrodes which is suggestive of a decrease in irregularity in the frequency-time domain. We conclude that all three complexity measures can be vital features useful for EEG analysis which could be incorporated in POCT systems.

Systematic reviews of machine learning in healthcare: a literature review

INTRODUCTION

The increasing availability of data and computing power has made machine learning (ML) a viable approach to faster, more efficient healthcare delivery.

METHODS

A systematic literature review (SLR) of published SLRs evaluating ML applications in healthcare settings published between1 January 2010 and 27 March 2023 was conducted.

RESULTS

In total 220 SLRs covering 10,462 ML algorithms were reviewed. The main application of AI in medicine related to the clinical prediction and disease prognosis in oncology and neurology with the use of imaging data. Accuracy, specificity, and sensitivity were provided in 56%, 28%, and 25% SLRs respectively. Internal and external validation was reported in 53% and less than 1% of the cases respectively. The most common modeling approach was neural networks (2,454 ML algorithms), followed by support vector machine and random forest/decision trees (1,578 and 1,522 ML algorithms, respectively).

EXPERT OPINION

The review indicated considerable reporting gaps in terms of the ML's performance, both internal and external validation. Greater accessibility to healthcare data for developers can ensure the faster adoption of ML algorithms into clinical practice.

The now and future of ChatGPT and GPT in psychiatry

ChatGPT has sparked extensive discussions within the healthcare community since its November 2022 release. However, potential applications in the field of psychiatry have received limited attention. Deep learning has proven beneficial to psychiatry, and GPT is a powerful deep learning-based language model with immense potential for this field. Despite the convenience of ChatGPT, this advanced chatbot currently has limited practical applications in psychiatry. It may be used to support psychiatrists in routine tasks such as completing medical records, facilitating communications between clinicians and with patients, polishing academic writings and presentations, and programming and performing analyses for research. The current training and application of ChatGPT require using appropriate prompts to maximize appropriate outputs and minimize deleterious inaccuracies and phantom errors. Moreover, future GPT advances that incorporate empathy, emotion recognition, personality assessment, and detection of mental health warning signs are essential for its effective integration into psychiatric care. In the near future, developing a fully-automated psychotherapy system trained for expert communication (such as psychotherapy verbatim) is conceivable by building on foundational GPT technology. This dream system should integrate practical 'real world' inputs and friendly AI user and patient interfaces via clinically validated algorithms, voice comprehension/generation modules, and emotion discrimination algorithms based on facial expressions and physiological inputs from wearable devices. In addition to the technology challenges, we believe it is critical to establish generally accepted ethical standards for applying ChatGPT-related tools in all mental healthcare environments, including telemedicine and academic/training settings.

Machine Learning-Based Detection of Parkinson's Disease From Resting-State EEG: A Multi-Center Study

Resting-state EEG (rs-EEG) has been demonstrated to aid in Parkinson's disease (PD) diagnosis. In particular, the power spectral density (PSD) of low-frequency bands (δ and θ) and high-frequency bands (α and β) has been shown to be significantly different in patients with PD as compared to subjects without PD (non-PD). However, rs-EEG feature extraction and the interpretation thereof can be time-intensive and prone to examiner variability. Machine learning (ML) has the potential to automatize the analysis of rs-EEG recordings and provides a supportive tool for clinicians to ease their workload. In this work, we use rs-EEG recordings of 84 PD and 85 non-PD subjects pooled from four datasets obtained at different centers. We propose an end-to-end pipeline consisting of preprocessing, extraction of PSD features from clinically-validated frequency bands, and feature selection. Following, we assess the classification ability of the features via ML algorithms to stratify between PD and non-PD subjects. Further, we evaluate the effect of feature harmonization, given the multi-center nature of the datasets. Our validation results show, on average, an improvement in PD detection ability (69.6% vs. 75.5% accuracy) by logistic regression when harmonizing the features and performing univariate feature selection (k = 202 features). Our final results show an average global accuracy of 72.2% with balanced accuracy results for all the centers included in the study: 60.6%, 68.7%, 77.7%, and 82.2%, respectively.Clinical relevance- We present an end-to-end pipeline to extract clinically relevant features from rs-EEG recordings that can facilitate the analysis and detection of PD. Further, we provide an ML system that shows a good performance in detecting PD, even in the presence of centers with different acquisition protocols. Our results show the relevance of harmonizing features and provide a good starting point for future studies focusing on rs-EEG analysis and multi-center data.

Deep Learning in the Identification of Electroencephalogram Sources Associated with Sexual Orientation

INTRODUCTION

It is unclear if sexual orientation is a biological trait that has neurofunctional footprints. With deep learning, the power to classify biological datasets without an a priori selection of features has increased by magnitudes. The aim of this study was to correctly classify resting-state electroencephalogram (EEG) data from males with different sexual orientation using deep learning and to explore techniques to identify the learned distinguishing features.

METHODS

Three cohorts (homosexual men, heterosexual men, and a mixed sex cohort), one pretrained network on sex classification, and one newly trained network for sexual orientation classification were used to classify sex. Further, Grad-CAM methodology and source localization were used to identify the spatiotemporal patterns that were used for differentiation by the networks.

RESULTS

Using a pretrained network for classification of males and females, no differences existed between classification of homosexual and heterosexual males. The newly trained network was able, however, to correctly classify the cohorts with a total accuracy of 83%. The retrograde activation using Grad-CAM technology yielded distinctive functional EEG patterns in the Brodmann area 40 and 1 when combined with Fourier analysis and a source localization.

DISCUSSION

This study shows that electrophysiological trait markers of male sexual orientation can be identified using deep learning. These patterns are different from the differentiating signatures of males and females in a resting-state EEG.

Discriminating between bipolar and major depressive disorder using a machine learning approach and resting-state EEG data

OBJECTIVE

Distinguishing major depressive disorder (MDD) from bipolar disorder (BD) is a crucial clinical challenge as effective treatment is quite different for each condition. In this study electroencephalography (EEG) was explored as an objective biomarker for distinguishing MDD from BD using an efficient machine learning algorithm (MLA) trained by a relatively large and balanced dataset.

METHODS

A 3 step MLA was applied: (1) a multi-step preprocessing method was used to improve the quality of the EEG signal, (2) symbolic transfer entropy (STE), an effective connectivity measure, was applied to the resultant EEG and (3) the MLA used the extracted STE features to distinguish MDD (N = 71) from BD (N = 71) subjects.

RESULTS

14 connectivity features were selected by the proposed algorithm. Most of the selected features were related to the frontal, parietal, and temporal lobe electrodes. The major involved regions were the Broca region in the frontal lobe and the somatosensory association cortex in the parietal lobe. These regions are near electrodes FC5 and CPz and are involved in processing language and sensory information, respectively. The resulting classifier delivered an evaluation accuracy of 88.5% and a test accuracy of 89.3%, using 80% of the data for training and evaluation and the remaining 20% for testing, respectively.

CONCLUSIONS

The high evaluation and test accuracies of our algorithm, derived from a large balanced training sample suggests that this method may hold significant promise as a clinical tool.

SIGNIFICANCE

The proposed MLA may provide an inexpensive and readily available tool that clinicians may use to enhance diagnostic accuracy and shorten time to effective treatment.

Sex differences in rTMS treatment response: A deep learning-based EEG investigation

INTRODUCTION

The present study aimed to investigate sex differences in response to repetitive transcranial magnetic stimulation (rTMS) in Major Depressive Disorder (MDD) patients. Identifying the factors that mediate treatment response to rTMS in MDD patients can guide clinicians to administer more appropriate, reliable, and personalized interventions.

METHODS

In this paper, we developed a novel pipeline based on convolutional LSTM-based deep learning (DL) to classify 25 female and 25 male patients based on their rTMS treatment response.

RESULTS

Five different classification models were generated, namely pre-/post-rTMS female (model 1), pre-/post-rTMS male (model 2), pre-rTMS female responder versus pre-rTMS female nonresponders (model 3), pre-rTMS male responder vs. pre-rTMS male nonresponder (model 4), and pre-rTMS responder versus nonresponder of both sexes (model 5), achieving 93.3%, 98%, 95.2%, 99.2%, and 96.6% overall test accuracy, respectively.

CONCLUSION

These results indicate the potential of our approach to be used as a response predictor especially regarding sex-specific antidepressant effects of rTMS in MDD patients.

Multimodal Assessment of Schizophrenia and Depression Utilizing Video, Acoustic, Locomotor, Electroencephalographic, and Heart Rate Technology: Protocol for an Observational Study

Background

Current standards of psychiatric assessment and diagnostic evaluation rely primarily on the clinical subjective interpretation of a patient’s outward manifestations of their internal state. While psychometric tools can help to evaluate these behaviors more systematically, the tools still rely on the clinician’s interpretation of what are frequently nuanced speech and behavior patterns. With advances in computing power, increased availability of clinical data, and improving resolution of recording and sensor hardware (including acoustic, video, accelerometer, infrared, and other modalities), researchers have begun to demonstrate the feasibility of cutting-edge technologies in aiding the assessment of psychiatric disorders.

Objective

We present a research protocol that utilizes facial expression, eye gaze, voice and speech, locomotor, heart rate, and electroencephalography monitoring to assess schizophrenia symptoms and to distinguish patients with schizophrenia from those with other psychiatric disorders and control subjects.

Methods

We plan to recruit three outpatient groups: (1) 50 patients with schizophrenia, (2) 50 patients with unipolar major depressive disorder, and (3) 50 individuals with no psychiatric history. Using an internally developed semistructured interview, psychometrically validated clinical outcome measures, and a multimodal sensing system utilizing video, acoustic, actigraphic, heart rate, and electroencephalographic sensors, we aim to evaluate the system’s capacity in classifying subjects (schizophrenia, depression, or control), to evaluate the system’s sensitivity to within-group symptom severity, and to determine if such a system can further classify variations in disorder subtypes.

Results

Data collection began in July 2020 and is expected to continue through December 2022.

Conclusions

If successful, this study will help advance current progress in developing state-of-the-art technology to aid clinical psychiatric assessment and treatment. If our findings suggest that these technologies are capable of resolving diagnoses and symptoms to the level of current psychometric testing and clinician judgment, we would be among the first to develop a system that can eventually be used by clinicians to more objectively diagnose and assess schizophrenia and depression with the possibility of less risk of bias. Such a tool has the potential to improve accessibility to care; to aid clinicians in objectively evaluating diagnoses, severity of symptoms, and treatment efficacy through time; and to reduce treatment-related morbidity.

International Registered Report Identifier (IRRID)

DERR1-10.2196/36417

A Machine Learning Algorithm to Discriminating Between Bipolar and Major Depressive Disorders Based on Resting EEG Data

Distinguishing major depressive disorder (MDD) from bipolar disorder (BD) is a crucial clinical challenge due to the lack of known biomarkers. Conventional methods of diagnosis rest exclusively on symptomatic presentation, and personal and family history. As a result, BD-depressed episode (BD-DE) is often misdiagnosed as MDD, and inappropriate therapy is given. Electroencephalography (EEG) has been widely studied as a potential source of biomarkers to differentiate these disorders. Previous attempts using machine learning (ML) methods have delivered insufficient sensitivity and specificity for clinical use, likely as a consequence of the small training set size, and inadequate ML methodology. We hope to overcome these limitations by employing a training dataset of resting-state EEG from 71 MDD and 71 BD patients. We introduce a robust 3 steps ML technique: 1) a multi-step preprocessing method is used to improve the quality of the EEG signal 2) symbolic transfer entropy (STE), which is an effective connectivity measure, is applied to the resultant EEG signals 3) the ML algorithm uses the extracted STE features to distinguish MDD from BD patients. Clinical Relevance--- The accuracy of our algorithm, derived from a large sample of patients, suggests that this method may hold significant promise as a clinical tool. The proposed method delivered total accuracy, sensitivity, and specificity of 84.9%, 83.4%, and 87.1%, respectively.

Utility of Cognitive Neural Features for Predicting Mental Health Behaviors

Cognitive dysfunction underlies common mental health behavioral symptoms including depression, anxiety, inattention, and hyperactivity. In this study of 97 healthy adults, we aimed to classify healthy vs. mild-to-moderate self-reported symptoms of each disorder using cognitive neural markers measured with an electroencephalography (EEG). We analyzed source-reconstructed EEG data for event-related spectral perturbations in the theta, alpha, and beta frequency bands in five tasks, a selective attention and response inhibition task, a visuospatial working memory task, a Flanker interference processing task, and an emotion interference task. From the cortical source activation features, we derived augmented features involving co-activations between any two sources. Logistic regression on the augmented feature set, but not the original feature set, predicted the presence of psychiatric symptoms, particularly for anxiety and inattention with >80% sensitivity and specificity. We also computed current flow closeness and betweenness centralities to identify the “hub” source signal predictors. We found that the Flanker interference processing task was the most useful for assessing the connectivity hubs in general, followed by the inhibitory control go-nogo paradigm. Overall, these interpretable machine learning analyses suggest that EEG biomarkers collected on a rapid suite of cognitive assessments may have utility in classifying diverse self-reported mental health symptoms.

Application of Artificial Intelligence on Psychological Interventions and Diagnosis: An Overview

BACKGROUND

Innovative technologies, such as machine learning, big data, and artificial intelligence (AI) are approaches adopted for personalized medicine, and psychological interventions and diagnosis are facing huge paradigm shifts. In this literature review, we aim to highlight potential applications of AI on psychological interventions and diagnosis.

METHODS

This literature review manifest studies that discuss how innovative technology as deep learning (DL) and AI is affecting psychological assessment and psychotherapy, we performed a search on PUBMED, and Web of Science using the terms "psychological interventions," "diagnosis on mental health disorders," "artificial intelligence," and "deep learning." Only studies considering patients' datasets are considered.

RESULTS

Nine studies met the inclusion criteria. Beneficial effects on clinical symptoms or prediction were shown in these studies, but future study is needed to determine the long-term effects.

LIMITATIONS

The major limitation for the current study is the small sample size, and lies in the lack of long-term follow-up-controlled studies for a certain symptom.

CONCLUSIONS

AI such as DL applications showed promising results on clinical practice, which could lead to profound impact on personalized medicine for mental health conditions. Future studies can improve furthermore by increasing sample sizes and focusing on ethical approvals and adherence for online-therapy.