Deep Learning Representation from Electroencephalography of Early-Stage Creutzfeldt-Jakob Disease and Features for Differentiation from Rapidly Progressive Dementia

EEG-based classification of individuals with neuropsychiatric disorders using deep neural networks: A systematic review of current status and future directions

The use of deep neural networks for electroencephalogram (EEG) classification has rapidly progressed and gained popularity in recent years, but automatic feature extraction from EEG signals remains a challenging task. The classification of neuropsychiatric disorders demands the extraction of neuro-markers for use in automated EEG classification. Numerous advanced deep learning algorithms can be used for this purpose. In this article, we present a comprehensive review of the main factors and parameters that affect the performance of deep neural networks in classifying different neuropsychiatric disorders using EEG signals. We also analyze the EEG features used for improving classification performance. Our analysis includes 82 scientific journal papers that applied deep neural networks for subject-wise classification based on EEG signals. We extracted information on the EEG dataset and types of disorders, deep neural network structures, performance, and hyperparameters. The results show that most studies have focused on clinical classification, achieving an average accuracy of 91.83 ± 7.34, with convolutional neural networks (CNNs) being the most frequently used network architecture and resting-state EEG signals being the most commonly used data type. Additionally, the review reveals that depression (N = 18), Alzheimer's (N = 11), and schizophrenia (N = 11) were studied more frequently than other types of neuropsychiatric disorders. Our review provides insight into the performance of deep neural networks in EEG classification and highlights the importance of EEG feature extraction in improving classification accuracy. By identifying the main factors and parameters that affect deep neural network performance in EEG classification, our review can guide future research in this area. We hope that our findings will encourage further exploration of deep learning methods for EEG classification and contribute to the development of more accurate and effective methods for diagnosing and monitoring neuropsychiatric disorders using EEG signals.

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.

Depression screening using hybrid neural network

Depression is a common cause of increased suicides worldwide, and studies have shown that the number of patients suffering from major depressive disorder (MDD) increased several-fold during the COVID-19 pandemic, highlighting the importance of disease detection and depression management, while increasing the need for effective diagnostic tools. In recent years, machine learning and deep learning methods based on electroencephalography (EEG) have achieved significant results in the field of automatic depression detection. However, most current studies have focused on a small number of EEG signal channels, and experimental data require special processing by professionals. In this study, 128 channels of EEG signals were simply filtered and 24-fold leave-one-out cross-validation experiments were performed using 2DCNN-LSTM classifier, support vector machine, K-nearest neighbor and decision tree. The current results show that the proposed 2DCNN-LSTM model has an average classification accuracy of 95.1% with an AUC of 0.98 for depression detection of 6-second participant EEG signals, and the model is much better than 72.05%, 79.7% and 79.49% for support vector machine, K nearest neighbor and decision tree. In addition, we found that the model achieved a 100% probability of correctly classifying the EEG signals of 300-second participants.

Resting-state EEG-based convolutional neural network for the diagnosis of depression and its severity

Purpose: The study aimed to assess the value of the resting-state electroencephalogram (EEG)-based convolutional neural network (CNN) method for the diagnosis of depression and its severity in order to better serve depressed patients and at-risk populations.

Methods: In this study, we used the resting state EEG-based CNN to identify depression and evaluated its severity. The EEG data were collected from depressed patients and healthy people using the Nihon Kohden EEG-1200 system. Analytical processing of resting-state EEG data was performed using Python and MATLAB software applications. The questionnaire included the Self-Rating Anxiety Scale (SAS), Self-Rating Depression Scale (SDS), Symptom Check-List-90 (SCL-90), and the Eysenck Personality Questionnaire (EPQ).

Results: A total of 82 subjects were included in this study, with 41 in the depression group and 41 in the healthy control group. The area under the curve (AUC) of the resting-state EEG-based CNN in depression diagnosis was 0.74 (95%CI: 0.70–0.77) with an accuracy of 66.40%. In the depression group, the SDS, SAS, SCL-90 subscales, and N scores were significantly higher in the major depression group than those in the non-major depression group (p < 0.05). The AUC of the model in depression severity was 0.70 (95%CI: 0.65–0.75) with an accuracy of 66.93%. Correlation analysis revealed that major depression AI scores were significantly correlated with SAS scores (r = 0.508, p = 0.003) and SDS scores (r = 0.765, p < 0.001).

Conclusion: Our model can accurately identify the depression-specific EEG signal in terms of depression diagnosis and severity identification. It would eventually provide new strategies for early diagnosis of depression and its severity.

An EEG-based systematic explainable detection framework for probing and localizing abnormal patterns in Alzheimer's disease

OBJECTIVE

Electroencephalography (EEG) is a potential source of downstream biomarkers for the early diagnosis of Alzheimer's disease (AD) due to its low-cost, non-invasive, and portable advantages. Accurately detecting AD-induced patterns from EEG signals is essential for understanding AD-related neurodegeneration at the EEG level and further evaluating the risk of AD at an early stage. This paper proposes a deep learning-based, functional explanatory framework that probes AD abnormalities from short-sequence EEG data.

APPROACH

The framework is a learning-based automatic detection system consisting of three encoding pathways that analyze EEG signals in frequency, complexity, and synchronous domains. We integrated the proposed EEG descriptors with the neural network components into one learning system to detect AD patterns. A transfer learning-based model was used to learn the deep representations, and a modified generative adversarial module was attached to the model to overcome feature sparsity. Furthermore, we utilized activation mapping to obtain the AD-related neurodegeneration at brain rhythm, dynamic complexity, and functional connectivity levels.

MAIN RESULTS

The proposed framework can accurately (100%) detect AD patterns based on our raw EEG recordings without delicate preprocessing. Meanwhile, the system indicates that 1) the power of different brain rhythms exhibits abnormal in the frontal lobes of AD patients, and such abnormality spreads to central lobes in the alpha and beta rhythms, 2) the difference in nonlinear complexity varies with the temporal scales, and 3) all the connections of pair-wise brain regions except bilateral temporal connectivity are weak in AD patterns. The proposed method outperforms other related methods in detection performance.

SIGNIFICANCE

We provide a new method for revealing abnormalities and corresponding localizations in different feature domains of EEG from AD patients. This study is a significant foundation for our future work on identifying individuals at high risk of AD at an early stage.

Multi-Method Analysis of Medical Records and MRI Images for Early Diagnosis of Dementia and Alzheimer’s Disease Based on Deep Learning and Hybrid Methods

Dementia and Alzheimer’s disease are caused by neurodegeneration and poor communication between neurons in the brain. So far, no effective medications have been discovered for dementia and Alzheimer’s disease. Thus, early diagnosis is necessary to avoid the development of these diseases. In this study, efficient machine learning algorithms were assessed to evaluate the Open Access Series of Imaging Studies (OASIS) dataset for dementia diagnosis. Two CNN models (AlexNet and ResNet-50) and hybrid techniques between deep learning and machine learning (AlexNet+SVM and ResNet-50+SVM) were also evaluated for the diagnosis of Alzheimer’s disease. For the OASIS dataset, we balanced the dataset, replaced the missing values, and applied the t-Distributed Stochastic Neighbour Embedding algorithm (t-SNE) to represent the high-dimensional data in the low-dimensional space. All of the machine learning algorithms, namely, Support Vector Machine (SVM), Decision Tree, Random Forest and K Nearest Neighbours (KNN), achieved high performance for diagnosing dementia. The random forest algorithm achieved an overall accuracy of 94% and precision, recall and F1 scores of 93%, 98% and 96%, respectively. The second dataset, the MRI image dataset, was evaluated by AlexNet and ResNet-50 models and AlexNet+SVM and ResNet-50+SVM hybrid techniques. All models achieved high performance, but the performance of the hybrid methods between deep learning and machine learning was better than that of the deep learning models. The AlexNet+SVM hybrid model achieved accuracy, sensitivity, specificity and AUC scores of 94.8%, 93%, 97.75% and 99.70%, respectively.

Creutzfeldt-Jakob disease with neuroleptic malignant syndrome

Creutzfeldt‐Jakob disease (CJD) is a rare rapidly progressive fatal neurodegenerative disease. Neuroleptic malignant syndrome (NMS) is a complication of antipsychotic medications which may be used to treat neuropsychiatric symptoms of CJD. We present a case of a 51‐year‐ old woman with CJD who developed NMS after being prescribed quetiapine.

Identification of Autism Subtypes Based on Wavelet Coherence of BOLD FMRI Signals Using Convolutional Neural Network

The functional connectivity (FC) patterns of resting-state functional magnetic resonance imaging (rs-fMRI) play an essential role in the development of autism spectrum disorders (ASD) classification models. There are available methods in literature that have used FC patterns as inputs for binary classification models, but the results barely reach an accuracy of 80%. Additionally, the generalizability across multiple sites of the models has not been investigated. Due to the lack of ASD subtypes identification model, the multi-class classification is proposed in the present study. This study aims to develop automated identification of autism spectrum disorder (ASD) subtypes using convolutional neural networks (CNN) using dynamic FC as its inputs. The rs-fMRI dataset used in this study consists of 144 individuals from 8 independent sites, labeled based on three ASD subtypes, namely autistic disorder (ASD), Asperger’s disorder (APD), and pervasive developmental disorder not otherwise specified (PDD-NOS). The blood-oxygen-level-dependent (BOLD) signals from 116 brain nodes of automated anatomical labeling (AAL) atlas are used, where the top-ranked node is determined based on one-way analysis of variance (ANOVA) of the power spectral density (PSD) values. Based on the statistical analysis of the PSD values of 3-level ASD and normal control (NC), putamen_R is obtained as the top-ranked node and used for the wavelet coherence computation. With good resolution in time and frequency domain, scalograms of wavelet coherence between the top-ranked node and the rest of the nodes are used as dynamic FC feature input to the convolutional neural networks (CNN). The dynamic FC patterns of wavelet coherence scalogram represent phase synchronization between the pairs of BOLD signals. Classification algorithms are developed using CNN and the wavelet coherence scalograms for binary and multi-class identification were trained and tested using cross-validation and leave-one-out techniques. Results of binary classification (ASD vs. NC) and multi-class classification (ASD vs. APD vs. PDD-NOS vs. NC) yielded, respectively, 89.8% accuracy and 82.1% macro-average accuracy, respectively. Findings from this study have illustrated the good potential of wavelet coherence technique in representing dynamic FC between brain nodes and open possibilities for its application in computer aided diagnosis of other neuropsychiatric disorders, such as depression or schizophrenia.

Hysteresis Modeling in Iron-Dominated Magnets Based on a Multi-Layered Narx Neural Network Approach

A full-fledged neural network modeling, based on a Multi-layered Nonlinear Autoregressive Exogenous Neural Network (NARX) architecture, is proposed for quasi-static and dynamic hysteresis loops, one of the most challenging topics for computational magnetism. This modeling approach overcomes drawbacks in attaining better than percent-level accuracy of classical and recent approaches for accelerator magnets, that combine hybridization of standard hysteretic models and neural network architectures. By means of an incremental procedure, different Deep Neural Network Architectures are selected, fine-tuned and tested in order to predict magnetic hysteresis in the context of electromagnets. Tests and results show that the proposed NARX architecture best fits the measured magnetic field behavior of a reference quadrupole at CERN. In particular, the proposed modeling framework leads to a percent error below 0.02% for the magnetic field prediction, thus outperforming state of the art approaches and paving a very promising way for future real time applications.

One-Shot Neural Architecture Search by Dynamically Pruning Supernet in Hierarchical Order

Neural Architecture Search (NAS), which aims at automatically designing neural architectures, recently draw a growing research interest. Different from conventional NAS methods, in which a large number of neural architectures need to be trained for evaluation, the one-shot NAS methods only have to train one supernet which synthesizes all the possible candidate architectures. As a result, the search efficiency could be significantly improved by sharing the supernet's weights during the candidate architectures' evaluation. This strategy could greatly speed up the search process but suffer a challenge that the evaluation based on sharing weights is not predictive enough. Recently, pruning the supernet during the search has been proven to be an efficient way to alleviate this problem. However, the pruning direction in complex-structured search space remains unexplored. In this paper, we revisited the role of path dropout strategy, which drops the neural operations instead of the neurons, in supernet training, and several interesting characters of the supernet trained with dropout are found. Based on the observations, a Hierarchically-Ordered Pruning Neural Architecture Search (HOPNAS) algorithm is proposed by dynamically pruning the supernet with a proper pruning direction. Experimental results indicate that our method is competitive with state-of-the-art approaches on CIFAR10 and ImageNet.

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

In recent medical research, tremendous progress has been made in the application of deep learning (DL) techniques. This article systematically reviews how DL techniques have been applied to electroencephalogram (EEG) data for diagnostic and predictive purposes in conducting research on mental disorders. EEG-studies on psychiatric diseases based on the ICD-10 or DSM-V classification that used either convolutional neural networks (CNNs) or long -short-term-memory (LSTMs) networks for classification were searched and examined for the quality of the information they contained in three domains: clinical, EEG-data processing, and deep learning. Although we found that the description of EEG acquisition and pre-processing was sufficient in most of the studies, we found, that many of them lacked a systematic characterization of clinical features. Furthermore, many studies used misguided model selection procedures or flawed testing. It is recommended that the study of psychiatric disorders using DL in the future must improve the quality of clinical data and follow state of the art model selection and testing procedures so as to achieve a higher research standard and head toward a clinical significance.

A Deep Fourier Neural Network for Seizure Prediction Using Convolutional Neural Network and Ratios of Spectral Power

Epileptic seizure prediction is one of the most used therapeutic adjuvant strategies for drug-resistant epilepsy. Conventional methods usually adopt handcrafted features and manual parameter setting. The over-reliance on the expertise of specialists may lead to weak exploitation of features and low popularization of clinical application. This paper proposes a novel parameterless patient-specific method based on Fourier Neural Network (FNN), where the Fourier transform and backpropagation learning are synthesized to make the predictor more efficient and practical. The employment of FNN is the first attempt in the field of seizure prediction due to its automatic extraction of immanent spectra in epileptic signals. Despite the self-adaptive superiority of FNN, we introduce Convolutional Neural Network (CNN) to further improve its search capability in high-dimensional feature spaces. The study also develops a multi-layer module to estimate spectral power ratios of raw recordings, which optimizes the prediction by enhancing feature diversity. Based on these modules, this paper proposes a two-channel deep neural network: Fourier Ratio Convolutional Neural Network (FRCNN). To demonstrate the reliability of the model, we explain the mathematical meaning of hidden-layer neurons in FRCNN theoretically. This approach is evaluated on both intracranial and scalp EEG datasets. It shows that the predictor achieved a sensitivity of 91.2% and a false prediction rate (FPR) of 0.06[Formula: see text]h[Formula: see text] across intracranial subjects and a sensitivity of 85.4% and an FPR of 0.14[Formula: see text]h[Formula: see text] over scalp subjects. The results indicate that FRCNN enables the convenience of epilepsy treatments while preserving a high degree of precision. In the end, a detailed comparison with the previous methods demonstrates that FRCNN has achieved higher performance and generalization ability.

Classification of Relaxation and Concentration Mental States with EEG

In this paper, we study the use of EEG (Electroencephalography) to classify between concentrated and relaxed mental states. In the literature, most EEG recording systems are expensive, medical-graded devices. The expensive devices limit the availability in a consumer market. The EEG signals are obtained from a toy-grade EEG device with one channel of output data. The experiments are conducted in two runs, with 7 and 10 subjects, respectively. Each subject is asked to silently recite a five-digit number backwards given by the tester. The recorded EEG signals are converted to time-frequency representations by the software accompanying the device. A simple average is used to aggregate multiple spectral components into EEG bands, such as α, β, and γ bands. The chosen classifiers are SVM (support vector machine) and multi-layer feedforward network trained individually for each subject. Experimental results show that features, with α+β+γ bands and bandwidth 4 Hz, the average accuracy over all subjects in both runs can reach more than 80% and some subjects up to 90+% with the SVM classifier. The results suggest that a brain machine interface could be implemented based on the mental states of the user even with the use of a cheap EEG device.

Spatial distribution of abnormal EEG activity in Creutzfeldt-Jakob disease

OBJECTIVES

Electroencephalogram (EEG) pattern in Creutzfeldt-Jakob disease (CJD) is characterized by diffuse abnormal activity, although lateralization to one hemisphere has been described in the first stages of the disease. This study aimed to determine whether abnormal EEG activity predominantly occurs in anterior versus posterior brain regions.

METHODS

As part of a prospective study, the demographics, clinical features and MRI findings of genetic E200K CJD patients were collected. EEG was performed and the recordings reviewed for the typical periodic sharp wave complex (PSWC) and non-specific slow activity. Data were analyzed using the qEEG tool, and the activity in anterior and posterior regions of the brain compared.

RESULTS

Eleven genetic E200K CJD patients were included in the study (67% women). The average age was 59.1 ± 8.4 SD years and the average disease duration was 2.4 ± 2.1 months. EEG showed the classic PSWC pattern in 5/11 (45%) of the patients, and slow activity was seen in 9/11 (82%). EEG was normal in 2 patients. PSWC activity was diffuse in 2/5 patients and unilateral in 3/5 patients; slow activity was diffuse in 9 patients. Quantitative analysis of PSWC and slow activity showed no significant difference between anterior and posterior distribution.

CONCLUSION

The abnormal EEG activity in CJD is diffuse with no clear spatial predominance in anterior or posterior brain regions.

A New Recognition Method for the Auditory Evoked Magnetic Fields

Magnetoencephalography (MEG) is a persuasive tool to study the human brain in physiology and psychology. It can be employed to obtain the inference of change between the external environment and the internal psychology, which requires us to recognize different single trial event-related magnetic fields (ERFs) originated from different functional areas of the brain. Current recognition methods for the single trial data are mainly used for event-related potentials (ERPs) in the electroencephalography (EEG). Although the MEG shares the same signal sources with the EEG, much less interference from the other brain tissues may give the MEG an edge in recognition of the ERFs. In this work, we propose a new recognition method for the single trial auditory evoked magnetic fields (AEFs) through enhancing the signal. We find that the signal strength of the single trial AEFs is concentrated in the primary auditory cortex of the temporal lobe, which can be clearly displayed in the 2D images. These 2D images are then recognized by an artificial neural network (ANN) with 100% accuracy, which realizes the automatic recognition for the single trial AEFs. The method not only may be combined with the source estimation algorithm to improve its accuracy but also paves the way for the implementation of the brain-computer interface (BCI) with the MEG.

Clinical manifestations and polysomnography-based analysis in nine cases of probable sporadic Creutzfeldt-Jakob disease

PURPOSE

To summarize the clinical characteristics of patients with sporadic Creutzfeldt-Jakob disease (sCJD), analyze its sleep disorder characteristics using polysomnography (PSG), and compare sleep disturbances with those of fatal familial insomnia (FFI).

PATIENTS AND METHODS

We retrospectively reviewed the sleep disturbances; cerebrospinal fluid (CSF) protein 14-3-3 (CSF-14-3-3 protein); prion protein gene, PRNP; magnetic resonance imaging; and electroencephalogram (EEG) of nine sCJD patients RESULTS: Of the nine sCJD patients, six were positive for CSF-14-3-3 protein. In the eight patients who completed diffusion-weighted imaging, seven showed cortical "ribbons sign" and two showed high signal in the basal ganglia. All nine patients had an EEG, which showed an increase in background slow waves; moreover, four showed typical periodic sharp wave complexes. The codon diversity at position 129, 219 of nine patients were MM, EE. Almost all nine patients had sleep disturbances such as insomnia, hypersomnia, and periodic limb movement disorder (PLMD). Five patients completed PSG, which demonstrated severe sleep structure disorder, prolonged total waking time, significantly reduced sleep efficiency, and absent rapid eye movement in some severe patients.

CONCLUSION

Sleep disturbances are common in sCJD patients, manifested as insomnia, lethargy, and PLMD. The sCJD patients often demonstrate severe sleep structure disorder through PSG, which is similar to patients with FFI.

Major Depressive Disorder Classification Based on Different Convolutional Neural Network Models: Deep Learning Approach

The human brain is characterized by complex structural, functional connections that integrate unique cognitive characteristics. There is a fundamental hurdle for the evaluation of both structural and functional connections of the brain and the effects in the diagnosis and treatment of neurodegenerative diseases. Currently, there is no clinically specific diagnostic biomarker capable of confirming the diagnosis of major depressive disorder (MDD). Therefore, exploring translational biomarkers of mood disorders based on deep learning (DL) has valuable potential with its recently underlined promising outcomes. In this article, an electroencephalography (EEG)-based diagnosis model for MDD is built through advanced computational neuroscience methodology coupled with a deep convolutional neural network (CNN) approach. EEG recordings are analyzed by modeling 3 different deep CNN structure, namely, ResNet-50, MobileNet, Inception-v3, in order to dichotomize MDD patients and healthy controls. EEG data are collected for 4 main frequency bands (Δ, θ, α, and β, accompanying spatial resolution with location information by collecting data from 19 electrodes. Following the pre-processing step, different DL architectures were employed to underline discrimination performance by comparing classification accuracies. The classification performance of models based on location data, MobileNet architecture generated 89.33% and 92.66% classification accuracy. As to the frequency bands, delta frequency band outperformed compared to other bands with 90.22% predictive accuracy and area under curve (AUC) value of 0.9 for ResNet-50 architecture. The main contribution of the study is the delineation of distinctive spatial and temporal features using various DL architectures to dichotomize 46 MDD subjects from 46 healthy subjects. Exploring translational biomarkers of mood disorders based on DL perspective is the main focus of this study and, though it is challenging, with its promising potential to improve our understanding of the psychiatric disorders, computational methods are highly worthy for the diagnosis process and valuable in terms of both speed and accuracy compared with classical approaches.

A survey on deep learning-based non-invasive brain signals: recent advances and new frontiers

Brain signals refer to the biometric information collected from the human brain. The research on brain signals aims to discover the underlying neurological or physical status of the individuals by signal decoding. The emerging deep learning techniques have improved the study of brain signals significantly in recent years. In this work, we first present a taxonomy of non-invasive brain signals and the basics of deep learning algorithms. Then, we provide a comprehensive survey of the frontiers of applying deep learning for non-invasive brain signals analysis, by summarizing a large number of recent publications. Moreover, upon the deep learning-powered brain signal studies, we report the potential real-world applications which benefit not only disabled people but also normal individuals. Finally, we discuss the opening challenges and future directions.

Use of machine learning in geriatric clinical care for chronic diseases: a systematic literature review

Objectives

Geriatric clinical care is a multidisciplinary assessment designed to evaluate older patients’ (age 65 years and above) functional ability, physical health, and cognitive well-being. The majority of these patients suffer from multiple chronic conditions and require special attention. Recently, hospitals utilize various artificial intelligence (AI) systems to improve care for elderly patients. The purpose of this systematic literature review is to understand the current use of AI systems, particularly machine learning (ML), in geriatric clinical care for chronic diseases.

Materials and Methods

We restricted our search to eight databases, namely PubMed, WorldCat, MEDLINE, ProQuest, ScienceDirect, SpringerLink, Wiley, and ERIC, to analyze research articles published in English between January 2010 and June 2019. We focused on studies that used ML algorithms in the care of geriatrics patients with chronic conditions.

Results

We identified 35 eligible studies and classified in three groups: psychological disorder (n = 22), eye diseases (n = 6), and others (n = 7). This review identified the lack of standardized ML evaluation metrics and the need for data governance specific to health care applications.

Conclusion

More studies and ML standardization tailored to health care applications are required to confirm whether ML could aid in improving geriatric clinical care.

Middle-Level Features for the Explanation of Classification Systems by Sparse Dictionary Methods

Machine learning (ML) systems are affected by a pervasive lack of transparency. The eXplainable Artificial Intelligence (XAI) research area addresses this problem and the related issue of explaining the behavior of ML systems in terms that are understandable to human beings. In many explanation of XAI approaches, the output of ML systems are explained in terms of low-level features of their inputs. However, these approaches leave a substantive explanatory burden with human users, insofar as the latter are required to map low-level properties into more salient and readily understandable parts of the input. To alleviate this cognitive burden, an alternative model-agnostic framework is proposed here. This framework is instantiated to address explanation problems in the context of ML image classification systems, without relying on pixel relevance maps and other low-level features of the input. More specifically, one obtains sets of middle-level properties of classification inputs that are perceptually salient by applying sparse dictionary learning techniques. These middle-level properties are used as building blocks for explanations of image classifications. The achieved explanations are parsimonious, for their reliance on a limited set of middle-level image properties. And they can be contrastive, because the set of middle-level image properties can be used to explain why the system advanced the proposed classification over other antagonist classifications. In view of its model-agnostic character, the proposed framework is adaptable to a variety of other ML systems and explanation problems.

Sensitivity and specificity evaluation of multiple neurodegenerative proteins for Creutzfeldt-Jakob disease diagnosis using a deep-learning approach

The diagnosis of sporadic Creutzfeldt-Jakob disease (sCJD) can only be confirmed by abnormal protease-resistant prion protein accumulation in post-mortem brain tissue. The relationships between sCJD and cerebrospinal fluid (CSF) proteins such as 14–3-3, tau, and α-synuclein (a-syn) have been investigated for their potential value in pre-mortem diagnosis. Recently, deep-learning (DL) methods have attracted attention in neurodegenerative disease research. We established DL-aided pre-mortem diagnostic methods for CJD using multiple CSF biomarkers to improve their discriminatory sensitivity and specificity. Enzyme-linked immunosorbent assays were performed on phospho-tau (p-tau), total-tau (t-tau), a-syn, and β-amyloid (1–42), and western blot analysis was performed for 14–3-3 protein from CSF samples of 49 sCJD and 256 non-CJD Korean patients, respectively. The deep neural network structure comprised one input, five hidden, and one output layers, with 20, 40, 30, 20 and 12 hidden unit numbers per hidden layer, respectively. The best performing DL model demonstrated 90.38% accuracy, 83.33% sensitivity, and 92.5% specificity for the three-protein combination of t-tau, p-tau, and a-syn, and all other patients in a separate CSF set (n = 15) with other neuronal diseases were correctly predicted to not have CJD. Thus, DL-aided pre-mortem diagnosis may provide a suitable tool for discriminating CJD patients from non-CJD patients.

Higher Sensitivity and Reproducibility of Wavelet-Based Amplitude of Resting-State fMRI

The fast Fourier transform (FFT) is a widely used algorithm used to depict the amplitude of low-frequency fluctuation (ALFF) of resting-state functional magnetic resonance imaging (RS-fMRI). Wavelet transform (WT) is more effective in representing the complex waveform due to its adaptivity to non-stationary or local features of data and many varieties of wavelet functions with different shapes being available. However, there is a paucity of RS-fMRI studies that systematically compare between the results of FFT versus WT. The present study employed five cohorts of datasets and compared the sensitivity and reproducibility of FFT-ALFF with those of Wavelet-ALFF based on five mother wavelets (namely, db2, bior4.4, morl, meyr, and sym3). In addition to the conventional frequency band of 0.0117-0.0781 Hz, a comparison was performed in sub-bands, namely, Slow-6 (0-0.0117 Hz), Slow-5 (0.0117-0.0273 Hz), Slow-4 (0.0273-0.0742 Hz), Slow-3 (0.0742-0.1992 Hz), and Slow-2 (0.1992-0.25 Hz). The results indicated that the Wavelet-ALFF of all five mother wavelets was generally more sensitive and reproducible than FFT-ALFF in all frequency bands. Specifically, in the higher frequency band Slow-2 (0.1992-0.25 Hz), the mean sensitivity of db2-ALFF results was 1.54 times that of FFT-ALFF, and the reproducibility of db2-ALFF results was 2.95 times that of FFT-ALFF. The findings suggest that wavelet-ALFF can replace FFT-ALFF, especially in the higher frequency band. Future studies should test more mother wavelets on other RS-fMRI metrics and multiple datasets.

A Comprehensive Machine-Learning-Based Software Pipeline to Classify EEG Signals: A Case Study on PNES vs. Control Subjects

The diagnosis of psychogenic nonepileptic seizures (PNES) by means of electroencephalography (EEG) is not a trivial task during clinical practice for neurologists. No clear PNES electrophysiological biomarker has yet been found, and the only tool available for diagnosis is video EEG monitoring with recording of a typical episode and clinical history of the subject. In this paper, a data-driven machine learning (ML) pipeline for classifying EEG segments (i.e., epochs) of PNES and healthy controls (CNT) is introduced. This software pipeline consists of a semiautomatic signal processing technique and a supervised ML classifier to aid clinical discriminative diagnosis of PNES by means of an EEG time series. In our ML pipeline, statistical features like the mean, standard deviation, kurtosis, and skewness are extracted in a power spectral density (PSD) map split up in five conventional EEG rhythms (delta, theta, alpha, beta, and the whole band, i.e., 1-32 Hz). Then, the feature vector is fed into three different supervised ML algorithms, namely, the support vector machine (SVM), linear discriminant analysis (LDA), and Bayesian network (BN), to perform EEG segment classification tasks for CNT vs. PNES. The performance of the pipeline algorithm was evaluated on a dataset of 20 EEG signals (10 PNES and 10 CNT) that was recorded in eyes-closed resting condition at the Regional Epilepsy Centre, Great Metropolitan Hospital of Reggio Calabria, University of Catanzaro, Italy. The experimental results showed that PNES vs. CNT discrimination tasks performed via the ML algorithm and validated with random split (RS) achieved an average accuracy of 0.97 ± 0.013 (RS-SVM), 0.99 ± 0.02 (RS-LDA), and 0.82 ± 0.109 (RS-BN). Meanwhile, with leave-one-out (LOO) validation, an average accuracy of 0.98 ± 0.0233 (LOO-SVM), 0.98 ± 0.124 (LOO-LDA), and 0.81 ± 0.109 (LOO-BN) was achieved. Our findings showed that BN was outperformed by SVM and LDA. The promising results of the proposed software pipeline suggest that it may be a valuable tool to support existing clinical diagnosis.

Predicting Improved Daily Use of the More Affected Arm Poststroke Following Constraint-Induced Movement Therapy

BACKGROUND

Constraint-induced movement therapy (CI therapy) produces, on average, large and clinically meaningful improvements in the daily use of a more affected upper extremity in individuals with hemiparesis. However, individual responses vary widely.

OBJECTIVE

The study objective was to investigate the extent to which individual characteristics before treatment predict improved use of the more affected arm following CI therapy.

DESIGN

This study was a retrospective analysis of 47 people who had chronic (> 6 months) mild to moderate upper extremity hemiparesis and were consecutively enrolled in 2 CI therapy randomized controlled trials.

METHODS

An enhanced probabilistic neural network model predicted whether individuals showed a low, medium, or high response to CI therapy, as measured with the Motor Activity Log, on the basis of the following baseline assessments: Wolf Motor Function Test, Semmes-Weinstein Monofilament Test of touch threshold, Motor Activity Log, and Montreal Cognitive Assessment. Then, a neural dynamic classification algorithm was applied to improve prognostic accuracy using the most accurate combination obtained in the previous step.

RESULTS

Motor ability and tactile sense predicted improvement in arm use for daily activities following intensive upper extremity rehabilitation with an accuracy of nearly 100%. Complex patterns of interaction among these predictors were observed.

LIMITATIONS

The fact that this study was a retrospective analysis with a moderate sample size was a limitation.

CONCLUSIONS

Advanced machine learning/classification algorithms produce more accurate personalized predictions of rehabilitation outcomes than commonly used general linear models.

Scaled Subprofile Modeling and Convolutional Neural Networks for the Identification of Parkinson’s Disease in 3D Nuclear Imaging Data

Over the last years convolutional neural networks (CNNs) have shown remarkable results in different image classification tasks, including medical imaging. One area that has been less explored with CNNs is Positron Emission Tomography (PET). Fluorodeoxyglucose Positron Emission Tomography (FDG-PET) is a PET technique employed to obtain a representation of brain metabolic function. In this study we employed 3D CNNs in FDG-PET brain images with the purpose of discriminating patients diagnosed with Parkinson’s disease (PD) from controls. We employed Scaled Subprofile Modeling using Principal Component Analysis as a preprocessing step to focus on specific brain regions and limit the number of voxels that are used as input for the CNNs, thereby increasing the signal-to-noise ratio in our data. We performed hyperparameter optimization on three CNN architectures to estimate the classification accuracy of the networks on new data. The best performance that we obtained was [Formula: see text] and area under the receiver operating characteristic curve [Formula: see text] on the test set. We believe that, with larger datasets, PD patients could be reliably distinguished from controls by FDG-PET scans alone and that this technique could be applied to more clinically challenging tasks, like the differential diagnosis of neurological disorders with similar symptoms, such as PD, Progressive Supranuclear Palsy (PSP) and Multiple System Atrophy (MSA).

Performing Multi-Target Regression via a Parameter Sharing-Based Deep Network

Multi-target regression (MTR) comprises the prediction of multiple continuous target variables from a common set of input variables. There are two major challenges when addressing the MTR problem: the exploration of the inter-target dependencies and the modeling of complex input–output relationships. This paper proposes a neural network model that is able to simultaneously address these two challenges in a flexible way. A deep architecture well suited for learning multiple continuous outputs is designed, providing some flexibility to model the inter-target relationships by sharing network parameters as well as the possibility to exploit target-specific patterns by learning a set of nonshared parameters for each target. The effectiveness of the proposal is analyzed through an extensive experimental study on 18 datasets, demonstrating the benefits of using a shared representation that exploits the commonalities between target variables. According to the experimental results, the proposed model is competitive with respect to the state-of-the-art in MTR.

Descriptor Selection Improvements for Quantitative Structure-Activity Relationships

Molecular descriptor selection is an essential procedure to improve a predictive quantitative structure–activity relationship (QSAR) model. However, within the QSAR model, there are a number of redundant, noisy and irrelevant descriptors. In this study, we propose a novel descriptor selection framework using self-paced learning (SPL) via sparse logistic regression (LR) with Logsum penalty (SPL-Logsum), which can simultaneously adaptively identify the simple and complex samples and avoid over-fitting. SPL is inspired by the learning process of humans or animals gradually learned from simple and complex samples to train models, and the Logsum penalized LR helps to select a small subset of significant molecular descriptors for improving the QSAR models. Experimental results on some simulations and three public QSAR datasets show that our proposed SPL-Logsum framework outperforms other existing sparse methods regarding the area under the curve, sensitivity, specificity, accuracy, and [Formula: see text]-values.

Reachability Analysis of Neural Masses and Seizure Control Based on Combination Convolutional Neural Network

Epileptic seizures arise from synchronous firing of multiple spatially separated neural masses; therefore, many synchrony measures are used for seizure detection and characterization. However, synchrony measures reflect only the overall interaction strength among populations of neurons but cannot reveal the coupling strengths among individual populations, which is more important for seizure control. The concepts of reachability and reachable cluster were proposed to denote the coupling strengths of a set of neural masses. Here, we describe a seizure control method based on coupling strengths using combination convolutional neural network (CCNN) modeling. The neurophysiologically based neural mass model (NMM), which can bridge signal processing and neurophysiology, was used to simulate the proposed controller. Although the adjacency matrix and reachability matrix could not be identified perfectly, the vast majority of adjacency values were identified, reaching 95.64% using the CCNN with an optimal threshold. For cases of discrete and continuous coupling strengths, the proposed controller maintained the average reachable cluster strengths at about 0.1, indicating effective seizure control.

Segmentation and clustering in brain MRI imaging

Clustering is a vital task in magnetic resonance imaging (MRI) brain imaging and plays an important role in the reliability of brain disease detection, diagnosis, and effectiveness of the treatment. Clustering is used in processing and analysis of brain images for different tasks, including segmentation of brain regions and tissues (grey matter, white matter, and cerebrospinal fluid) and clustering of the atrophy in different parts of the brain. This paper presents a state-of-the-art review of brain MRI studies that use clustering techniques for different tasks.

Tonic Cold Pain Detection Using Choi–Williams Time-Frequency Distribution Analysis of EEG Signals: A Feasibility Study

Detecting pain based on analyzing electroencephalography (EEG) signals can enhance the ability of caregivers to characterize and manage clinical pain. However, the subjective nature of pain and the nonstationarity of EEG signals increase the difficulty of pain detection using EEG signals analysis. In this work, we present an EEG-based pain detection approach that analyzes the EEG signals using a quadratic time-frequency distribution, namely the Choi–Williams distribution (CWD). The use of the CWD enables construction of a time-frequency representation (TFR) of the EEG signals to characterize the time-varying spectral components of the EEG signals. The TFR of the EEG signals is analyzed to extract 12 time-frequency features for pain detection. These features are used to train a support vector machine classifier to distinguish between EEG signals that are associated with the no-pain and pain classes. To evaluate the performance of our proposed approach, we have recorded EEG signals for 24 healthy subjects under tonic cold pain stimulus. Moreover, we have developed two performance evaluation procedures—channel- and feature-based evaluation procedures—to study the effect of the utilized EEG channels and time-frequency features on the accuracy of pain detection. The experimental results show that our proposed approach achieved an average classification accuracy of 89.24% in distinguishing between the no-pain and pain classes. In addition, the classification performance achieved using our proposed approach outperforms the classification results reported in several existing EEG-based pain detection approaches.

Multivariate Approach for Alzheimer's Disease Detection Using Stationary Wavelet Entropy and Predator-Prey Particle Swarm Optimization

BACKGROUND

The number of patients with Alzheimer's disease is increasing rapidly every year. Scholars often use computer vision and machine learning methods to develop an automatic diagnosis system.

OBJECTIVE

In this study, we developed a novel machine learning system that can make diagnoses automatically from brain magnetic resonance images.

METHODS

First, the brain imaging was processed, including skull stripping and spatial normalization. Second, one axial slice was selected from the volumetric image, and stationary wavelet entropy (SWE) was done to extract the texture features. Third, a single-hidden-layer neural network was used as the classifier. Finally, a predator-prey particle swarm optimization was proposed to train the weights and biases of the classifier.

RESULTS

Our method used 4-level decomposition and yielded 13 SWE features. The classification yielded an overall accuracy of 92.73±1.03%, a sensitivity of 92.69±1.29%, and a specificity of 92.78±1.51%. The area under the curve is 0.95±0.02. Additionally, this method only cost 0.88 s to identify a subject in online stage, after its volumetric image is preprocessed.

CONCLUSION

In terms of classification performance, our method performs better than 10 state-of-the-art approaches and the performance of human observers. Therefore, this proposed method is effective in the detection of Alzheimer's disease.

Deep learning-based electroencephalography analysis: a systematic review

CONTEXT

Electroencephalography (EEG) is a complex signal and can require several years of training, as well as advanced signal processing and feature extraction methodologies to be correctly interpreted. Recently, deep learning (DL) has shown great promise in helping make sense of EEG signals due to its capacity to learn good feature representations from raw data. Whether DL truly presents advantages as compared to more traditional EEG processing approaches, however, remains an open question.

OBJECTIVE

In this work, we review 154 papers that apply DL to EEG, published between January 2010 and July 2018, and spanning different application domains such as epilepsy, sleep, brain-computer interfacing, and cognitive and affective monitoring. We extract trends and highlight interesting approaches from this large body of literature in order to inform future research and formulate recommendations.

METHODS

Major databases spanning the fields of science and engineering were queried to identify relevant studies published in scientific journals, conferences, and electronic preprint repositories. Various data items were extracted for each study pertaining to (1) the data, (2) the preprocessing methodology, (3) the DL design choices, (4) the results, and (5) the reproducibility of the experiments. These items were then analyzed one by one to uncover trends.

RESULTS

Our analysis reveals that the amount of EEG data used across studies varies from less than ten minutes to thousands of hours, while the number of samples seen during training by a network varies from a few dozens to several millions, depending on how epochs are extracted. Interestingly, we saw that more than half the studies used publicly available data and that there has also been a clear shift from intra-subject to inter-subject approaches over the last few years. About [Formula: see text] of the studies used convolutional neural networks (CNNs), while [Formula: see text] used recurrent neural networks (RNNs), most often with a total of 3-10 layers. Moreover, almost one-half of the studies trained their models on raw or preprocessed EEG time series. Finally, the median gain in accuracy of DL approaches over traditional baselines was [Formula: see text] across all relevant studies. More importantly, however, we noticed studies often suffer from poor reproducibility: a majority of papers would be hard or impossible to reproduce given the unavailability of their data and code.

SIGNIFICANCE

To help the community progress and share work more effectively, we provide a list of recommendations for future studies and emphasize the need for more reproducible research. We also make our summary table of DL and EEG papers available and invite authors of published work to contribute to it directly. A planned follow-up to this work will be an online public benchmarking portal listing reproducible results.

A Machine Learning Approach to Reveal the NeuroPhenotypes of Autisms

Although much research has been undertaken, the spatial patterns, developmental course, and sexual dimorphism of brain structure associated with autism remains enigmatic. One of the difficulties in investigating differences between the sexes in autism is the small sample sizes of available imaging datasets with mixed sex. Thus, the majority of the investigations have involved male samples, with females somewhat overlooked. This paper deploys machine learning on partial least squares feature extraction to reveal differences in regional brain structure between individuals with autism and typically developing participants. A four-class classification problem (sex and condition) is specified, with theoretical restrictions based on the evaluation of a novel upper bound in the resubstitution estimate. These conditions were imposed on the classifier complexity and feature space dimension to assure generalizable results from the training set to test samples. Accuracies above [Formula: see text] on gray and white matter tissues estimated from voxel-based morphometry (VBM) features are obtained in a sample of equal-sized high-functioning male and female adults with and without autism ([Formula: see text], [Formula: see text]/group). The proposed learning machine revealed how autism is modulated by biological sex using a low-dimensional feature space extracted from VBM. In addition, a spatial overlap analysis on reference maps partially corroborated predictions of the “extreme male brain” theory of autism, in sexual dimorphic areas.

A Machine Learning Approach to Characterize the Modulation of the Hippocampal Rhythms Via Optogenetic Stimulation of the Medial Septum

The medial septum (MS) is a potential target for modulating hippocampal activity. However, given the multiple cell types involved, the changes in hippocampal neural activity induced by MS stimulation have not yet been fully characterized. We combined MS optogenetic stimulation with local field potential (LFP) recordings from the hippocampus and leveraged machine learning techniques to explore how activating or inhibiting multiple MS neuronal subpopulations using different optical stimulation parameters affects hippocampal LFP biomarkers. First, of the seven different optogenetic viral vectors used for modulating different neuronal subpopulations, only two induced a substantial change in hippocampal LFP. Second, we found hippocampal low-gamma band to be most effectively modulated by the stimulation. Third, the hippocampal biomarkers were sensitive to the optogenetic virus type and the stimulation frequency, establishing those parameters as the critical ones for the regulation of hippocampal biomarker activity. Last, we built a Gaussian process regression model to describe the relationship between stimulation parameters and activity of the biomarker as well as to identify the optimal parameters for biomarker modulation. This new machine learning approach can further our understanding of the effects of neural stimulation and guide the selection of optimal parameters for neural control.

Early Detection of Parkinson’s Disease by Using SPECT Imaging and Biomarkers

Precise and timely diagnosis of Parkinson’s disease is important to control its progression among subjects. Currently, a neuroimaging technique called dopaminergic imaging that uses single photon emission computed tomography (SPECT) with 123I-Ioflupane is popular among clinicians for detecting Parkinson’s disease in early stages. Unlike other studies, which consider only low-level features like gray matter, white matter, or cerebrospinal fluid, this study explores the non-linear relation between different biomarkers (SPECT + biological) using deep learning and multivariate logistic regression. Striatal binding ratios are obtained using 123I-Ioflupane SPECT scans from four brain regions which are further integrated with five biological biomarkers to increase the diagnostic accuracy. Experimental results indicate that this investigated approach can differentiate subjects with 100% accuracy. The obtained results outperform the ones reported in the literature. Furthermore, logistic regression model has been developed for estimating the Parkinson’s disease onset probability. Such models may aid clinicians in diagnosing this disease.

Latent Phase Detection of Hypoxic-Ischemic Spike Transients in the EEG of Preterm Fetal Sheep Using Reverse Biorthogonal Wavelets & Fuzzy Classifier

Hypoxic-ischemic (HI) studies in preterms lack reliable prognostic biomarkers for diagnostic tests of HI encephalopathy (HIE). Our group's observations from in utero fetal sheep models suggest that potential biomarkers of HIE in the form of developing HI micro-scale epileptiform transients emerge along suppressed EEG/ECoG background during a latent phase of 6-7h post-insult. However, having to observe for the whole of the latent phase disqualifies any chance of clinical intervention. A precise automatic identification of these transients can help for a well-timed diagnosis of the HIE and to stop the spread of the injury before it becomes irreversible. This paper reports fusion of Reverse-Biorthogonal Wavelets with Type-1 Fuzzy classifiers, for the accurate real-time automatic identification and quantification of high-frequency HI spike transients in the latent phase, tested over seven in utero preterm sheep. Considerable high performance of 99.78 ± 0.10% was obtained from the Rbio-Wavelet Type-1 Fuzzy classifier for automatic identification of HI spikes tested over 42h of high-resolution recordings (sampling-freq:1024Hz). Data from post-insult automatic time-localization of high-frequency HI spikes reveals a promising trend in the average rate of the HI spikes, even in the animals with shorter occlusion periods, which highlights considerable higher number of transients within the first 2h post-insult.