electroencephalography-network-indices-as-biomarkers-upper-limb

Electroencephalography Network Indices as Biomarkers of Upper Limb Impairment in Chronic Stroke

Alteration of electroencephalography (EEG) signals during task-specific movement of the impaired limb has been reported as a potential biomarker for the ...

Korea University

Abnormal facial emotion recognition is considered as one of the key symptoms of schizophrenia. Only few studies have considered deficits in the spatial frequency (SF)-dependent visual pathway leading to abnormal facial emotion recognition in schizophrenia. Twenty-one patients with schizophrenia and 19 matched healthy controls (HC) were recruited for this study. Event-related potentials (ERP) were measured during presentation of SF-modulated face stimuli and their source imaging was analyzed. The patients showed reduced P100 amplitude for low-spatial frequency (LSF) pictures of fearful faces compared with the HC group. The P100 amplitude for high-spatial frequency (HSF) pictures of neutral faces was increased in the schizophrenia group, but not in the HC group. The neural source activities of the LSF fearful faces and HSF neutral faces led to hypo- and hyperactivation of the frontal lobe of subjects from the schizophrenia group and HC group, respectively. In addition, patients with schizophrenia showed enhanced N170 activation in the right hemisphere in the LSF condition, while the HC group did not. Our results suggest that deficits in the LSF-dependent visual pathway, which involves magnocellular neurons, impair early visual processing leading to dysfunctional facial emotion recognition in schizophrenia. Moreover, it suggests impaired bottom-up processing rather than top-down dysfunction for facial emotion recognition in these patients.

Early visual processing deficits in patients with schizophrenia during spatial frequency-dependent facial affect processing.

Women tend to respond to emotional stimuli differently from men. This study aimed at investigating whether neural responses to perceptually "invisible" emotional stimuli differ between men and women by exploiting event-related potential (ERP). Forty healthy participants (21 women) were recruited for the main experiment. A control experiment was conducted by excluding nine (7 women) participants from the main experiment and replacing them with additional ten (6 women) participants (total 41 participants) where Beck's Anxiety Inventory (BAI) and Beck's Depression Inventory (BDI) scores were controlled. Using the visual backward masking paradigm, either a fearful or a neutral face stimulus was presented in varied durations (subthreshold, near-threshold, or suprathreshold) followed by a mask. Participants performed a two-alternative forced choice (2-AFC) emotion discrimination task on each face. Behavioral analysis showed that participants were unaware of masked stimuli of which duration was the shortest and, therefore, processed at subthreshold. Nevertheless, women showed significantly larger response in P100 amplitude to subthreshold fearful faces than men. This result remained consistent in the control experiment. Our findings indicate gender-differences in neural response to subthreshold emotional face, which is reflected in the early processing stage.

Gender Differences in Neural Responses to Perceptually Invisible Fearful Face-An ERP Study.

Pathologies of schizophrenia and bipolar disorder have been poorly understood. Brain network analysis could help understand brain mechanisms of schizophrenia and bipolar disorder. This study investigates the source-level brain cortical networks using resting-state electroencephalography (EEG) in patients with schizophrenia and bipolar disorder.

Altered Cortical Functional Networks in Patients With Schizophrenia and Bipolar Disorder: A Resting-State Electroencephalographic Study.

To what extent are different levels of expertise reflected in the functional connectivity of the brain? We addressed this question by using resting-state functional magnetic resonance imaging (fMRI) in mathematicians versus non-mathematicians. To this end, we investigated how the two groups of participants differ in the correlation of their spontaneous blood oxygen level-dependent fluctuations across the whole brain regions during resting state. Moreover, by using the classification algorithm in machine learning, we investigated whether the resting-state fMRI networks between mathematicians and non-mathematicians were distinguished depending on features of functional connectivity. We showed diverging involvement of the frontal-thalamic-temporal connections for mathematicians and the medial-frontal areas to precuneus and the lateral orbital gyrus to thalamus connections for non-mathematicians. Moreover, mathematicians who had higher scores in mathematical knowledge showed a weaker connection strength between the left and right caudate nucleus, demonstrating the connections' characteristics related to mathematical expertise. Separate functional networks between the two groups were validated with a maximum classification accuracy of 91.19% using the distinct resting-state fMRI-based functional connectivity features. We suggest the advantageous role of preconfigured resting-state functional connectivity, as well as the neural efficiency for experts' successful performance.

Resting-State Functional Connectivity in Mathematical Expertise.

In recent years, machine learning techniques have been frequently applied to uncovering neuropsychiatric biomarkers with the aim of accurately diagnosing neuropsychiatric diseases and predicting treatment prognosis. However, many studies did not perform cross validation (CV) when using machine learning techniques, or others performed CV in an incorrect manner, leading to significantly biased results due to overfitting problem. The aim of this study is to investigate the impact of CV on the prediction performance of neuropsychiatric biomarkers, in particular, for feature selection performed with high-dimensional features. To this end, we evaluated prediction performances using both simulation data and actual electroencephalography (EEG) data. The overall prediction accuracies of the feature selection method performed outside of CV were considerably higher than those of the feature selection method performed within CV for both the simulation and actual EEG data. The differences between the prediction accuracies of the two feature selection approaches can be thought of as the amount of overfitting due to selection bias. Our results indicate the importance of correctly using CV to avoid biased results of prediction performance of neuropsychiatric biomarkers.

Inflated prediction accuracy of neuropsychiatric biomarkers caused by data leakage in feature selection.

 Impaired movement after stroke is closely associated with altered brain functions, and thus the investigation on neural substrates of patients with stroke can pave a way for not only understanding the underlying mechanisms of neuropathological traits, but also providing an innovative solution for stroke rehabilitation. The objective of this study was to precisely investigate altered brain functions in terms of power spectral and brain network analyses.  Altered brain function was investigated by using electroencephalography (EEG) measured while 34 patients with chronic stroke performed movement tasks with the affected and unaffected hands. The relationships between functional brain network indices and Fugl-Meyer Assessment (FMA) scores were also investigated.  A stronger low-beta event-related desynchronization was found in the contralesional hemisphere for both affected and unaffected movement tasks compared with that of the ipsilesional hemisphere. More efficient whole-brain networks (increased strength and clustering coefficient, and prolonged path length) in the low-beta frequency band were revealed when moving the unaffected hand compared with when moving the affected hand. In addition, the brain network indices of the contralesional hemisphere indicated higher efficiency and cost-effectiveness than those of the ipsilesional hemisphere in both the alpha and low-beta frequency bands. Moreover, the alpha network indices (strength, clustering coefficient, path length, and small-worldness) were significantly correlated with the FMA scores.  Efficient functional brain network indices are associated with better motor outcomes in patients with stroke and could be useful biomarkers to monitor stroke recovery during rehabilitation.

Altered Functional Networks of Alpha and Low-Beta Bands During Upper Limb Movement and Association with Motor Impairment in Chronic Stroke.

Editorial: AI-Based Computer-Aided Diagnosis and Prognosis for Psychiatric Disorders.

Electroencephalography (EEG)-based diagnosis of psychiatric diseases using machine-learning approaches has made possible the objective diagnosis of various psychiatric diseases. The objective of this study was to improve the performance of a resting-state EEG-based computer-aided diagnosis (CAD) system to diagnose post-traumatic stress disorder (PTSD), by optimizing the frequency bands used to extract EEG features. We used eyes-closed resting-state EEG data recorded from 77 PTSD patients and 58 healthy controls (HC). Source-level power spectrum densities (PSDs) of the resting-state EEG data were extracted from 6 frequency bands (delta, theta, alpha, low-beta, high-beta, and gamma), and the PSD features of each frequency band and their combinations were independently used to discriminate PTSD and HC. The classification performance was evaluated using support vector machine with leave-one-out cross validation. The PSD features extracted from slower-frequency bands (delta and theta) showed significantly higher classification performance than those of relatively higher-frequency bands. The best classification performance was achieved when using delta PSD features (86.61%), which was significantly higher than that reported in a recent study by about 13%. The PSD features selected to obtain better classification performances could be explained from a neurophysiological point of view, demonstrating the promising potential to develop a clinically reliable EEG-based CAD system for PTSD diagnosis.

Enhanced Performance by Interpretable Low-Frequency Electroencephalogram Oscillations in the Machine Learning-Based Diagnosis of Post-traumatic Stress Disorder.

Impaired cognitive function related to intrusive memories of traumatic experiences is the most noticeable characteristic of post-traumatic stress disorder (PTSD); nevertheless, the brain mechanism involved in the cognitive processing is still elusive. To improve the understanding of the neuropathology in PTSD patients, we investigated functional cortical networks that are based on graph theory, by using electroencephalogram (EEG). EEG signals, elicited by an auditory oddball paradigm, were recorded from 53 PTSD patients and 39 healthy controls (HCs). Source signals in 68 regions of interests were estimated using EEG data for each subject using minimum-norm estimation. Then, using source signals of each subject, time-frequency analysis was conducted, and a functional connectivity matrix was constructed using the imaginary part of coherence, which was used to evaluate three global-level (strength, clustering coefficient, and path length) and two nodal-level (strength and clustering coefficients) network indices in four frequency bands (theta, alpha, low-beta, and high-beta). The relationships between the network indices and symptoms were evaluated using Pearson's correlation. Compared with HCs, PTSD patients showed significantly reduced spectral powers around P300 periods and significantly altered network indices (diminished strength and clustering coefficient, and prolonged path length) in theta frequency band. In addition, the nodal strengths and nodal clustering coefficients in theta band of PTSD patients were significantly reduced, compared with those of HCs, and the reduced nodal clustering coefficients in parieto-temporo-occipital regions had negative correlations with the symptom scores (Impact of Event Scale-Revises, Beck Depression Inventory, and Beck Anxiety Inventory). The characterization of this disrupted pattern improves the understanding of the neuropathophysiology underlying the impaired cognitive function in PTSD patients.

Impaired functional cortical networks in the theta frequency band of patients with post-traumatic stress disorder during auditory-cognitive processing.

The use of EEG for evaluating and diagnosing neurological abnormalities related to psychiatric diseases and identifying human emotions has been improved by deep learning advancements. This research aims to categorize individuals with schizophrenia (SZ), their biological relatives (REL), and healthy controls (HC) using resting EEG brain source signal data defined by regions of interest (ROIs). The proposed solution is a deep neural network for the cortical source signals of the ROIs, incorporating a Squeeze-and-Excitation Block and multiple CNNs designed for eyes-open and closed resting states. The model, called EEG Temporal Spatial Network (ETSNet), has two variants: ETSNet and ETSNet. Two evaluations were conducted to show the effectiveness of the proposed model. The average accuracy for the classification of SZ, REL, and HC using EEG resting data was 99.57% (ETSNet), and the average accuracy for the classification of eyes-open (EO) and eyes-closed (EC) resting states was 93.15% (ETSNet). An ablation study was also conducted using two public datasets for intellectual and developmental disorders and emotional states, showing improved classification accuracy compared to advanced EEG classification algorithms when using ETSNet.

ETSNet: A deep neural network for EEG-based temporal-spatial pattern recognition in psychiatric disorder and emotional distress classification.

A machine-learning-based computer-aided diagnosis (CAD) system can complement the traditional diagnostic error for major depressive disorder (MDD) using trait-like neurophysiological biomarkers. Previous studies have shown that the CAD system has the potential to differentiate between female MDD patients and healthy controls. The aim of this study was to develop a practically useful resting-state electroencephalography (EEG)-based CAD system to assist in the diagnosis of drug-naïve female MDD patients by considering both the drug and gender effects. In addition, the feasibility of the practical use of the resting-state EEG-based CAD system was evaluated using a channel reduction approach.

Article	Year
Early visual processing deficits in patients with schizophrenia during spatial frequency-dependent facial affect processing. Schizophrenia research\| PubMed ID: 25553978	2015
Gender Differences in Neural Responses to Perceptually Invisible Fearful Face-An ERP Study. Frontiers in behavioral neuroscience\| PubMed ID: 28184189	2017
Altered Cortical Functional Networks in Patients With Schizophrenia and Bipolar Disorder: A Resting-State Electroencephalographic Study. Frontiers in psychiatry\| PubMed ID: 32774308	2020
Resting-State Functional Connectivity in Mathematical Expertise. Brain sciences\| PubMed ID: 33800679	2021
Inflated prediction accuracy of neuropsychiatric biomarkers caused by data leakage in feature selection. Scientific reports\| PubMed ID: 33846489	2021
Altered Functional Networks of Alpha and Low-Beta Bands During Upper Limb Movement and Association with Motor Impairment in Chronic Stroke. Brain connectivity\| PubMed ID: 34269616	2021
Editorial: AI-Based Computer-Aided Diagnosis and Prognosis for Psychiatric Disorders. Frontiers in human neuroscience\| PubMed ID: 35517988	2022
Enhanced Performance by Interpretable Low-Frequency Electroencephalogram Oscillations in the Machine Learning-Based Diagnosis of Post-traumatic Stress Disorder. Frontiers in neuroinformatics\| PubMed ID: 35571868	2022
Impaired functional cortical networks in the theta frequency band of patients with post-traumatic stress disorder during auditory-cognitive processing. Frontiers in psychiatry\| PubMed ID: 36032254	2022
ETSNet: A deep neural network for EEG-based temporal-spatial pattern recognition in psychiatric disorder and emotional distress classification. Computers in biology and medicine\| PubMed ID: 37044046	2023
Toward practical machine-learning-based diagnosis for drug-naïve women with major depressive disorder using EEG channel reduction approach. Journal of affective disorders\| PubMed ID: 37302509	2023

Miseon Shim

Biography

JoVE Journal Publications

Other Publications