Machine-learning-based diagnosis of schizophrenia using combined sensor-level and source-level EEG features

Shim, Miseon; Hwang, Han-Jeong; Kim, Do-Won; Lee, Seung Hwan; Im, Chang‐Hwan

doi:10.1016/j.schres.2016.05.007

Cited by 152 publications

(90 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…36 Data were filtered using a 1-50-Hz bandpass filter with pre-stimulus baseline correction and were epoched from late-onset 1-500 ms. To calculate the P300 alpha frequency, a minimum epoch length above 500 ms has been recommended. 37 The absolute power was calculated using a fast Fourier transform: delta (1-4 Hz), theta (4-8 Hz), low-frequency alpha (8-10 Hz), highfrequency alpha (10-12 Hz), low-frequency beta (12-18 Hz), highfrequency beta (18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), and gamma (30-50 Hz) signals. Artifacts exceeding AE100 μV were excluded at all electrode sites.…”

Section: Resting-state Eeg Paradigmmentioning

confidence: 99%

Unsuccessful reduction of high‐frequency alpha activity during cognitive activation in schizophrenia

Jang

Jung

et al. 2019

Psychiatry Clin Neurosci

Self Cite

View full text Add to dashboard Cite

Aims: Electroencephalogram (EEG) alpha activity during resting state reflects the 'readiness' of an individual to respond to the environment; this includes the performance of cognitive processes. Alpha activity is reported to be attenuated in schizophrenia (SCZ). Understanding the interaction between alpha activity during rest and when cognitively engaged may provide insights into the neural circuitry, which is dysfunctional in SCZ. This study investigated the changes of alpha activity between resting state and cognitive engagement in SCZ patients.Methods: Thirty-four SCZ patients and 29 healthy controls (HC) were recruited. EEG was performed in the resting state and during an auditory P300 task. All experimental procedures followed the relevant institutional guidelines and regulations.Results: In SCZ, high-frequency alpha activity was reduced in the resting state. High-frequency alpha source density was decreased in both the resting-state and a P300 task condition in patients, compared to HC. HC, but not SCZ patients, showed a reduction in high-frequency alpha source density during the P300 task compared to the resting state. The negative correlation between high-frequency alpha source density in the resting state and positive symptoms was significant.Conclusions: High-frequency alpha activity in SCZ patients and its unsuccessful reduction during cognitive processing may be biological markers of SCZ.

show abstract

Section: Resting-state Eeg Paradigmmentioning

confidence: 99%

Unsuccessful reduction of high‐frequency alpha activity during cognitive activation in schizophrenia

Jang

Jung

et al. 2019

Psychiatry Clin Neurosci

Self Cite

View full text Add to dashboard Cite

show abstract

“…Machine learning algorithms have been employed for automated classification of altered brain activity in SZ using EEG and fMRI data, primarily based on traditional classifiers such as support vector machine (SVM) [21][22][23][24], kernel discriminant analysis (KDA) [25] and adaptive boosting [26]. Moreover, most previous works on EEG-based classification of SZ used time-frequency features from single EEG channels such as band-specific spectral power and univariate autoregression model coefficients [26,27].…”

Section: Introductionmentioning

confidence: 99%

A Multi-Domain Connectome Convolutional Neural Network for Identifying Schizophrenia From EEG Connectivity Patterns

Phang

Noman

Hussain

et al. 2020

IEEE J. Biomed. Health Inform.

136

View full text Add to dashboard Cite

We exploit altered patterns in brain functional connectivity as features for automatic discriminative analysis of neuropsychiatric patients. Deep learning methods have been introduced to functional network classification only very recently for fMRI, and the proposed architectures essentially focused on a single type of connectivity measure. Methods: We propose a deep convolutional neural network (CNN) framework for classification of electroencephalogram (EEG)-derived brain connectome in schizophrenia (SZ). To capture complementary aspects of disrupted connectivity in SZ, we explore combination of various connectivity features consisting of time and frequencydomain metrics of effective connectivity based on vector autoregressive model and partial directed coherence, and complex network measures of network topology. We design a novel multidomain connectome CNN (MDC-CNN) based on a parallel ensemble of 1D and 2D CNNs to integrate the features from various domains and dimensions using different fusion strategies. We also consider an extension to dynamic brain connectivity using the recurrent neural networks. Results: Hierarchical latent representations learned by the multiple convolutional layers from EEG connectivity reveals apparent group differences between SZ and healthy controls (HC). Results on a large restingstate EEG dataset show that the proposed CNNs significantly outperform traditional support vector machine classifier. The MDC-CNN with combined connectivity features further improves performance over single-domain CNNs using individual features, achieving remarkable accuracy of 91.69% with a decision-level fusion. Conclusion: The proposed MDC-CNN by integrating information from diverse brain connectivity descriptors is able to accurately discriminate SZ from HC. Significance: The new framework is potentially useful for developing diagnostic tools for SZ and other disorders.

show abstract

“…Em uma análise mais generalista da produção cientí ca com foco na avaliação psicológica computadorizada, percebe-se que a grande maioria dos modelos preditivos construídos utilizam dados provenientes de exames bioquímicos, EEGs (eletroencefalogramas) e ressonâncias magnéticas (Hosseinifard et al;2013;Foland-Ross et al;Patel et al;Jiang et al;2016;Shim et al;2016;Li et al;2016;Hilbert et al;Zheng et al;. Apenas um dos trabalhos encontrados pelos autores empregou testes baseados em produção de desenhos como ferramenta de psicodiagnóstico (in Kim, Kang, Chung and joo Hong; 2012).…”

Section: Trabalhos Relacionadosunclassified

Melampus: um modelo deep learning para triagem psicológica infantil

Silva¹,

Raeder²

2018

RBCA

View full text Add to dashboard Cite

Embora problemas relacionados à saúde mental usualmente tenham início durante a infância ou adolescência, apenas uma pequena parcela desta população recebe diagnóstico e tratamento adequado. Uma das causas para a baixa taxa de identificação de desordens mentais é a falta de instrumentos especializados nesta tarefa, especialmente ferramentas que reduzam o custo e o tempo necessário para a execução de processos de triagem psicológica. Na literatura recente, muitos autores vêm analisando como o aprendizado de máquina pode contribuir para a construção de instrumentos de avaliação psicológica, contudo poucas pesquisas se propõem a construir ferramentas válidas para grupos compostos majoritariamente por crianças. O presente trabalho propõe um modelo de ferramenta para apoio à triagem psicológica infantil baseada em testes clínicos e deep learning. Tal modelo foi avaliado através de uma implementação que combina o uso de Redes Neurais Convolucionais e um sistema de escalas clínicas para avaliação do Desenho da Figura Humana. Os resultados apresentados pelos modelos de classificação treinados demonstraram bons índices de acerto considerando-se a pequena amostra disponível, o que sugere que ferramentas de deep learning podem ser adequadas para o cenário proposto.

show abstract

Machine-learning-based diagnosis of schizophrenia using combined sensor-level and source-level EEG features

Cited by 152 publications

References 42 publications

Unsuccessful reduction of high‐frequency alpha activity during cognitive activation in schizophrenia

Unsuccessful reduction of high‐frequency alpha activity during cognitive activation in schizophrenia

A Multi-Domain Connectome Convolutional Neural Network for Identifying Schizophrenia From EEG Connectivity Patterns

Melampus: um modelo deep learning para triagem psicológica infantil

Contact Info

Product

Resources

About