Classification of Low and High Schizotypy Levels via Evaluation of Brain Connectivity

Abstract

Schizotypy is a latent cluster of personality traits that denote a vulnerability for schizophrenia or a type of spectrum disorder. The aim of the study is to investigate parametric effective brain connectivity features for classifying high versus low schizotypy (LS) status. Electroencephalography (EEG) signals are recorded from 13 high schizotypy (HS) and 11 LS participants during an emotional auditory odd-ball task. The brain connectivity signals for machine learning are taken after the settlement of event-related potentials. A multivariate autoregressive (MVAR)-based connectivity measure is estimated from the EEG signals using the directed transfer functions (DTFs) method. The values of DTF power in five standard frequency bands are used as features. The support vector machines (SVMs) revealed significant differences between HS and LS. The accuracy, specificity, and sensitivity of the results using SVM are as high as 89.21%, 90.3%, and 88.2%, respectively. Our results demonstrate that the effective brain connectivity in prefrontal/parietal and prefrontal/frontal brain regions considerably changes according to schizotypal status. These findings prove that the brain connectivity indices offer valuable biomarkers for detecting schizotypal personality. Further monitoring of the changes in DTF following the diagnosis of schizotypy may lead to the early identification of schizophrenia and other spectrum disorders.

Auditory P300 in individuals with high schizotypy: associations of schizotypal traits with amplitude and latency under different oddball conditions

“…These findings are similar to those from clinical individuals with schizophrenia-spectrum disorders ( Wang et al, 2003 ; Bramon et al, 2004 ; Higuchi et al, 2013 , 2021 ; Kim et al, 2015 , 2020 ), and consistent with the previous non-clinical report ( Klein et al, 1999 ), indicating schizophrenia-related impairment of attention and working memory in acoustic information process ( Kim et al, 2015 ). This process impairment may be related to the poor information flow in individuals with high schizotypy ( Hu et al, 2020 ), which was supported by the evidence of machine learning study ( Zandbagleh et al, 2022a ), that reduced connectivity between prefrontal and parietal regions in the beta band, and decreased frontal connectivity in the alpha band in participants with high schizotypy.…”

Tensor factorization approach for ERP-based assessment of schizotypy in a novel auditory oddball task on perceived family stress

“…On the other hand, any delay in diagnosis and treatment of psychosis-like signs contributes to poorer outcomes in psychosis [55]. Technological advances in signal processing exploit the multi-modal brain responses to accurately classify the people with schizotypy, which is a sub-clinical personality trait akin to psychosis [37]. In this study, tensor factorization, a state-of-the-art method, has been used for accurate detection of P300 subcomponents, namely P3a and P3b.…”

“…Then, this procedure is repeated for all the participants. Some performance measures namely, accuracy, specificity, sensitivity, and F1-score are derived as follows [44]: × 100 (16) where TP is the number of HS participants classified correctly in the HS class, FP is the number of LS participants classified incorrectly as HS class, TN is the number of LS participants recognized correctly in the LS class, and FN is the number of HS participants recognized incorrectly as LS class.…”

“…They examined a combination of features including weighted phase lag index (wPLI) as a measure of brain connectivity to distinguish between two schizotypy groups. Quite recently, the authors have used directed transfer function (DTF) derived from multivariate autoregressive (MVAR) coefficients and achieved an accuracy of 89.21% to effectively classify the subjects into HS and LS categories for the first time [37].…”

A systematic review of EEG based automated schizophrenia classification through machine learning and deep learning

“…The studies in Vázquez et al (2021) and ( Guo et al, 2022 ), employing Butterworth and Vietoris–Rips filtering for preprocessing and RF classifier and Bottleneck and Wasserstein distances, respectively, for classification of SCZ, did not report their results in terms of accuracy. The methods in Jahmunah et al (2019) ; ( Khare and Bajaj, 2021 ; Rajesh and Sunil Kumar, 2021 ; Luján et al, 2022 ; Zandbagleh et al, 2022 ), and ( Aksöz et al, 2022 ) achieved moderate accuracy with different preprocessing techniques. However, the approaches in Neuhaus et al (2013) and ( Du et al, 2020 ) secure lower accuracies but still surpass the performance of Devia et al (2019) .…”