PhD Scientific Days 2023

Budapest, 22-23 June 2023

Mental Health Sciences II.

Multivariate Analysis of Cognitive Deficits and Their Electrophysiological Correlates in Schizophrenia

Melinda Becske1, Csilla Marosi1, Hajnalka Molnár1, Zsuzsanna Fodor1, László Tombor1, Gábor Csukly1
1 Semmelweis University, Department of Psychiatry and Psychotherapy, Budapest

Text of the abstract

Introduction: Although previous research has identified a number of potential behavioral and electroencephalographic (EEG) markers of schizophrenia, substantial individual differences exist in the patient population in terms of these markers, as well as clinical symptoms.
Aims: Our main goal was to identify the features in both modalities that best discriminate between patients and controls. We also aimed to investigate the relationship between a set of behavioral and EEG variables.
Method: 47 patients with schizophrenia and 42 matched controls participated in our research, under which we examined certain aspects of basic visual processing and face perception, as well as visual and acoustic Mismatch Negativity, and processes related to higher-level cognitive functions such as emotional face recognition and working memory. Resting state EEG functional connectivity strength and Minimum Spanning Tree measures were also analyzed. For statistical analysis, we applied multivariate methods that can effectively deal with high-dimensional data (regularized canonical correlation, group regularized canonical correlation, and stepwise logistic regression), and applied some machine learning algorithms.
Results: Based on the results of the stepwise logistic regression analysis, we have chosen 5 out of 76 variables. 10-fold cross-validation resulted in a classification accuracy of 82.5% (95% CI = 77% - 88%).
Conclusion: Our preliminary results show that disturbed visual processing (specifically related to magnocellular pathway deficit), as well as weaker resting-state functional connectivity in the alpha and delta frequency ranges, and some variables related to distractibility can be the most important features in the classification.
Funding: Supported by the ÚNKP-22-3-II New National Excellence Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund; Hungarian Research Found - OTKA PD 115837; Bolyai Research Fellowship Program of the Hungarian Academy of Sciences; Higher Education Institutional Excellence Programme of the Ministry of Human Capacities in Hungary, within the framework of the Neurology thematic programme of Semmelweis University.