Research Repository

Towards Decoding of Depersonalisation Disorder Using EEG: A Time Series Analysis Using CDTW

Salami, Abbas and Andreu-Perez, Javier and Gillmeister, Helge (2020) Towards Decoding of Depersonalisation Disorder Using EEG: A Time Series Analysis Using CDTW. In: IEEE Symposium Series on Computational Intelligence, 2020-12-01 - 2020-12-04, Canberra, Australia. (In Press)

[img]
Preview
Text
Towards-Decoding-Salami.pdf - Accepted Version

Download (899kB) | Preview

Abstract

Depersonalisation/derealisation refers to a transient psychological condition characterised by losing the sense of body ownership and feeling detached from the outside world. It is often accompanied by a lack of emotional responsiveness and sometimes memory fragmentation. Studies have shown the temporary occurrence of this condition among 34-70% of the general population during their life span. However, if the symptoms become consistent, they can be intolerable and can profoundly affect the quality of life in such an extent that it would be considered as one type of the dissociative disorders, depersonalisation disorder (DPD). Currently, there is no laboratory method to diagnose DPD, and studies have expressed a period of seven to 12 years for the correct diagnosis of DPD. We recently aimed to investigate DPD and its symptoms based on inexpensive and convenient electroencephalogram (EEG) neuroimaging technique, using calculation of event-related-potentials (ERPs) over the somatosensory cortex. We showed that DPD symptoms could be as a result of impairment in early (implicit) stages of information processing in the brain. We also introduced P45 as a potential electrophysiological biomarker to study DPD. In this paper, we first replicated our results and then used P45 as a feature to discriminate between individuals with high and low tendency to DPD symptoms. We used Continuous Dynamic Time Warping (CDTW) to address the possible time shift and distortion in the ERP signals and to reach better classification performance. We reached 85% accuracy (Kappa 0.7) using leave-one-subject-out cross-validation, which confirms the feasibility for discrimination between DPD patients and a control group using EEG signals.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Published proceedings: _not provided_
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Faculty of Science and Health > Psychology, Department of
Depositing User: Elements
Date Deposited: 04 Dec 2020 19:36
Last Modified: 04 Dec 2020 20:15
URI: http://repository.essex.ac.uk/id/eprint/28920

Actions (login required)

View Item View Item