Brain functional connectivity patterns for emotional state classification in Parkinson’s disease patients without dementia

• Published in: Behavioural brain research Vol. 298; no. Pt B; pp. 248 - 260
• Main Authors: Yuvaraj, R., Murugappan, M., Acharya, U. Rajendra, Adeli, Hojjat, Ibrahim, Norlinah Mohamed, Mesquita, Edgar
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.02.2016
• Subjects: Brain - physiopathology; Coherence; Correlation; EEG; Electroencephalography - methods; Emotion; Emotions - physiology; Female; Humans; Male; Middle Aged; Neural Pathways - physiopathology; Parkinson Disease - classification; Parkinson Disease - physiopathology; Parkinson Disease - psychology; Parkinson’s disease; PSI; Signal Processing, Computer-Assisted; Support Vector Machine; SVM; Correlation; EEG; Coherence; PSI; Emotion; SVM; Parkinson’s disease
• ISSN: 0166-4328; 1872-7549
• DOI: 10.1016/j.bbr.2015.10.036

•Recognize different emotional states using brain functional connectivity.•EEG change is significantly different among emotional states of PD patients.•Highest classification results for the proposed bispectral functional connectivity index.•PD patients exists decline in cortical connectivity during emotion processing. Successful emotional communication is crucial for social interactions and social relationships. Parkinson’s Disease (PD) patients have shown deficits in emotional recognition abilities although the research findings are inconclusive. This paper presents an investigation of six emotions (happiness, sadness, fear, anger, surprise, and disgust) of twenty non-demented (Mini-Mental State Examination score >24) PD patients and twenty Healthy Controls (HCs) using Electroencephalogram (EEG)-based Brain Functional Connectivity (BFC) patterns. The functional connectivity index feature in EEG signals is computed using three different methods: Correlation (COR), Coherence (COH), and Phase Synchronization Index (PSI). Further, a new functional connectivity index feature is proposed using bispectral analysis. The experimental results indicate that the BFC change is significantly different among emotional states of PD patients compared with HC. Also, the emotional connectivity pattern classified using Support Vector Machine (SVM) classifier yielded the highest accuracy for the new bispectral functional connectivity index. The PD patients showed emotional impairments as demonstrated by a poor classification performance. This finding suggests that decrease in the functional connectivity indices during emotional stimulation in PD, indicating functional disconnections between cortical areas.