Poster No:
1633
Submission Type:
Abstract Submission
Authors:
Taegyeong Lee1, Kang-min Choi1,2, Seung-Hwan Lee2,3, Chang-Hwan Im1,4
Institutions:
1School of Electronic Engineering, Hanyang University, Seoul, Republic of Korea, 2Clinical Emotion and Cognition Research Laboratory, Inje University, Goyang, Korea, Republic of, 3Department of Psychiatry, Ilsan Paik Hospital, Inje University College of Medicine, Goyang, Korea, Republic of, 4Department of Biomedical Engineering, Hanyang University, Seoul, Korea, Republic of
First Author:
Taegyeong Lee
School of Electronic Engineering, Hanyang University
Seoul, Republic of Korea
Co-Author(s):
Kang-min Choi
School of Electronic Engineering, Hanyang University|Clinical Emotion and Cognition Research Laboratory, Inje University
Seoul, Republic of Korea|Goyang, Korea, Republic of
Seung-Hwan Lee
Clinical Emotion and Cognition Research Laboratory, Inje University|Department of Psychiatry, Ilsan Paik Hospital, Inje University College of Medicine
Goyang, Korea, Republic of|Goyang, Korea, Republic of
Chang-Hwan Im
School of Electronic Engineering, Hanyang University|Department of Biomedical Engineering, Hanyang University
Seoul, Republic of Korea|Seoul, Korea, Republic of
Introduction:
Several patients with major depressive disorder (MDD) do not respond to the antidepressant. Prediction of antidepressant response is regarded as one of the necessary issues because pharmacotherapy for treatment-resistant patients could cause unnecessary side effects. For several years, numerous neuroimaging studies have demonstrated the feasibility that characteristics of functional brain networks could serve as predictors for pharmacotherapeutic responses in patients with MDD. Most of these studies have assumed that the brain would maintain a functionally static state; however, as well known, functional brain networks indeed dynamically fluctuate even in the resting state [1]. Electroencephalography (EEG) has comparative advantages for exploring dynamic functional brain networks over the other modalities, due to its excellent temporal resolution. In this study, we compared the dynamic patterns of the resting-state EEG-derived functional brain networks between patients with non-remitted MDD (nrMDD) and remitted MDD (rMDD).
Methods:
Forty-nine drug-naïve patients with MDD (aged 45.47 ± 12.00, W 46) and twenty-two healthy controls (HC, aged 44.05 ± 13.80, W 18) participated in this study. After receiving vortioxetine treatment for 8 weeks, patients were divided into two subgroups based on Hamilton Depression Rating Scales: (i) nrMDD (Ham-Dw8 ≥ 8) and rMDD (otherwise). EEG signals were acquired during the resting state while participants closed their eyes for 3 min. After signal pre-processing, a 30-s noise-free segment was manually selected individually. To construct the dynamic functional brain network, the 30-s segments were divided into 2-s time bins with 95 % overlap. Thirty-one regions of interest were determined as nodes. For edges, functional connectivity was dynamically evaluated for each time bin using the weighted phase-lag index (wPLI), for 6 frequency bands, including theta, low alpha, high alpha, low beta, high beta, and gamma bands. The k-means clustering was implemented to identify the states of the reoccurring functional network (FN). For clustering each network derived from the time bin, the L1 norm-based distance was employed for the clustering analysis. The number of states, k, was determined by the elbow method. It is noted that the clustering was conducted using all networks for the participants, including nrMDD, rMDD, and HC. The fraction time (FT) and dwell time (DT) of each state were evaluated to assess the dynamic characteristics of the FN states. To compare the dynamic FN indices between the nrMDD and rMDD, the permutation test was conducted, with the significance level adjusted by the number of frequency bands (i.e., α = 0.05/6).
Results:
The optimal k values were determined to be 4, except for the high-alpha band (k = 5). In the high alpha band, nrMDD exhibited decreased FT (nrMDD = 0.18 vs. rMDD = 0.24, p = 0.0060) and DT (nrMDD = 2.93 vs. rMDD = 3.48, p = 0.0034) in the state 3, which is characterized with overall hypoconnectivities particularly for the right hemispheric temporo-parieto-occipital regions: right visual primary cortex, right intraparietal sulcus, right angular gyrus, right inferior parietal lobule, right middle temporal gyrus, and right supramarginal gyrus.
Conclusions:
In this study, we suggested that aberrant dynamicity in the resting-state functional brain networks could predict the treatment efficacy of antidepressants in patients with MDD. Non-remitted patients tended to spend less time and escape earlier for the functional state in which the right hemispheric temporo-parieto-occipital regions are suppressed in the high-alpha band.
Disorders of the Nervous System:
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 2
Modeling and Analysis Methods:
EEG/MEG Modeling and Analysis 1
Keywords:
Electroencephaolography (EEG)
Psychiatric Disorders
Other - resting-state functional brain network
1|2Indicates the priority used for review
Provide references using author date format
Fox, Michael D. (2005), 'The human brain is intrinsically organized into dynamic, anticorrelated functional networks', Proceedings of the National Academy of Sciences, vol. 102, no. 27 pp. 9673-9678.