Investigation of common EEG features between PD and MDD
Poster No:
196
Submission Type:
Abstract Submission
Authors:
Chia-Yen Yang1, Fan-Ning Kuo1
Institutions:
1Department of Biomedical Engineering, Ming-Chuan University, Taoyuan, Taiwan
First Author:
Co-Author:
Introduction:
Parkinson's disease (PD) is the second most common neurodegenerative disease [1]. While its exact cause remains elusive, it's widely thought to intertwine with genetic, environmental, and neurological factors [2]. Complex clinical profiles of PD patients include motor and non-motor symptoms. Numerous studies have highlighted certain non-motor symptoms (such as depression) that manifest years prior to disease onset [e.g., 3], potentially aiding early diagnosis.
Depression, a prevalent mental illness with diverse symptoms, stems from intricate causes like psychology, life stress, genetics, personality traits, physical ailments, and brain abnormalities [4]. Although depression and PD are distinct, many studies have identified some shared physiological mechanisms [5-7], such as mitochondrial dysfunction, the monoamine hypothesis, and the inflammation hypothesis.
The identified abnormalities might alter brain activity. Utilizing brainwave analysis, shared traits between the two diseases could be further investigated for assessment applications. Therefore, the aim of this study was to identify distinctive differences in resting-state electroencephalography (EEG) between PD patients and healthy controls (HCs), as well as between patients with major depressive disorder (MDD) and HCs, while also to explore common features between PD and MDD.
Depression, a prevalent mental illness with diverse symptoms, stems from intricate causes like psychology, life stress, genetics, personality traits, physical ailments, and brain abnormalities [4]. Although depression and PD are distinct, many studies have identified some shared physiological mechanisms [5-7], such as mitochondrial dysfunction, the monoamine hypothesis, and the inflammation hypothesis.
The identified abnormalities might alter brain activity. Utilizing brainwave analysis, shared traits between the two diseases could be further investigated for assessment applications. Therefore, the aim of this study was to identify distinctive differences in resting-state electroencephalography (EEG) between PD patients and healthy controls (HCs), as well as between patients with major depressive disorder (MDD) and HCs, while also to explore common features between PD and MDD.
Methods:
2.1 Datasets
All EEG data used in this study were downloaded from the Patient Repository of EEG Data + Computational Tools (PRED+CT) developed by Cavanagh et al. [8, 9]. 27 PD patients, 21 MDD patients and 27 HCs were included. Participants were instructed to minimize movements and to remain thoughtless for 3 or 5 min in the eye closed state. EEG signals were recorded with a sampling rate of 500 Hz.
2.2 EEG Processing
There were two steps for pre-processing of EEG signals: detrending and 0.5–50 Hz bandpass filtering. After preprocessing, the signals were decomposed into five frequency bands through discrete wavelet decomposition: delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), and gamma (30–50 Hz). Five features were then calculated: mean frequency, frequency power, alpha interhemispheric asymmetry, sample entropy and detrended fluctuation analysis (DFA).
2.3 Statistical analysis
The Mann-Whitney U test was used to compare the EEG features between the patients with PD and HCs and between the patients with MDD and HCs (p < 0.05 for significance). Furthermore, Pearson correlation coefficient was used to evaluate the similarity of features between PD and MDD patients (R > 0.8 and p < 0.05 for significance).
All EEG data used in this study were downloaded from the Patient Repository of EEG Data + Computational Tools (PRED+CT) developed by Cavanagh et al. [8, 9]. 27 PD patients, 21 MDD patients and 27 HCs were included. Participants were instructed to minimize movements and to remain thoughtless for 3 or 5 min in the eye closed state. EEG signals were recorded with a sampling rate of 500 Hz.
2.2 EEG Processing
There were two steps for pre-processing of EEG signals: detrending and 0.5–50 Hz bandpass filtering. After preprocessing, the signals were decomposed into five frequency bands through discrete wavelet decomposition: delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), and gamma (30–50 Hz). Five features were then calculated: mean frequency, frequency power, alpha interhemispheric asymmetry, sample entropy and detrended fluctuation analysis (DFA).
2.3 Statistical analysis
The Mann-Whitney U test was used to compare the EEG features between the patients with PD and HCs and between the patients with MDD and HCs (p < 0.05 for significance). Furthermore, Pearson correlation coefficient was used to evaluate the similarity of features between PD and MDD patients (R > 0.8 and p < 0.05 for significance).
Results:
We first identified distinct EEG features between PD patients and HCs, and between MDD patients and HCs. Then, we intersected these features, considering those showing consistent trends across groups as candidate of common EEG features between the two diseases. Figure 1 displays the types and quantities of significant features identified.
Additionally, we conducted correlations of EEG features between PD and MDD. Figure 2 illustrates three significantly correlated features, i.e., relative power, sample entropy and DFA.
The combined results of comparison and correlation analyses suggested potential common EEG features encompassing delta, alpha, beta, and gamma bands in relative power, delta, theta and alpha bands in sample entropy, and delta in DFA. Specifically, delta, alpha, and beta power, along with delta entropy, emerge as pronounced common EEG traits. These findings might aid in understanding the neurophysiological connections between PD and MDD.
Additionally, we conducted correlations of EEG features between PD and MDD. Figure 2 illustrates three significantly correlated features, i.e., relative power, sample entropy and DFA.
The combined results of comparison and correlation analyses suggested potential common EEG features encompassing delta, alpha, beta, and gamma bands in relative power, delta, theta and alpha bands in sample entropy, and delta in DFA. Specifically, delta, alpha, and beta power, along with delta entropy, emerge as pronounced common EEG traits. These findings might aid in understanding the neurophysiological connections between PD and MDD.
·Figure 1. Significantly different EEG features between patients and HCs and the intersection between them with the same trend (marked with a gray background).
·Figure 2. Topographic maps of averaged relative power (a), sample entropy (b) and DFA (c) of EEG signals from PD and MDD patients and their significant correlation coefficients (R).
Conclusions:
This study demonstrated that there were certain similarities in EEG features between PD and MDD diseases, i.e., traits in delta, alpha, and beta power, alongside delta entropy. The results may facilitate in future applications for transfer learning models between the two diseases, or even in model training for other rare diseases.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 1
Psychiatric (eg. Depression, Anxiety, Schizophrenia) 2
Keywords:
Data analysis
Electroencephaolography (EEG)
Other - Parkinson's disease; Depression; Features
1|2Indicates the priority used for review
Provide references using author date format
[2] Lu, J., Wu, M., & Yue, Z. (2020), Autophagy and Parkinson’s Disease. Advances in Experimental Medicine and Biology, vol 1207, pp. 21:51.
[3] Cheng, Y.C and Su, C.H. (2020), Evidence Supports PA Prescription for Parkinson’s Disease: Motor Symptoms and Non-Motor Features: A Scoping Review. International Journal of Environmental Research and Public Health (IJERPH), vol. 17, no. 8, pp. 2894.
[4] Pitsilloun, E., Bresnehan, S.M., Kagarakis, E.A., & et al. (2019), The cellular and molecular basis of major depressive disorder: towards a unified model for understanding clinical depression. Molecular Biology Reports, vol. 47, pp. 753–770.
[5] Yang, L., Youngblood, H., Wu, C., et al. (2020), Mitochondria as a target for neuroprotection: role of methylene blue and photobiomodulation. Translational Neurodegeneration, vol. 9, pp. 19.
[6] Dallé E. and Mabandla M.V. (2018), Early Life Stress, Depression And Parkinson’s Disease: A New Approach. Mol Brain, vol. 11, no.1, pp. 18.
[7] Lessie E.N. (2021), Neurobiological similarities between depression and parkinson’s disease: etiological and therapeutic implications (bachelor's thesis, university of arizona, tucson, usa).
[8] Cavanag, J.F.h., Bismark, A.W., Frank, M.J., Allen, J.J.B. (2019), Multiple dissociations between comorbid Depression and anxiety on reward and punishment processing: evidence from computationally informed EEG. Computational Psychiatry, vol. 3, no.1, pp. 1–17.
[9] Cavanag, J.F.h., Kumar, P., Mueller, A.A., Richardson, S.P., & et al. (2018), Diminished EEG habituation to novel events effectively classifies Parkinson’s patients. Clinical Neurophysiology, vol. 129, no. 2, pp. 409–418.