Dynamic Brain Network States in Alzheimer’s Disease
Poster No:
298
Submission Type:
Abstract Submission
Authors:
Tien-Tse Huang1, Shih-Jen Tsai2, Ching-Po Lin3, Chiung-Chih Chang4, Chun-Yi Zac Lo1
Institutions:
1Biomedical Engineering Department,Chung Yuan Christian University, Taoyuan, Taiwan, 2Department of Psychiatry, Taipei Veterans General Hospital, Taipei,Taiwan, Taipei, Taiwan, 3Institute of Neuroscience, National Yang Ming Chiao Tung University, Taipei, Taiwan, 4Department of Neurology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Kaohsiung, Taiwan, Kaohsiung, Taiwan
First Author:
Co-Author(s):
Shih-Jen Tsai
Department of Psychiatry, Taipei Veterans General Hospital, Taipei,Taiwan
Taipei, Taiwan
Department of Psychiatry, Taipei Veterans General Hospital, Taipei,Taiwan
Taipei, Taiwan
Chiung-Chih Chang
Department of Neurology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Kaohsiung, Taiwan
Kaohsiung, Taiwan
Department of Neurology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Kaohsiung, Taiwan
Kaohsiung, Taiwan
Introduction:
Alzheimer's disease is a neurodegenerative disorder, with disrupted system-level brain network organization [1]. With the development of medical imaging equipment and analysis techniques, functional Magnetic Resonance Imaging (fMRI) has increasingly become a vital tool for observing brain activity. It enables the construction of dynamic functional connectivity networks in the brain [2]. Three states of the brain's dynamic functional networks have been demonstrated which are associated with sensory , somatomotor, and internal mentation state, and highly correlated with aging [3]. In this study, we investigated the dynamic brain network states in Alzheimer's disease, and it may be as potential biomarkers in biomedical imaging.
Methods:
MR data were obtained from Kaohsiung Chang Gung Memorial Hospital, including T1-weighted images (T1WI) and fMRI, with two sets of imaging acquisition protocols. Total 43 healthy volunteers (HV ) and 22 AD patient MR data were acquired using a Siemens 3T skyra scanner, including 34 participants data (HV: 18 AD:16 ) with protocol as T1WI: voxel size: 1.0*1.0*1.0 mm³ fMRI: TR=3660ms, TE=45ms, slice number = 50, voxel size = 3.0*3.0*3.0 mm³, 120 volumes; and 31 participants data (HV: 25, AD: 6) T1WI: voxel size: 0.5*0.5*1.0 mm³; fMRI: TR=2500ms, TE=27ms, slice number=43, voxel size=3.4*3.4*3.4 mm³, 200 volumes.
After preprocessing the T1-weighted and fMRI images separately [4], the fMRI data was first registered to its corresponding T1 space. Subsequently, utilizing previously computed transformation matrices, the T1 images were registered to the standard brain space (MNI152). Finally, a 6mm full-width at half-maximum Gaussian blur was applied to smooth the images.
The fMRI signals were segmented into brain regions using the Automated Anatomical Labeling 2 (AAL2) atlas [5]. A sliding-window approach was used for dynamic network estimation. Employing a 20 frame time window, correlation were performed at different time window for each brain region. With that data we can derive the similarities between feature matrix of the three brain networks. Three predominant states of brain networks were estimated by feature score with the feature metrics into: sensory, somatomotor, and internal mentation [3] (Fig.1). The occurrence rates of each person in these three modes were calculated. Due to different protocols, ComBat was applied in statistical analyses [6]. Analysis of covariance (ANCOVA) was used to test the difference between groups with age, gender and education years as covariates. The presence of significant differences P < 0.05 was set.
After preprocessing the T1-weighted and fMRI images separately [4], the fMRI data was first registered to its corresponding T1 space. Subsequently, utilizing previously computed transformation matrices, the T1 images were registered to the standard brain space (MNI152). Finally, a 6mm full-width at half-maximum Gaussian blur was applied to smooth the images.
The fMRI signals were segmented into brain regions using the Automated Anatomical Labeling 2 (AAL2) atlas [5]. A sliding-window approach was used for dynamic network estimation. Employing a 20 frame time window, correlation were performed at different time window for each brain region. With that data we can derive the similarities between feature matrix of the three brain networks. Three predominant states of brain networks were estimated by feature score with the feature metrics into: sensory, somatomotor, and internal mentation [3] (Fig.1). The occurrence rates of each person in these three modes were calculated. Due to different protocols, ComBat was applied in statistical analyses [6]. Analysis of covariance (ANCOVA) was used to test the difference between groups with age, gender and education years as covariates. The presence of significant differences P < 0.05 was set.
Results:
As illustrated in Figure 2, there are significant differences in the occurrence rates of the Sensory and Somatomotor states between AD and HV group (p=0.018; p=0.014). There is no significant difference in terms of internal mentation state between groups. The occurrence rates of these two states are primarily influenced by the presence or absence of Alzheimer's disease. The findings of this study reveal that the dynamic states of the brain network in Alzheimer's disease.
·Fig2:The boxplots of occurrence rate. The significant group difference was found in the sensory and somatomotor state between AD and HV (denoted by *, P < 0.05).
Conclusions:
Alzheimer's patients showed lower occurrence rate in Sensory and Somatomotor states. Dynamic brain networks might be capable of quantifying the brain activity, serving as a reference indicator in the interpretation of Alzheimer's disease. The quantitative metrics of brain activity states assessed in this study may be as potential biomarkers in biomedical imaging.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) 1
Modeling and Analysis Methods:
fMRI Connectivity and Network Modeling 2
Keywords:
FUNCTIONAL MRI
Other - Alzheimer's Disease, Dynamic Functional Connectivity
1|2Indicates the priority used for review
Provide references using author date format
2. Zhou, Q.J., et al., Tracking the Main States of Dynamic Functional Connectivity in Resting State. Frontiers in Neuroscience, 2019. 13.
3. Zhang, L., et al., Sensory, somatomotor and internal mentation networks emerge dynamically in the resting brain with internal mentation predominating in older age. Neuroimage, 2021. 237: p. 118188.
4. Alorf, A. and M.U.G. Khan, Multi-label classification of Alzheimer's disease stages from resting-state fMRI-based correlation connectivity data and deep learning. Comput Biol Med, 2022. 151(Pt A): p. 106240.
5. Rolls, E.T., et al., Effective Connectivity in Depression. Biological Psychiatry-Cognitive Neuroscience and Neuroimaging, 2018. 3(2): p. 187-197.
6. Johnson, W.E., C. Li, and A. Rabinovic, Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics, 2007. 8(1): p. 118-127.