Abnormal Functional Connectivity In Subjects With Mild Cognitive Impairment And Alzheimer’s Disease

Human brain is the most important organ of body. We rely on it for thinking, perception, memory and completing daily cognition. Meanwhile, human brain is a complex and mysterious system to us. To date, we know little about the brain. How the brain works and real new insights about what causes illnesses and problems are rare. Recently, investigating human brain through brain imaging has become a hotspot of research in the world. Functional magnetic resonance imaging (fMRI) detects altered magnetic field caused by blood oxygen level dependent (BOLD) signal and provides an indirect in vivo correlate of neural activity. Functional connectivity captures patterns of deviations from statistical independence between BOLD signal of topographically-separated regions. It reflects coherence of spontaneous activity between brain regions. fMRI has been applied to investigating nervous system diseases more and more, such as Alzheimer’s disease (AD), depression and so on. Investigating human brain through fMRI strengths our understanding about potential mechanism of disease and facilitates early diagnosis and intervention and finally help to delay disease progression.AD is a neurodegeneration disease with symptoms typically including deficits in episodic memory, resoning, language, spatial attention as well as motor and behavior function decline. Extracellular P-amyloid deposition and hyperphosphorylation of tau and subsequent formation of intracellular neurofibrillary tangles are pivotal features of pathophysiology of AD patients. The cellular pathology results in loss of neuron and synaptics and induces deficits in cognition. Mild cognitive impairment (MCI) is a transition stage between healthy elder and AD with memory decline and daily function intact. MCI is belived to be a high risk condtion for converting to AD and 50% MCI will convert to AD in 3-5 years. While healthy elder tend to convert to AD at a rate of approximately 1%-2% per year. AD has a long period of latency, and pathophysiology changes decades before the appearance of clinical symptoms. At prodromal stage (MCI), patients showed deficits in memory, hippocampus atrophy and regional hypometabolism. At later stage of AD, structural changes including hippocampus, entorphinal cortex atrophy and atrophy in temporal, frontal and neocortical. AD is an irreversible process and few available pharmacological treatments with limit efficacy. Given that pathphysiology changes at early stage and structural lesion at later stage of AD, fMRI detects functional dysfunctions that help to early diagnosis and intervention of AD.In this dissertation, we focused on the following three topics:impaired intra-and inter-network brain connectivity in subjects at high risk for Alzheimer’s disease; impaired episodic memory network in subjects at high risk for Alzheimer’s disease; the last one is cluster-based statistics for aberrant functional connectivity in Alzheimer’s disease. The main contents and contributions of the dissertation are as follows:☆ Impaired episodic memory network in subjects at high risk for Alzheimer’s diseaseEpisodic memory is a cognitive system that enables an individual to record (or encode), store, and retrieve information about personal experiences and the temporal and spatial contexts of those experiences. Deficit in episodic memory is one of the hallmark symptoms in AD and MCI. The core regions related to episodic memory including hippocampus, parahippocampus, entorphinal cortex, perirhinal cortex and so on. Investigating connectivity between regions relate to episodic memory shed light on the impaired episodic memory function in AD and MCI.Data were downloaded from Alzheimer’s disease Neuroimaging Initiative (ADNI) dataset. According to our screen criterias,68 participants with mean scan interval about half year were included,23 normal controls (NC),26 early MCI (EMCI) and 19 late MCI (LMCI). The episodic memory network was extracted based on an automated brain mapping framework, NeuroSynth. It extracted regions related to episodic memory based on meta-analysis of 393 studies. Voxel-wise connectivity was obtained for each subjects and connect sparsity from 0.05 to 0.20 with increment of 0.05 were applied to retaining the strongest connections and binarization. Voxel-wise connectivity degree was calculated by summing the number of connections that directly link to it in the resultant matrices. Subsequent, repeated measure ANOVA was performed for a group-wise comparison between three groups. The identified abnormal regions were extracted as seed of interest and estimate the voxel-wise connectivity between seed region and rest voxels within episodic memory. A repeated measure ANOVA was performed for a group-wise difference in functional connectivity between three groups.Right fusiform showed group difference in connectivity degree between three groups. Comparing to NC subjects, MCI patients showed affected connectivity between right fusiform and right precuneus due to between-subject factor. Futhermore, MCI patients exhibited abnormal connectivity between right fusiform and extend seed region (right fusiform) comparing to NC.The observed alteration in functional connectivity in MCI subjects might strengthen the idea that mild changes in episodic memory would lead to functional connectivity differences in visual processing tasks in the MCI subjects. These findings provide a new insight in understanding the impaired episodic memory function in MCI and AD.☆ Impaired intra-and inter-network brain connectivity in subjects at high risk for Alzheimer’s diseasePrevious fMRI studies indicated that default mode network (DMN) is preferentially affected by AD/MCI. With disease progression, AD patients showed abnormal connectivity widely distributed in whole brain. Recently, reseachers investigate AD dysfunction from resting-state network (RSN) aspect and found aberrant RSN beyond DMN, such as salience network (SAL) and dorsal attention network (DAN). Moreover, results indicated that not only intra-network connectivity dysfunction but also inter-network. But little is known for MCI patients. Therefore, studying the organization of intra-and inter-network connectivity in the context of cognitive changes observed in individuals at high risk for AD may shed light on the neurological basis of cognitive decline in AD.We extracted ROI template from Brier and colleagues achievement which reprensting DMN, DAN, control network (CON), SAL and sensory-motor network (SMN). ROI based functional connectivity was constructed for each aforementioned 68 subjects. We investigated disease-related alterations in functional connectivity with three different levels of analysis:’nodal integration’;’RSN profile’; and ’large-scale connectivity’. We also tested for continuous relationships between brain network measurements and cognitive ability (Alzheimer’s disease Assessment Scale-Cognitive section, ADAS-Cog).The "nodal integration" results showed that functional connectivity of right insula and right putamen was significant altered in MCI patients in comparing to NC. We observed that connectivity of left visual cortex and medial thalamu was significantly correlated with ADAS-Cog. The "RSN profile" results showed that change in intra-network connectivity within DMN was correlated with change in ADAS-Cog. The change in correlation strength of the DMN-SAL pair showed group difference between three groups. The "large-scale" connectivity results showed that the aberrant connecitivity mostly involved in DMN, SMN and SAL.Our findings showed that with disease progression, abnormal connectivity widely distributed in whole brain and the aberrant functional connetivities of the selected RSNs mainly involved in the DMN and SAL networks. The abnormal functional connectivities correlated with the changes ADAS-Cog scores. These results suggest that abnormalities in connectivity may facilitate the early diagnosis of AD.☆ Cluster-based statistics for aberrant functional connectivity in Alzheimer’s diseaseHuman brain is a complex and efficient system in natrue. Not only can it processing specific signal, but also integrating signals from different brain region. Given the nature that interconnected of sub-networks are responsible for daily behavior rather than single pair of functional connectivity. It is valid to exploit the clustering structure of connectivity alterations in AD/MCI.We explored abnormal connectivity based on resting-state fMRI connectivity in a sample of patients with AD (N=35), MCI (N=27) and age-matched healthy subjects (N=27). A ROI template which is comprised of 264 ROIs representing 14 RSNs was extracted from Power and colleague. First, the abnormal network components were identified using network-based statistic (NBS) approach between AD and NC. In order to test the relationship between connectivity strength and disease severity, connectivity strength of the identified network components were compared between MCI and NC, MCI and AD. Finally, Pearson’s correlation analyses were performed between the mean strengths of the identified connectivity components and the clinical variables (MMSE) in the MCI and AD patients.The results demonstrated that patients had reduced functional connectivity strength in several components including the default mode network, sensori-motor network, visual-sensory network and visual-attention network. In patients with AD, functional connectivity of these subnetworks exhibited greater attenuation than that in the MCI subjects compared to NC. Greater degree of cognitive impairment was correlated with greater decreased functional connectivity in the identified subnetworks.These results indicate that neurodegenerative disruption of fMRI connectivity widely distributed in several networks in AD/MCI. Our findings suggest that we should pay more attention on sensory and motor nervous system dsyfuntion in elder. These profiles deepen our understanding of the neural basis of AD/MCI dysfunction and indicate the potential of resting-state fMRI measures as biomarkers or predictors of disease progression in AD.