Classification And Prediction Of Alzheimer’s Disease Spectrum And Early Repetitive Transcranial Magnetic Stimulation Based On Neuroimaging Technology

Chapter one:Moderating role of the default mode network in the relationship between the APOE genotype and cognition and individualizes identification across the Alzheimer’s disease spectrumObjective:Apolipoprotein E(APOE)is closely related to cognitive function,and the changes in the default mode network(DMN)can be used to monitor the progression of Alzheimer’s disease(AD).The present study aimed to investigate how the APOE genotype regulates the DMN and subsequently affects cognitive decline in the AD spectrum.Method:This study included 46 cognitive normal(CN),52 early mild cognitive impairment(EMCI),41 late mild cognitive impairment(LMCI)and 30 AD subjects with resting-state functional magnetic resonance imaging(rs-fMRI)data,APOE genotype test and assessment of cognitive function.Posterior cingulate cortex(PCC)was selected as seed to construct DMN across all subjects.We investigated the main effects and interaction of the APOE genotype and disease status on the DMN.Multiple linear regression analysis was also employed to investigate the behavioral significance of the altered DMN connectivity.A moderation analysis was conducted to investigate the relationship among the APOE genotype,DMN connectivity and cognition.Additionally,the pair-wised DMN connectivity was used to classify AD spectrum.Results:Compared to APOE ε4 non-carriers,APOEε4 carriers showed the opposite trajectory of DMN connectivity across the AD spectrum.The strengths in the posterior cingulate cortex(PCC)connecting with the right precuneus,right insular,and right fusiform area(RFFA)were positively correlated with general cognition scores in APOE ε4 non-carriers but not in APOE ε4 carriers.Furthermore,PCC-RFFA connectivity could moderate the negative effect of the APOE genotype on general cognition scores across the disease groups.These altered connectivities produced strong classification powers in the AD spectrum.Conclusions:Intrinsic DMN connectivity moderated the effect of the APOE genotype on cognition and provided a neuroimaging biomarker for early differentiation of the AD spectrum.Chapter two:The effect of endocytosis-pathway multilocus genetic profile on hippocampal network connectivity and individualized identification across the Alzheimer’s disease spectrumObjective:Hippocampus are important for cognitive processing.Endocytosis dysfunction may play an important role in the pathogenesis of AD.The present study aimed to investigate the neural mechanisms underlying the polygenic effects in the endocytosis pathway on the hippocampal network across the AD spectrum using imaging genetic approach.Method:This study included 35 CN,23 EMCI and 19 LMCI subjects.All the participants had endocytosis-pathway multilocus gene test,structural MRI,rs-fMRI data and assessment of cognitive function.Hippocampal volumes were examined among the three group subjects.Multivariate linear regression analysis was employed to measure the effects of disease and endocytosis-based multilocus genetic risk scores(MGRS)on the hippocampal network which was constructed using the bilateral hippocampal regions.Additionally,the model composed of pair-wised hippocampal network connectivity and hippocampal volume was used to classify EMCI and LMCI subjects.Results:Hippocampal volumes in LMCI group were smaller than those in CN and EMCI groups.Endocytosis-based MGRS widely influenced the neural structures within the hippocampal network,mainly in the prefrontal-occipital regions and striatum.Compared to low endocytosis-based MGRS carriers,high MGRS carriers showed the opposite trajectory of hippocampal network FC across the prodromal stages of AD.Furthermore,a model composed of selected hippocampal FCs and hippocampal volume produced strong classification powers of EMCI and LMCI.Conclusion:Endocytosis-pathway multilocus genetic profile made a considerable contribution to hippocampal network across the AD spectrum.Altered connectivities and hippocampal volume produced strong classification powers across the AD spectrum.Chapter three:Connectome-based models predict delayed recall performance across the Alzheimer’s disease spectrumObjective:Early clinical symptoms of AD is episodic memory impairment.Using delayed recall to quantify episodic memory impairment is very useful for detection prodromal stages of AD.The present study aimed to explore whether the whole brain resting-state functional connectivity(rs-FC)could predict delayed recall performance across the AD spectrum.Method:This study included 33 CN,26 subjective cognitive decline(SCD)and 27 amnestic mild cognitive impairment(aMCI)subjects.All of the participants had rs-fMRI and assessment of cognitive function.Connectome-based predictive modeling(CPM)based on the rs-FC data was used to predict the auditory verbal learning test-delayed recall(AVLT-DR)scores,which measured episodic memory impaiment in individuals.Significant correlation FCs were obtained by constructing correlation of all brain connections and AVLT-DR scores across the SCD and aMCI participants.Two linear regression prediction models were trained according to positive/negative correlation.Brain connections with predictive characteristics were selected and fed back to the trained positive/negative correlation models respectively,for predicting the target AVLT-DR score.Results:CPM could predict individual delayed recall performance from rs-FC across the AD spectrum.Key nodes that contributed to the prediction model were mainly located in the prefrontal cortex,frontal cortex,parietal cortex and temporal lobe.Conclusion:Our findings demonstrated that rs-FC among multiple neural systems could predict delayed recall performance across the AD spectrum.Chapter four:Abnormal dynamic functional network connectivity and individualized identification across Alzheimer’s disease spectrumObjective:Most previous studies on AD spectrum have focused on the characteristics of static brain activity,ignoring the brain activity dynamics.The study aimed to investigate dynamic FC(dFC)and variability of topological metrics of functional network across the AD spectrum.Method:This study included 33 CN,30 SCD and 30 aMCI subjects.All of the participants had rs-fMRI and assessment of cognitive function.The sliding window approach and k-means algorithm were used to identify connectivity states.Temporal properties and FC differences between groups in each state were analyzed.Graph-theoretical analysis for the variability of network topological organization was performed.Additionally,group difference of dynamic topological metrics was used to classify SCD and aMCI subjects.Results:Two connectivity states(strongly-connected state and weakly-connected state)were identified by k-means clustering in AD spectrum.Compared with CN group,aMCI group showed increased mean dwell time in the weakly-connected state and stronger FC within frontal-parietal-occipital-striatal subnetwork in the strongly-connected state.Compared with the SCD group,the aMCI group showed reduced time variability of degree centrality(DC)in the right hippocampus and could be used to classify the SCD and aMCI subjects.Conclusion:These findings provided new perspectives for understanding the abnormal dFC across the AD spectrum.And variability of DC in the right hippocampus could be used as a neuroimaging biomarker for SCD and aMCI classification.Chapter five:The effect of long-term repetitive transcranial magnetic stimulation on the cognitive function in amnestic mild cognitive impairmentObjective:To assess the effect of long-term high frequency repetitive transcranial magnetic stimulation(rTMS)on cognitive function in aMCI and explore the brain-neural-mechanisms of the rTMS for aMCI.Method:18 aMCI subjects received sixty sessions of rTMS(90%motor-threshold;left dorsolateral prefrontal cortex at 10 Hz;five sessions each week).Cognitive function assessment and MRI were performed at baseline and the last rTMS session,which was compared with 21 CN subjects.Mean fractional amplitude of low-frequency fluctuation(mfALFF)and FC were measured to reflect resting state brain functional activity.Additionally,the model composed of altered mfALFF and connectivity was used to classify aMCI from CN at baseline.Results:AVLT-DR scores improved for aMCI subjects after rTMS treatment.Increased mfALFF was found in right posterior middle temporal gyms/inferior temporal gyrus(RpMTG/ITG)and left posterior cerebellum(LpCBLM)after rTMS treatment.RpMTG/ITG network and LpCBLM network connection strength also increased in aMCI subjects after rTMS treatment.In addition,the altered strengths of LpCBLM connecting with the bilateral superior frontal gyrus,right hippocampal/parahippocampal gyrus and RMTG/ITG were positively correlated with the changes of AVLT-DR scores in aMCI subjects.Furthermore,the model composed of selected mfALFF and FC of RpMTG/ITG could classify aMCI from CN at baseline.Conclusions:The long-term high frequency rTMS might be an effective treatment for aMCI subjects.It could influence the local brain activity and remodel the brain network to improve memory function.Altered mfALFF and FC of RpMTG/ITG produced strong classification powers of aMCI and CN.