There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Hao Luo |
| Institution: | Sun Yat-sen University |
| Department: | School of Public Health |
| Country: | |
| Proposed Analysis: | Purpose: Briefly speaking, our purpose is establish a prediction model to predict the progression from Mild Cognitive Impairment(MCI) to Alzheimer's Disease(AD) by analyzing longitudinal changes in rest-state functional connectivity from fMRI. Background: Many previous papers suggested the potential for functional connectivity to distinguish patients who progress from MCI to AD from who do not.While most of those studies used only once scan, namely cross-sectional data, to achieve that. Considering that the process of progression from MDI to AD is long, in which the structure of brains in patients could experience significant changes. Accordingly, we assume that the information in longitudinal changes in the functional connectivity patterns could enhance prediction performance by absorbing temporal information. Based on the above, we attempted to apply for the access to ADNI database, which possesses relatively large size of cohorts to analyse. Expected methods: To be brief, the problem in this study can be described as the classification problem of high-dimension time series with missing data and unbalanced intervals of measures.To solve that, firstly, mature or emerged feature extraction methods will be used to extract features which have the most useful information for prediction.Secondly, machine learning , or further deep learning techniques, will be adopted by us to serve as the prediction model. Meaning: 1) By analyzing the longitudinal functional connectivity data, we expect to obtain a considerable performance for progression prediction, which would outperform the model using only the baseline data. 2) We attempt to find out the the most proper time window to predict(in other words, before when the performance will maintain relatively stable) and figure out the least number of scans needed to obtain a relatively high performance(the less number, the less burden patients will have. 3) If our outcomes is as expected, we hope to transplant our outcomes in quantitative Electroencephalogram(EEG), where the rest-state functional connectivity analysis are also popular. If our outcomes in fMRI can be replicated by using EEG, our research will advance the progression prediction in AD patients owing to relative cheapness and convenient accessibility for EEG in the most developing country. |
| Additional Investigators | |
| Investigator's Name: | Jinxin Zhang |
| Proposed Analysis: | Purpose: Briefly speaking, our purpose is establish a prediction model to predict the progression from Mild Cognitive Impairment(MCI) to Alzheimer's Disease(AD) by analyzing longitudinal changes in rest-state functional connectivity from fMRI. Background: Many previous papers suggested the potential for functional connectivity to distinguish patients who progress from MCI to AD from who do not.While most of those studies used only once scan, namely cross-sectional data, to achieve that. Considering that the process of progression from MDI to AD is long, in which the structure of brains in patients could experience significant changes. Accordingly, we assume that the information in longitudinal changes in the functional connectivity patterns could enhance prediction performance by absorbing temporal information. Based on the above, we attempted to apply for the access to ADNI database, which possesses relatively large size of cohorts to analyse. Expected methods: To be brief, the problem in this study can be described as the classification problem of high-dimension time series with missing data and unbalanced intervals of measures.To solve that, firstly, mature or emerged feature extraction methods will be used to extract features which have the most useful information for prediction.Secondly, machine learning , or further deep learning techniques, will be adopted by us to serve as the prediction model. Meaning: 1) By analyzing the longitudinal functional connectivity data, we expect to obtain a considerable performance for progression prediction, which would outperform the model using only the baseline data. 2) We attempt to find out the the most proper time window to predict(in other words, before when the performance will maintain relatively stable) and figure out the least number of scans needed to obtain a relatively high performance(the less number, the less burden patients will have. 3) If our outcomes is as expected, we hope to transplant our outcomes in quantitative Electroencephalogram(EEG), where the rest-state functional connectivity analysis are also popular. If our outcomes in fMRI can be replicated by using EEG, our research will advance the progression prediction in AD patients owing to relative cheapness and convenient accessibility for EEG in the most developing country. |
