Investigator:Tammie Benzinger, M.D. Ph.D.
Title:Development of a multi-modality deep convolutional neural network feature extraction algorithm for prediction of longitudinal biomarker evolution in Alzheimer’s disease.
Date of Request:6/28/2017
ID:DIAN-D1719
Aim 1:Develop and implement an in-lab processing pipeline for using deep neural networks for analysis of cross-sectional and longitudinal MRI, PET, clinical, and psychometric data, with both single and multi-modality implementations.
Aim 2:Replicate findings from previous studies showing successful cross-sectional disease-state classification of subjects using single modality (volumetric MRI) and multi-modality data in the DIAN, Knight ADRC, ADNI reprocessed, and AIBL reprocessed (when available) cohorts.
Aim 3:Utilize the developed platform for deeper exploration of the dynamics of Alzheimer’s biomarkers between ADAD and LOAD cohorts.
Aim 4Explore the utility of deep neural networksfor examining longitudinal relationships between biomarkers and disease progression.