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: | Jian Dong |
| Institution: | Harvard Medical School |
| Department: | Radiology |
| Country: | |
| Proposed Analysis: | In 2018, the National Institute on Aging (NIA) and Alzheimer’s Association (AA) shifted the research framework of Alzheimer Disease (AD) from a syndromic definition to a biological definition comprised of beta amyloid deposition (A), pathologic tau (T), and neurodegeneration (N), together known as the ‘ATN’ biomarker groups. Many of the ATN biomarker groups are rooted in molecular or anatomical imaging, including amyloid PET (for A), tau PET (for T), and anatomical MRI and FDG PET (for N). As such, staging along the AD pathobiological continuum is now possible using ATN imaging biomarkers, much like the TNM system is used for cancer staging. Along the continuum, cerebral amyloidosis is believed to induce or facilitate the spread of pathologic tau, and pathological tau is believed to be temporally proximate to neurodegeneration, which is believed to be a proximal cause of cognitive decline. While the interpretation of pathological distribution of pathological tau as assessed by tau PET is currently under investigation, the pathological cerebral amyloid deposition as assessed by amyloid PET is better understood with respect to the anatomical locations of pathological uptake [including anterior and posterior cingulate, medial orbitofrontal lobe, parietal lobe, and medial temporal lobe], as well as the quantification of pathological uptake [various literature have defined mean SUVr greater than 1.15 or 1.17], enabling the interpretation of beta amyloid status. The application of convolutional neural networks (CNN) to medical imaging has enabled machine learning (ML) image classification for various purposes including disease diagnosis. In the area of AD research, multiple recent literatures have successfully used CNNs to categorize based on cognitive status or imaging biomarkers such as FDG-PET. We aim to develop a ML approach composed of CNNs as well as unsupervised machine learning to enable automated classification based on the quantification of cerebral amyloid deposition. Further, in keeping with the pathobiological continuum of AD, we aim to use ML to further delineate the pattern of pathological tau as assessed by tau PET, in those positive for beta amyloid as compared to those who are negative for beta amyloid, as those with A- and T+ profiles are attributed to non-Alzheimer dementia. |
| Additional Investigators |
