Fatin Zainul Abidin - Peter Wijeratne - CMIC/WEISS joint seminar series

Fatin Zainul Abidin

Title: Neuroimaging genetics of brains in older adults with hearing loss

Abstract: Hearing loss is one of the most common geriatric diseases. Looking into neural correlates of hearing loss and their genetic underpinnings is potentially beneficial in developing interventions and treatment for hearing loss. I will present an ongoing study where we investigate glucose metabolism (using FDG-PET) in 1000 people with and without hearing loss from the Alzheimer's Disease Neuroimaging Iniative (ADNI) database; the largest study of its kind so far. Furthermore, I will show results from our investigation into the contribution of genetic markers to glucose metabolism in brain regions associated with hearing loss.

Peter Wijeratne

Title: Gaussian Process Progression Modelling of structural MRI changes in Huntington’s disease

Abstract:

Longitudinal measurements of brain atrophy using imaging data can provide powerful markers for clinical trials in neurodegenerative diseases. However, patient data can be confounded by effects such as inter-subject variability, measurement noise and individual disease stage. Disease progression modelling aims to untangle both observed and unobserved effects using probabilistic methods to learn patterns of disease-related changes directly from data. The trained disease progression model can then be used to stage new patients, providing diagnostic utility for clinical practice and stratification for clinical trials.

Here I will present one of the latest advances in longitudinal disease progression modelling, specifically the Gaussian Process Progression Model (GPPM) and its application to longitudinal structural magnetic resonance imaging (sMRI) measurements in Huntington’s disease (HD). We use the GPPM to infer individual-level patient trajectories; estimate, for the first time, the relative time-scale of sub-cortical atrophy changes in HD; and identify when sMRI provides additional information to genetic markers. We conclude that the GPPM could increase power over standard imaging markers and provide additional utility for clinical trials in HD