3669

Quantification of brain age using high-resolution 7T MR imaging and implications in major depressive disorder

Gaurav Verma¹, Yael Jacob¹, Laurel Morris², Priti Balchandani¹, and James Murrough²
¹Radiology, Icahn School of Medicine at Mount Sinai, NEW YORK, NY, United States, ²Psychiatry, Icahn School of Medicine at Mount Sinai, NEW YORK, NY, United States

A linear regression model to estimate brain age using imaging parameters was developed using data from 29 controls and applied to 35 patients diagnosed with major depression. The model estimated brain age with absolute error of 6 (HC) and 7.6 years (MDD) while over-estimating MDD age by 2.4 years.

Axial T1-weighted MRI of 38 year old male healthy control (left) whose age was overestimated by 2 years and 39 year old MDD patient (left) whose age was overestimated by 10 years by the model. Among the criteria that drove age over-estimation in the MDD subject were higher ventricle volume, lower inferior parietal lobe volume and lower supramarginal gyrus thickness than age-average.

Plots showing actual age versus projected age of the Brain age model along with linear regression trendlines with zero intercept.