INTRODUCTION: Calculating the gap between estimated brain and chronological age may support early identification of age-related brain disorders and differentiation between patient groups. Our findings suggest that brain regions derived from MRI using FreeSurfer’s ASEG segmentation and the A2009 atlas predict age accurately.

AIM(S): To explore machine learning models predicting brain age and assess deviations from the healthy aging trajectory across different hearing loss groups using brain morphometric features derived from the A2009 atlas and ASEG.

METHOD(S): Subjects were divided according to the BIAP hearing loss classification system based on tonal audiometry - 5 bilateral hearing groups, as well as a unilateral hearing loss group. For each brain feature, a polynomial function was calculated for the control group (separately for each sex). To account for the brain aging trend, a quadratic trend was included in the model. An OLS regression model was fitted with the scaled feature as a dependent variable and group status, age, and age² as independent variables.

RESULTS: The observed alterations included, among others, the bilateral planum temporale and the left suborbital sulcus, encompassing regions central to auditory and language processing as well as orbitofrontal networks involved in emotional and decision-related functions. Additionally, the left Heschl’s gyrus within the superior temporal gyrus showed alterations, highlighting regions crucial for primary auditory processing. However, unilateral hearing loss groups seem to have smaller deviations, which may reflect the integration of inputs across left and right auditory pathways. Nevertheless, several factors should be taken into account when making such comparisons, including the age distribution of each group and the duration of hearing impairment.

CONCLUSIONS: Notably, profound hearing loss is often accompanied by decreased social interaction and cognitive stimulation, potentially contributing to reduced cognitive reserve.