The brain age gap in social anxiety disorder

Blake, Kimberly Vanessa

The brain age gap in social anxiety disorder

Master Thesis

2021

Abstract

Background: When an individual's brain appears ‘older' than expected based upon their chronological age, they may be at an increased risk for developing brain-related diseases and cognitive decline. There is growing evidence of advanced brain ageing in neuropsychiatric diseases. Social anxiety disorder (SAD) is a disabling mental illness, which has been associated with both structural brain deficits and advanced biological ageing. However, brain age research has yet to be conducted in adults diagnosed with SAD. The present study investigated whether adults with SAD showed an advanced brain ageing process, compared to healthy controls (HCs), and whether brain ageing in SAD patients is associated with clinical characteristics. Method: A systematic review of the literature was conducted to identify knowledge gaps in brain age research in psychiatric disorders before commencing with the present dissertation. Hereafter, a secondary data analysis of a large multi-site dataset was performed. T1-weighted structural MRI scans of 387 participants (SAD n=174, HC n=213) between the ages 18 and 60 years were included. These structural scans were segmented using both FreeSurfer and SPM12, after which they underwent quality control procedures. Brain age was estimated by two different machine learning models – Tobias Kaufmann's brain age model and James Cole's BrainageR. The primary outcome for analysis was the brain age gap (BAG), calculated by subtracting a participants' chronological age from their estimated brain age. General linear models were run to determine whether there was a significantly larger positive BAG in the SAD group (Kaufmann model n=100, Cole model n=155) compared to the HC group (Kaufmann model n=138, Cole model n=197) after adjusting for age, mean centred age2 and sex. The association between BAG and comorbid depression and anxiety, as well as medication use and symptom severity, was also assessed. Results: In the present study sample, predicted age was more strongly associated with chronological age for the Cole model estimates than the Kaufmann model estimates (Cole: Pearson correlation = 0.828, MAE = 4.78, SD = 3.96, versus Kaufmann: Pearson correlation = 0.576, MAE = 11.93, SD = 6.93). With the Kaufmann model, the SAD group had a significantly larger BAG than the HC group of almost one year (mean difference = 0.943 year, SE = 0.40, p = .019). In addition, with the Kaufmann model, patients without psychiatric comorbidities had a significantly larger BAG than HCs, of more than one year (mean difference = 1.242 year, SE = 0.49, p = .038). No difference was observed in BAG between patients with comorbidities and HCs (mean difference = 0.983 year, SE = 0.85, p = .749). In contrast, with the Cole model, the SAD group did not have a significantly larger BAG than the HC group (mean difference = 0.513 year, SE = 0.49, p = .383). Moreover, the Cole model found no significant difference in BAG between SAD patients with and without comorbidities, or between each of these groups and HCs (all p > .708). Finally, no significant associations were observed between the BAG and symptom severity and the BAG and medication use in SAD patients in the Cole or Kaufmann models. Conclusion: This study observed contradictory evidence for a larger BAG between patients with SAD than HCs. The differences observed between the Cole model and the Kaufmann model may be a result of the different information used to estimate brain age (voxel-based volumetric data, compared to cortical thickness/surface area and subcortical/cerebellar volumes, respectively). The models demonstrated largely overlapping confidence intervals for group mean difference in BAG, suggesting that if there is a positive BAG in adults diagnosed with SAD, it is likely to be small. This should be verified in future research by using multiple different machine learning models based on different feature sets, to obtain more reliable and robust brain age estimates.

Keywords

Neuroscience

Reference:

Collections

Masters

Full item page