An investigation of amygdala and hippocampal subregions and their relation to ageing in anxiety and related disorders

Ntwatwa, Ziphozihle

An investigation of amygdala and hippocampal subregions and their relation to ageing in anxiety and related disorders

Thesis / Dissertation

2024

Publisher

University of Cape Town

Department

Department of Psychiatry and Mental Health

Faculty

Faculty of Health Sciences

Abstract

Background Obsessive-compulsive disorder (OCD) and social anxiety disorder (SAD) are debilitating disorders that are associated with (inconsistent) evidence of hippocampal and amygdala volumetric abnormalities. In addition, both OCD and SAD are associated with accentuated biological aging, as indexed by cellular and molecular markers. Nevertheless, little is known about brain aging in OCD and SAD, or the extent to which inconsistencies in hippocampal and amygdala volume findings in these disorders may be due to the differential effect of age on the subfields from which these structures are composed. Accordingly, this dissertation set out to characterise differences in hippocampal and amygdala subfield volumes between healthy controls (HCs) and participants with OCD and SAD in large-scale MRI datasets and relate these to whole and regional brain aging. Methods Hippocampal and amygdala subfield volumes and brain age estimates were derived from T1 weighted MRI images from the OCD Brain Imaging Consortium (De Wit et al., 2014) and the European and South African Research Network in Anxiety Disorders (Bas-Hoogendam et al., 2017). Subfield volumes were segmented using an automated segmentation algorithm from Freesurfer (v6.0). The brain age analysis was performed by using a previously trained machine learning algorithm that provides brain age estimates for the whole brain, as well as for regions of interest (occipital, frontal, temporal, parietal, cingulate, insula, or cerebellar–subcortical features) (Kaufmann et al., 2019). Differences in relative brain age (brain predicted age difference; brain-PAD) were calculated by subtracting chronological age from the predicted brain age. Between-group differences (diagnosis vs HCs) in volumetric and brain-PAD estimates were assessed using a mixed-effects (d) model adjusted for several covariates. Subgroup analyses were performed to determine the association of the main findings with clinical characteristics. Finally, unique associations between subfield volumes and whole brain age were estimated using partial correlation analysis. Results There was no evidence for a difference in subfield volumes between individuals with OCD and HCs. However, we found that psychotropic medication use was associated with significantly smaller hippocampal dentate gyrus (d=-0.26, pFDR=0.042), molecular layer (d=-0.29, pFDR=0.042) and larger lateral (d=0.23, pFDR=0.049) and basal (d=0.25, pFDR=0.049) amygdala subfields than HCs. Individuals with OCD without psychotropic medication use had significantly smaller hippocampal CA1 (d=-0.28, pFDR=0.016) compared to HCs. No association was found for symptom severity. In contrast to the findings for OCD, individuals with SAD demonstrated significantly smaller basal (d= 0.32, pFDR=0.022), accessory basal (d=-0.42, pFDR=0.005) and corticoamygdaloid transition area (d=0.37, pFDR=0.014) amygdala subfields overall compared to HCs, and larger hippocampal CA3 (d=0.34, pFDR=0.024), CA4 (d=0.44, pFDR= 0.007), dentate gyrus (d=0.35, pFDR= 0.022) and molecular layer (d=0.28, pFDR=0.033). In addition, individuals with SAD without comorbid anxiety disorder had smaller lateral amygdala and hippocampal amygdala transition area, compared to HCs. No association was found for psychotropic medication use and symptom severity. Individuals with OCD (n=375) had significantly higher whole brain-PAD (+1.6 years, pFDR=0.006, d=0.20) compared to HCs (n=335), but no differences were observed in the regional models. The effect on whole brain brain-PAD estimates was largely driven by psychotropic medication use as higher relative brain age was evident in individuals with OCD with psychotropic medication use (+2.98 years, d=0.38, p <0.001) compared to HCs, but not in individuals without psychotropic medication use (+0.57 years, d=0.07, p =0.374) compared to HCs. No association was found for symptom severity. Partial correlation analysis found a significant negative association between hippocampal and amygdala volume and whole brain PAD in the OCD group (R=-0.224, p=0.00001), but not in the HC group (R=0.081, p=0.138), specifically the lateral nucleus (R=-0.18), CAT(R=-0.13), hippocampal fimbria (R=0.17), and hippocampal fissure (R=0.17) were significant in OCD. Individuals with SAD (n=107) had significantly higher whole brain-PAD (+2.5 years, d=0.33, pFDR=0.010) compared to HCs (n=137), and significantly higher regional brain-PAD in the temporal (+3.80 years, d=0.37, pFDR=0.008,), parietal (+3.57 years, d=0.38, pFDR=0.008), occipital (+3.26 years, d = 0.33, pFDR=0.010), and frontal regions (+2.97 years, d=0.33, pFDR=0.010,) compared to HCs. Brain PAD was higher in SAD without comorbid anxiety disorder, without MDD, and without psychotropic medication use. No association was found for symptom severity. There was no partial correlation between subfields and brain age. Discussion & Conclusion The evidence presented in the thesis suggests that 1) differences in subfield volumes between OCD and HCs were influenced by psychotropic medication use, which is consistent with previous studies that suggest that psychotropic medication status is a strong confounder for subcortical brain volumes observed in OCD, 2) differences in subfield volumes between SAD and HCs were observed in the areas associated with sensory information processing and these differences were partially influenced by psychiatric comorbidity, 3) both OCD and SAD were associated with accentuated brain aging with differential patterns in the whole and regional brain, dependent on clinical characteristics, and 4) only OCD relative brain age was associated with subfield volumes. It is unclear whether our findings in OCD and SAD reflect an adaptive response or are a pre-existing risk factor to these disorders, or both. Future longitudinal analysis is required to investigate whether the observed differences in subfield volume and brain age remain over time.

Keywords

mental health

Reference:

Collections

PhD / Doctoral

Full item page