Neural correlates of obsessive-compulsive symptoms in children from the general population

In this thesis we investigated structural and functional brain characteristics related to obsessive-compulsive symptoms (OCS) in children from the general population. In Chapter 1, we gave an introduction on OCD and outlined the rationale and aims for this thesis. In Chapter 2, we studied the brain morphology related to OCS in 2551 children aged 9 to 12 years from the Generation R Study. We also defined a subgroup of 164 children with symptoms above a clinical cut-off (‘probable OCD’) and with controls. Consistent with clinical OCD studies, children with probable OCD had a larger thalamus than controls. Our results in the general population are highly consistent with prior clinical studies. In Chapter 3, we studied thalamic subregional volumes in relation to OCS in 9-12 year old children from the Generation R Study. We used an automated thalamic segmentation pipeline to compare volumes of the anterior, ventral, intralaminar/medial, lateral and pulvinar subregions. Children with probable OCD had a larger ventral subregion compared with controls. Severity of OCS was negatively associated with pulvinar volume after accounting for overall thalamic volume. These results did not survive multiple comparisons correction, but do suggest that thalamic enlargement in probable OCD may be explained in varying degrees by multiple rather than a single thalamic subregion. In Chapter 4, we studied thalamic subregions in relation to clinical OCD within the ENIGMA-OCD Working group. We used an automated thalamic segmentation pipeline to compare volume of the anterior, ventral, intralaminar/medial, lateral and pulvinar subregions of the thalamus between 2,649 OCD patients and 2,774 healthy controls. Unmedicated pediatric OCD patients under the age of 12 years had larger lateral, pulvinar, ventral and whole thalamus volumes. Adult OCD patients compared with controls had smaller volumes across all subregions and smaller whole thalamic volume, which was mostly accounted for by the medicated patients and associated with symptom severity. The anterior thalamus was smaller in patients after adjusting for total thalamus volume. Consistent with findings from the Generation R Study, we conclude that thalamic volume differences related to OCD are global rather than driven by particular subregions. In Chapter 5, we investigated the longitudinal relationship between brain morphology and OCS in children from the Generation R Study between the ages of 9 to 16 years. Longitudinal relationships between thalamus, cortical morphology and OCS were investigated. We found no difference in thalamic volume between adolescents (at the age of 13 to 16 years) with probable OCD and controls. Thalamic volume at baseline (age 9-12) predicted a relative persistence of OCS at follow-up (age 13-16). Linear-mixed effect models showed that baseline thalamic volume predicts the change in cortical thickness of 12 OCD-relevant cortical regions. These findings suggest OCS-related thalamic volume differences diminish with age in the general population, consistent with findings from cross-sectional clinical studies. Further, thalamic volume at age 9-12 years is associated with change in OCS between 9-12 years and 13-16 years in the general population. In Chapter 6, we studied the functional brain characteristics of children aged 9-12 years from the Generation R Study. We used a combination of graph theoretical and subnetwork connectivity analysis to study the brain network at multiple scales: global network topology, subnetwork connectivity and network participation of thalamic nodes. We showed that children endorsing at least one OCS had higher modularity, lower connectivity between frontoparietal, limbic and visual networks as well as altered participation of the lateral prefrontal thalamus node. Further analysis revealed that these associations were primarily accounted for by hand washing symptoms. These results suggest that the network characteristics of OCS in children from the general population are partly symptom-specific and severity-dependent.