Distinct spatial dimensions of changes in cortical structure during early adolescence.
Poster No:
1265
Submission Type:
Abstract Submission
Authors:
Mari Shishikura1, Alain Dagher2
Institutions:
1McGill University, Montreal, QC, 2Montreal Neurological Institute and Hospital, McGill University, Montreal, QC
First Author:
Co-Author:
Introduction:
Adolescence is a critical period when the brain undergoes structural and functional maturation. Previous studies have shown that mean cortical thickness peaks at 1-2 years of life and decreases thereafter, whereas total cortical surface area continues to expand during childhood until it peaks in late childhood/adolescence [1,2]. However, less is known about region-specific trajectories of these cortical structures during the maturation process. Moreover, we have yet to elucidate if different regions are coordinated in the way they mature.
In this study, we aimed to depict the region-specific developmental trajectory in the cortex during early adolescence. We further identified clusters of regions that had a common structural change pattern. Finally, we investigated whether these common change factors were associated with changes in impulsivity scores over the same period.
In this study, we aimed to depict the region-specific developmental trajectory in the cortex during early adolescence. We further identified clusters of regions that had a common structural change pattern. Finally, we investigated whether these common change factors were associated with changes in impulsivity scores over the same period.
Methods:
We used T1-weighted magnetic resonance images from Adolescent Brain Cognitive Development study [3], an ongoing longitudinal study where subjects were recruited at age 9-10 and are scanned every two years (baseline, 2-year-, and 4-year-follow-up data are currently available). We included subjects who were scanned at least twice, totaling 7904 subjects. Images were pre-processed and the cortex was segmented into 68 regions (DK atlas) by the ABCD working group. Site differences were corrected using Longitudinal ComBat [4].
For each region, we fitted a separate latent growth curve model [5] to obtain baseline measures (intercept) and annual change (slope) of cortical surface area and thickness. Estimate of change is reported as % change per annum. We then simultaneously estimated growth curves of all regions in the left hemisphere and obtained a latent correlation matrix, which represents how correlated slopes are among regions. Exploratory factor analysis (EFA) was conducted on the matrix to identify latent factors, which capture spatially distinct dimensions of cortical structural change. As a confirmatory factor analysis (CFA), we modeled growth curves simultaneously in the right hemisphere with the identified latent factors. Model fit was assessed based on Comparative Fit Index (CFI) and Standardized Root Mean Squared Residual (SRMR), with CFI>0.9 and SRMR<0.08 as a good fit. Finally, we assessed the Pearson correlation between the latent change factors and the change in UPPSP impulsivity scores [6].
For each region, we fitted a separate latent growth curve model [5] to obtain baseline measures (intercept) and annual change (slope) of cortical surface area and thickness. Estimate of change is reported as % change per annum. We then simultaneously estimated growth curves of all regions in the left hemisphere and obtained a latent correlation matrix, which represents how correlated slopes are among regions. Exploratory factor analysis (EFA) was conducted on the matrix to identify latent factors, which capture spatially distinct dimensions of cortical structural change. As a confirmatory factor analysis (CFA), we modeled growth curves simultaneously in the right hemisphere with the identified latent factors. Model fit was assessed based on Comparative Fit Index (CFI) and Standardized Root Mean Squared Residual (SRMR), with CFI>0.9 and SRMR<0.08 as a good fit. Finally, we assessed the Pearson correlation between the latent change factors and the change in UPPSP impulsivity scores [6].
Results:
Latent growth curve modeling revealed that while all regions undergo cortical thinning from age 10 to 14 (maximum in cuneus, 1.13%, minimum in entorhinal cortex, 0.13%), surface area show mixed direction of change. Specifically, regions in frontal lobe had surface area expansion (maximum in anterior cingulate, 0.54%), whereas regions in parietal and occipital lobe showed decrease in surface area (maximum in precuneus, 0.6%) (Fig 1). Although EFA suggested a two-factor solution for cortical thickness changes, CFA showed a moderate fit (CFI=0.81, SRMR=0.08). For cortical surface area, four-factor model had a good fit (CFI=0.94, SRMR=0.02). The loadings of the four factors are shown in Fig2. Factor 3, which had high loadings on frontal cortex, showed significant association with UPPSP lack of planning (r = -0.059, p=0.006).
Conclusions:
We showed that cortical regions can be grouped together in regard to the trajectory of development during early adolescence. These underlying spatial patterns may be crucial in assessing brain-behavior associations during development.
Lifespan Development:
Early life, Adolescence, Aging 1
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Normal Development 2
Keywords:
Cortex
Development
Modeling
STRUCTURAL MRI
Other - impulsivity
1|2Indicates the priority used for review
Provide references using author date format
2. Gilmore, J. H., Knickmeyer, R. C. & Gao (2018), W. 'Imaging structural and functional brain development in early childhood.' Nat Rev Neurosci 19, 123–137.
3. Karcher, N. R. & Barch, D. M. (2021), 'The ABCD study: understanding the development of risk for mental and physical health outcomes.' Neuropsychopharmacol. 46, 131–142.
4. Beer, J. C. et al. (2020), 'Longitudinal ComBat: A method for harmonizing longitudinal multi-scanner imaging data.' NeuroImage 220, 117129 (2020).
5. Burant, C. J. (2016), 'Latent Growth Curve Models: Tracking Changes Over Time.' Int J Aging Hum Dev 82, 336–350.
6. Lynam, D. R., Smith, G. T., Whiteside, S. P. & Cyders, M. A. (2006), 'The UPPS-P: Assessing Five Personality Pathways to Impulsive Behavior (Technical Report).'