Restructuring of amygdala subregion apportion across adolescence

Abstract

Total amygdala volumes develop in association with sex and puberty, and postmortem studies find neuronal numbers increase in a nuclei specific fashion across development. Thus, amygdala subregions and composition may evolve with age. Our goal was to examine if amygdala subregion absolute volumes and/or relative proportion varies as a function of age, sex, or puberty in a large sample of typically developing adolescents (N = 408, 43 % female, 10-17 years). Utilizing the in vivo CIT168 atlas, we quantified 9 subregions and implemented Generalized Additive Mixed Models to capture potential non-linear associations with age and pubertal status between sexes. Only males showed significant age associations with the basolateral ventral and paralaminar subdivision (BLVPL), central nucleus (CEN), and amygdala transition area (ATA). Again, only males showed relative differences in the proportion of the BLVPL, CEN, ATA, along with lateral (LA) and amygdalostriatal transition area (ASTA), with age. Using a best-fit modeling approach, age, and not puberty, was found to drive these associations. The results suggest that amygdala subregions show unique variations with age in males across adolescence. Future research is warranted to determine if our findings may contribute to sex differences in mental health that emerge across adolescence.

Keywords: Adolescent; Amygdala; Basolateral nuclear complex; Central amygdaloid nucleus; Corticomedial nuclear complex; Puberty; Sex characteristics.

Fig. 1

Probabilistic map of each amygdala subregion in a representative adolescent. Structural MRI with the left column showing the probabilistic maps of the 9 bilateral subregions shown in the coronal view (thresholded at probabilistic value of .15 for visualization purposes) and the right column showing the raw T1 weighted image in the same coronal slice. Key: LA, lateral nucleus; BLDI, basolateral dorsal and intermediate subdivision; BLVPL, basolateral ventral and paralaminar subdivision; BM, basomedial nucleus; CMN, cortical and medial nuclei; CEN, central nucleus; AAA, anterior amygdala area; ATA, amygdala transition area; ASTA, amygdalostriatal transition area.

Fig. 2

Outline of CIT168 segmentation on coronal slices through entire rostral-caudal view of the amygdala in the right hemisphere for four representative subjects. A maximum likelihood label was created for each subregion of the amygdala by creating a label based on a simple competition between probabilistic labels with a thresholded probabilistic value of .3 for visualization purposes; slices (1 mm) are sequential (no gap).

Fig. 3

Amygdala subregion volumes and relative volume fractions in adolescent males and females. Violin plots demonstrate the full distribution of the data by sex, by showing the frequency and extent of variation for each subregion for both A) absolute probabilistic volumes (mm³) and B) relative volume fraction (RVF; proportional to total amygdala volume) for each of the 9 bilateral amygdala subregion ROIs. Box plots demonstrate the median, first and third quartiles, and the upper and lower whiskers (3.5 times the interquartile range) for both C) absolute probabilistic volumes (mm³) and D) relative volume fraction (RVF; proportional to total amygdala volume).

Fig. 4

Sex differences in age associations with Absolute Probabilistic Estimate of the amygdala subregions. A) Basolateral ventral and paralaminar subdivision (BLVPL) and B) Central (CEN) and C) Amygdala transition area (ATA). Absolute Probabilistic Estimate plotted by age and sex (collapsed across hemispheres); solid lines reflect GAMM predicted fit estimates and dashed lines reflect 95 % confidence intervals; p-value represents the smooth age-by-sex interaction term.

Fig. 5

Sex differences in age associations with RVF of the amygdala subregions. A) Lateral nucleus (LA), B) Basolateral ventral and paralaminar subdivision (BLVPL) and C) Central (CEN) D) Amygdala transition area (ATA) and E) Amygdalostriatal transition area (ASTA). Relative Volume France (RVF) plotted by age and sex (collapsed across hemispheres); solid lines reflect GAMM predicted fit estimates and dashed lines reflect 95 % confidence intervals; p-value represents the smooth age-by-sex interaction term.