MRI Processing Pipeline Variability and Infant Brain Morphometry Associations to 4-Month Infant Temperament

Abstract

Negative reactive temperament, an infant temperament characterized by fear of novelty, is associated with adolescent amygdala volume (Filippi et al, 2020) and adult prefrontal cortex (PFC) thickness (Schwartz et al, 2010). However, it remains unknown whether these differences in brain morphometry emerge in infancy. Further, evaluating this possibility is a challenge because few pipelines are optimized for processing infant magnetic resonance imaging (MRI) data. Thus, evaluation of available infant MRI processing pipelines is necessary prior to examining associations between negative reactivity and brain morphometry. This study examines (1) which MRI pipeline performs best for 4-month-old infant MRI data and (2) associations between temperament and brain morphometry. Behavioral reactivity was assessed by presenting novel stimuli to infants. High-resolution structural MRI data was acquired a few weeks later. MRI data were processed using the iBEAT (Dai et al, 2013), dHCP (Makropoulos et al, 2018), and CIVET (Ad-Dab’bagh et al., 2006) pipelines to obtain estimates of amygdala and PFC volume. The quality of segmentations of the three pipelines was then assessed. The processing pipelines showed differences in terms of quality of gray/white segmentation and percentage of processing failures. Overall, iBEAT performed the best with the highest percent of useable data. Using the iBEAT output, we examined the associations between infant brain morphometry and reactivity. Results indicated no significant association between amygdala or PFC volume and reactivity.