The effect of template choice on morphometric analysis of pediatric brain data

Abstract

The present study investigated the "template effect" on the morphometric analysis of a pediatric brain MRI database obtained from 8-year-old children through various measures of surface and volumetric morphologies. We first constructed an age-appropriate template from an independent set of pediatric brain images and then compared it with a well-known adult brain template, the ICBM152, in terms of the morphometric features that resulted from our pediatric database. The individual cortical surface acquired based on the pediatric template exhibited, on average, a significantly thinner cortex (-1.66+/-1.60%, t=-15.18), larger cortical surface areas (0.31+/-0.70%, t=6.52), and a higher degree of cortical folding (0.08+/-0.13%, t=8.72) while compared with those based on the adult template. We also found a significant increase in the cerebrospinal fluid volume (-2.63+/-4.84) for the adult template based brains and the cortical gray matter (GM) volume (6.10+/-7.81) for the pediatric template based brains. The cross-correlation of pediatric template based individual brain data (0.95 without brain mask) was significantly higher than those of adult template based (0.88) and the amount of deformation during non-linear spatial normalization was significantly reduced when using the pediatric template (average magnitude of deformation in the cortical GM class: 1.71 mm vs. 2.23 mm, t=12.39). In addition, an "internal" pediatric template, taken from the study subjects themselves, was generated and compared with the "external" pediatric template for reference. There was no significant difference between these two pediatric brain templates and associated tissue probability maps. The results show that it is necessary to be cautious when interpreting results from pediatric imaging studies based on adult reference data.