Tract-based spatial statistics (TBSS): application to detecting white matter tract variation in mild hypoxic-ischemic neonates

Annu Int Conf IEEE Eng Med Biol Soc. 2012:2012:432-5. doi: 10.1109/EMBC.2012.6345960.

Abstract

The aim of this study is to employ tract-based spatial statistics (TBSS) to analyze the voxel-wise differences in DTI parameters between normal and mild hypoxic-ischemic (HI) neonatal brains. Forty-one full term neonates (24 normal controls and 17 with mild HI injury) and 31 preterm neonates (20 normal controls and 11 with mild HI injury) underwent T1 weighted imaging, T2 weighted imaging and diffusion tensor imaging (DTI) within 28 days after birth. The voxel differences of fractional anisotropy (FA), λ1, λ2, and λ3 values between mild HI group and control group were analyzed in preterm and full term neonates respectively. The significantly decreased FA with increased λ2, λ3 in corticospinal tract, genu of corpus callosum (GCC), external capsule (EC) and splenium of the corpus callosum (SCC) in mild HI neonates suggested deficits or delays in both myelination and premyelination. Such impaired corticospinal tract, in both preterm and term neonates, may directly lead to the subsequent poor motor performance. Impaired EC and SCC, the additional injured sites observed in full term neonates with mild HI injury, may be causally responsible for the dysfunction in coordination and integration. In conclusion, TBSS provides an objective, independent and sensitive method for DTI data analysis of neonatal white matter alterations after mild HI injury.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Anisotropy
  • Biostatistics
  • Brain / pathology*
  • Case-Control Studies
  • Corpus Callosum / pathology
  • Diffusion Tensor Imaging / statistics & numerical data*
  • Humans
  • Hypoxia-Ischemia, Brain / diagnosis*
  • Hypoxia-Ischemia, Brain / pathology
  • Image Interpretation, Computer-Assisted / methods
  • Infant, Newborn
  • Infant, Premature
  • Magnetic Resonance Imaging / statistics & numerical data