More insights into early brain development through statistical analyses of eigen-structural elements of diffusion tensor imaging using multivariate adaptive regression splines
| Autor: | Hongtu Zhu, Dinggang Shen, Yasheng Chen, Weili Lin, Hongyu An, John H. Gilmore, Diane Armao |
|---|---|
| Rok vydání: | 2013 |
| Předmět: |
Male
Histology Nerve Fibers Myelinated Brain mapping Article White matter Image Processing Computer-Assisted medicine Statistical inference Humans Applied mathematics Generalized estimating equation Brain Mapping Multivariate adaptive regression splines General Neuroscience Superior longitudinal fasciculus Infant Newborn Brain Infant Regression analysis Diffusion Magnetic Resonance Imaging Diffusion Tensor Imaging medicine.anatomical_structure Child Preschool Anisotropy Regression Analysis Female Anatomy Psychology Neuroscience Diffusion MRI |
| Zdroj: | Brain Structure and Function. 219:551-569 |
| ISSN: | 1863-2661 1863-2653 |
| Popis: | The aim of this study was to characterize the maturational changes of the three eigenvalues (λ1 ≥ λ2 ≥ λ3) of diffusion tensor imaging (DTI) during early postnatal life for more insights into early brain development. In order to overcome the limitations of using presumed growth trajectories for regression analysis, we employed Multivariate Adaptive Regression Splines (MARS) to derive data-driven growth trajectories for the three eigenvalues. We further employed Generalized Estimating Equations (GEE) to carry out statistical inferences on the growth trajectories obtained with MARS. With a total of 71 longitudinal datasets acquired from 29 healthy, full-term pediatric subjects, we found that the growth velocities of the three eigenvalues were highly correlated, but significantly different from each other. This paradox suggested the existence of mechanisms coordinating the maturations of the three eigenvalues even though different physiological origins may be responsible for their temporal evolutions. Furthermore, our results revealed the limitations of using the average of λ2 and λ3 as the radial diffusivity in interpreting DTI findings during early brain development because these two eigenvalues had significantly different growth velocities even in central white matter. In addition, based upon the three eigenvalues, we have documented the growth trajectory differences between central and peripheral white matter, between anterior and posterior limbs of internal capsule, and between inferior and superior longitudinal fasciculus. Taken together, we have demonstrated that more insights into early brain maturation can be gained through analyzing eigen-structural elements of DTI. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink