Rosella Tro' ^1,2* Monica Roascio ² Domenico Tortora ³ Mariasavina Severino ³ Andrea Rossi ^3,4 Eleftherios Garyfallidis ⁵ Gabriele Arnulfo ^2,6 Marco Fato ² Shreyas Fadnavis ⁷

• ¹ University of Genoa, Genoa, Italy
• ² Department of Computer Science, Bioengineering, Robotics and Systems Engineering, School of Mathematical, Physical and Natural Sciences, University of Genoa, Genoa, Italy
• ³ Center for Neuroradiology and Interventional Radiology, Giannina Gaslini Institute (IRCCS), Genoa, Italy
• ⁴ Department of Health Sciences, School of Medical and Pharmaceutical Sciences, University of Genoa, Genoa, Italy
• ⁵ Department of Intelligent Systems Engineering, Luddy School of Informatics, Computing and Engineering, Indiana University Bloomington, Bloomington, Indiana, United States
• ⁶ Neuroscience Center, University of Helsinki, Helsinki, Uusimaa, Finland
• ⁷ Harvard Medical School, Boston, Massachusetts, United States

Objective: High Angular Resolution Diffusion Imaging (HARDI) models have emerged as a valuable tool to investigate microstructure with a higher degree of detail than standard diffusion MRI. In this study, we explored the potential of multiple advanced microstructural diffusion models in investigating preterm birth to identify non-invasive markers of altered white matter development. Approach: Rather than focusing on a single MRI modality, we worked on a compound of HARDI techniques in 46 preterm babies studied on a 3T scanner at term equivalent age and in 23 control neonates born at term. Furthermore, we examined discriminative patterns of preterm birth using multiple analysis methods, drawn from two only seemingly divergent modeling goals, namely inference and prediction. We thus resorted to: (i) a traditional univariate voxelwise inferential method as the Tract-Based Spatial Statistics approach (TBSS); (ii) a univariate predictive approach as the Support Vector Machine (SVM) classification; (iii) a multivariate predictive Canonical Correlation Analysis (CCA). Main results: TBSS analysis showed significant differences between preterm and term cohorts in several white matter areas for multiple HARDI features. SVM classification on skeletonized HARDI measures produced satisfactory accuracy, particularly for highly informative parameters about fiber directionality. Assessment of the degree of overlap between the two methods in voting for the most discriminating features exhibited a good, though parameter-dependent rate of agreement. Finally, CCA identified joint changes precisely for those measures exhibiting less correspondence between TBSS and SVM. Significance: Our results suggest that a data-driven intramodal imaging approach is crucial for gathering deep and complementary information. The main contribution of this methodological outline consists of thoroughly investigating prematurity-related white matter changes through differing inquiry focuses, with a view to addressing this issue, both aiming towards mechanistic insight and optimizing predictive accuracy.