Skip to main content

Table 5 A list features to be used for lesion-free outcome prediction

From: Mining multi-site clinical data to develop machine learning MRI biomarkers: application to neonatal hypoxic ischemic encephalopathy

CategoriesDetails of features
I. Fiber tract featuresI.1. Histogram analysis (0, 25, 50, 75 and 100-percentile) of T1, T2, DWI, ADC, ZT1, ZT2, ZDWI, ZADC signal values within each of the 28 major fiber bundles as defined in the JHU atlas [123]
I.2. Skewness (asymmetry), kurtosis (flatness), uniformity and randomness (entropy and standard deviations) of T1, T2, DWI, ADC, ZT1, ZT2, ZDWI, ZADC signal values in each brain structures
II. Regional anatomy featuresII.1. Histogram analysis (0, 25, 50, 75 and 100-percentile) of T1, T2, DWI, ADC, ZT1, ZT2, ZDWI, ZADC signal values within the brain and each of the 61 auto-segmented brain structures/regions
II.2. Skewness (asymmetry), kurtosis (flatness), uniformity and randomness (entropy) of T1, T2, DWI, ADC, ZT1, ZT2, ZDWI, ZADC signal values in the brain and 61 auto-segmented regions
II.3. Volume of the 61 auto-segmented structures/regions as measured in T1 image
II.4. Left/right asymmetry in features II.1–II.3