Fig. 4.

Schematic of the quantitative CFE evaluations performed using simulated test-statistic images. A FOD template (a) was used to generate a whole-brain tractogram (b) that was then filtered using SIFT (c). The SRI24 atlas (d) was used to extract tract ROIs (e), which were combined with a FOD template-derived white matter fixel-mask (f) to define fixel ROIs (g). h) For all mask fixels, connectivity to other fixels was computed using the tractogram in c. For each ROI, 1000 simulated test-statistic images were created by adding noise (i) to all fixels in f and g, to generate a noise only (j) and signal + noise image (k). Fixel images were smoothed with a range of kernel extents (l–n) and CFE enhanced (o) using a range of values for parameters E, H and C. Enhanced images (p and q) were evaluated by computing the area-under the curve (AUC) (s) of an AFROC curve (r).