Fig. 9: Performance of ANTsX deep learning–based mouse brain parcellation. | Nature Communications

Fig. 9: Performance of ANTsX deep learning–based mouse brain parcellation.

From: The ANTsX ecosystem for mapping the mouse brain

Fig. 9

Evaluation of the ANTsX deep learning--based mouse brain parcellation on a diverse MRI cohort. a T2-weighted DevCCF P56 template with the six-region parcellation: cerebral cortex, nuclei, brain stem, cerebellum, main olfactory bulb, and hippocampal formation. b Example segmentation result from a representative subject (NR5, Day 0) using the proposed deep learning pipeline. c Box plots show Dice overlap across subjects for each registration approach and region. The centre line is the median; box bounds are the interquartile range (25th--75th percentiles); whiskers extend to the minimum and maximum values within 1.5 × IQR of the lower/upper quartiles; points beyond the whiskers are outliers.

Back to article page