Skip to main content
[Preprint]. 2024 May 23:2024.05.21.595190. [Version 1] doi: 10.1101/2024.05.21.595190

Table 3.

Dice score (average Dice ± standard deviation) comparison of single-modality and multi-modality models in five-fold cross-validation (Bold numbers represent the highest values among all models)

Subregions Single-modality model Multi-modality model without ModAug Multi-modality model with ModAug

Dice Dice

CA1 0.814 ± 0.036 0.817 ± 0.034 0.817 ± 0.036
CA2 0.735 ± 0.059 0.747 ± 0.054 0.741 ± 0.055
CA3 0.705 ± 0.064 ** 0.728 ± 0.058 0.724 ± 0.057
DG 0.870 ± 0.026 ** 0.871 ± 0.026 ** 0.866 ± 0.023
SUB 0.861 ± 0.026 0.862 ± 0.024 0.862 ± 0.026
Tail 0.845 ± 0.053 *** 0.859 ± 0.053 0.861 ± 0.052
ERC 0.846 ± 0.043 *** 0.845 ± 0.042 *** 0.855 ± 0.040
BA35 0.730 ± 0.080 *** 0.727 ± 0.072 *** 0.751 ± 0.073
BA36 0.812 ± 0.057 *** 0.802 ± 0.058 *** 0.825 ± 0.052
PHC 0.815 ± 0.065 0.818 ± 0.064 0.822 ± 0.062
**

p < 0.01;

***

p < 0.001 in paired t-test comparison between corresponding model and multi-modality model with ModAug