Skip to main content
. Author manuscript; available in PMC: 2022 Dec 26.
Published in final edited form as: Med Image Anal. 2022 Apr 22;79:102463. doi: 10.1016/j.media.2022.102463

Fig. 12.

Fig. 12.

We trained the model using (A) 60 CN, (B) 60 MCI, and (C) the mix of 60 CN and 60 MCI. For each model the same testing dataset (72 CN and 58MCI) was used. Each matrix in (A–C) represents the group-level MSE. The mean value of each MSE matrix was listed in (D).