Skip to main content

Table 2 Performance of the FCN models

From: Enhancing magnetic resonance imaging-driven Alzheimer’s disease classification performance using generative adversarial learning

(a)
1.5 T Accuracy Sensitivity Specificity F-1 MCC
ADNI test 0.8398 ± 0.0238 0.7363 ± 0.0514 0.9209 ± 0.0385 0.7972 ± 0.0325 0.6766 ± 0.0492
AIBL 0.8873 ± 0.0647 0.6309 ± 0.1393 0.9259 ± 0.0874 0.6126 ± 0.1016 0.5757 ± 0.1086
NACC 0.8157 ± 0.0224 0.6739 ± 0.0605 0.8989 ± 0.0616 0.7301 ± 0.0221 0.6032 ± 0.0425
(b)
3T* Accuracy Sensitivity Specificity F-1 MCC
ADNI test 0.8210 ± 0.0143 0.7411 ± 0.0312 0.8895 ± 0.0114 0.7923 ± 0.0195 0.6416 ± 0.0280
AIBL 0.9293 ± 0.0132 0.7143 ± 0.0000 0.9617 ± 0.0152 0.7276 ± 0.0365 0.6884 ± 0.0449
NACC 0.8429 ± 0.0069 0.7393 ± 0.0180 0.9037 ± 0.0134 0.7768 ± 0.0096 0.6585 ± 0.0149
  1. Accuracy, sensitivity, specificity, F1-score, and Matthew’s correlation coefficient are computed for the FCN models that used (a) 1.5-T scans and (b) 3-T* images, respectively
Back to article page

Alzheimer's Research & Therapy

ISSN: 1758-9193