Table 4 Cross-validated GBM performance measures according to the modality combination.

From: A multimodal machine learning model for predicting dementia conversion in Alzheimer’s disease

Modality combination

Metrics

BA

SE

SP

AUC

Mean ± SD

Mean ± SD

Mean ± SD

Mean ± SD

demo

0.768 ± 0.109

0.814 ± 0.170

0.721 ± 0.086

0.830 ± 0.079

demo + A

0.878 ± 0.075

0.925 ± 0.061

0.831 ± 0.147

0.937 ± 0.065

demo + AN

0.919 ± 0.048

0.949 ± 0.059

0.889 ± 0.084

0.977 ± 0.026

demo + ANV

0.923 ± 0.053

0.966 ± 0.044

0.881 ± 0.094

0.984 ± 0.016

  1. GBM; Gradient Boosting Models. BA; Balanced Accuracy, SE; Sensitivity, SP; Specificity, AUC; Area Under ROC Curve, demo; demographic characteristics, A; amyloid PET image features, N; T1-weigted image features, V; T2-FLAIR image features.
Back to article page