Kaio Felippe Secchinato1, Pedro Henrique Rodrigues da Silva1, Júlia Palaretti1, and Renata Ferranti Leoni1
1Departamento de Física, University of São Paulo, Ribeirão Preto, Brazil
1Departamento de Física, University of São Paulo, Ribeirão Preto, Brazil
Early detection of Alzheimer's disease (AD) increases the benefits of treatment. However, it is still a challenging question which biomarkers are useful for early diagnosis. We used supervised machine learning algorithms to separate groups that evolved into Alzeihmer's disease or not.
Figure 1: Average resting-state CVR maps for
both groups: A) cognitively normal elderly who convert to AD and B) control
group.
Figure
2: Box plot comparing model results with the training set using the Caret R
Package and the following features: a) NP measures, b) CVR measures, and c)
mixed measures (NP+CVR). Spec = Specificity; Sens = Sensibility; ROC = Receiver
Operating Characteristic.