Table 3 Selected variables of logistic LASSO regression models for discriminating prodromal Alzheimer’s disease (AD) from cognitively unimpaired amyloid-positive individual.

From: Exploration of neuroanatomical characteristics to differentiate prodromal Alzheimer’s disease from cognitively unimpaired amyloid-positive individuals

Regions of interest

Model 1

Model 2

Model 3

SUVR

Volume

SUVR

Volume

SUVR volume*

Temporal

L entorhinal cortex

 

 − 0.38

 

 − 0.36

 

R entorhinal cortex

 

 − 0.04

 

 − 0.06

 

Parietal

L superior parietal cortex

0.05

    

R inferior parietal cortex

 

 − 0.02

 

 − 0.01

 

R precuneus cortex

0.05

    

R superior parietal cortex

0.05

 

0.02

 

0.07

Subcortex

L Amygdala

 

 − 0.12

 

 − 0.12

 

L Hippocampus

 

 − 0.11

 

 − 0.13

 

R Amygdala

 

 − 0.10

 

 − 0.10

 

(Intercept)

1.10

1.19

1.19

Lambda

0.056

0.055

0.055

Accuracy

0.720

0.745

0.750

Precision

0.911

0.900

0.914

  1. L left, LASSO least absolute shrinkage and selection operator, R right, ROI region of interest, SUVR standardized uptake value ratio.
  2. Model 1 included SUVRs of all sub-ROIs; Model 2 included the volumes of all sub-ROIs, divided by intracranial volume; Model 3 included both the volumes and SUVRs of all sub-ROIs.
  3. *Values show β coefficients in each logistic LASSO models.
Back to article page