Table 3.
Hierarchical linear regression models predicting memory domain score
| Independent variable | ΔR2 | ΔF | Standardised Beta | Pvalue |
|---|---|---|---|---|
|
Model 1 (R2=0.39) | ||||
| LHCV |
0.15 |
7.9 |
0.42 |
0.008 |
|
Model 2 (R2= 0.49) | ||||
| AE |
0.24 |
17.1 |
0.51 |
0.0002 |
|
Model 3 (R2= 0.54) | ||||
| AE |
0.28 |
19.2 |
0.54 |
0.0001 |
| LHCV |
0.02 |
1.28 |
0.17 |
0.27 |
|
Model 4 (R2= 0.54) | ||||
| LHCV |
0.147 |
7.9 |
0.42 |
0.008 |
| AE | 0.151 | 10.5 | 0.46 | 0.003 |
LHCV: Left hippocampal volume adjusted for intracranial volume; AE: Atropine Effect.
In all models, age, gender, education were entered first, then baseline UPSIT. In model 3, AE was entered next, followed by LHCV. In model 4, this order was reversed. Restricted to non-demented (n = 39 for analyses including LHCV, n = 42 for model 2).