|
Analysis of variance
|
SS
|
DF
|
MS
|
F
|
p-value
| |
|---|
|
Regression
|
48.79
|
4
|
12.20
|
F(4,57) = 41.39
|
P < .0001****
| |
|
SWI: putamen (CNR)
|
15.46
|
1
|
15.46
|
F(1,57) = 52.46
|
P < .0001****
| |
|
SWI: caudate (CNR)
|
2.00
|
1
|
2.00
|
F(1,57) = 6.78
|
P = .0117*
| |
|
SWI: globus pallidus (CNR)
|
2.98
|
1
|
2.98
|
F(1,57) = 10.10
|
P = .0024**
| |
|
SWI: thalamus (CNR)
|
0.79
|
1
|
0.79
|
F(1,57) = 2.68
|
P = .1074
| |
|
Residual
|
16.80
|
57
|
0.29
| | | |
|
Total
|
65.58
|
61
| | | | |
|
Parameter Estimates
|
Variable
|
Estimate
|
SE
|
95% CI
|
T
|
p-value
|
|---|
|
β0
|
Intercept
|
1.36
|
0.31
|
0.74; 1.98
|
4.39
|
P < .0001****
|
|
β1
|
SWI: putamen (CNR)
|
4.44
|
0.61
|
3.21; 5,67
|
7.24
|
P < .0001****
|
|
β2
|
SWI: caudate (CNR)
|
0.27
|
0.10
|
0.06; 0.48
|
2.60
|
P = .0117*
|
|
β3
|
SWI: globus pallidus (CNR)
|
0.19
|
0.06
|
0.07; 0,31
|
3.18
|
P = .0024**
|
|
β4
|
SWI: thalamus (CNR)
|
0.10
|
0.05
|
0.00; 0.18
|
1.64
|
P = .1074
|
|
Goodness of Fit
| | | | | |
|---|
|
DF
|
57
| | | | | |
|
Multiple R
|
0.86
| | | | | |
|
R2
|
0.74
| | | | | |
|
R2adj
|
0.73
| | | | | |
|
SS
|
0.74
| | | | | |
|
RMSE
|
0.11
| | | | | |
|
Multicollinearity
|
Variable
|
VIF
|
R2 with other variables
| | | |
|---|
|
β0
|
Intercept
| | | | | |
|
β1
|
SWI: putamen (CNR)
|
1.59
|
0.37
| | | |
|
β2
|
SWI: caudate (CNR)
|
1.18
|
0.15
| | | |
|
β3
|
SWI: globus pallidus (CNR)
|
1.29
|
0.22
| | | |
|
β4
|
SWI: thalamus (CNR)
|
1.21
|
0.17
| | | |
|
Normality of Residuals
|
Statistics
|
p-value
|
Passed normality test (α = .05)?
| | | |
|---|
|
Anderson and Darling (1952)
|
0.58
|
P = 0.13
|
Yes
| | | |
|
D’Agostino et al. (1990)
|
2.23
|
P = 0.33
|
Yes
| | | |
|
Shapiro and Wilk (1965)
|
0.97
|
P = 0.16
|
Yes
| | | |
|
Massey (distance) (1951)
|
0.08
|
P > .10
|
Yes
| | | |
- The table summarizes the multiple regression model of (βATP + PCr)/iP vs. SWI CNR values, including descriptive analyses. Apart from the overall significance test of the model, parameter estimates, goodness of fit, and necessary assumptions for multiple regression models were tested (absence of multicollinearity and normality of residuals). */**/***/***: significance levels (*: P ≤ 0.05, **: P ≤ 0.01, ***: P ≤ 0.001; ****: P ≤ 0.0001)
- CI confidence interval, CNR contrast-to-noise ratio, DF degrees of freedom, MS mean square, R2 coefficient of determination, R2adj adjusted coefficient of determination, RMSE root mean square error, SE standard error, SS sum of squares, SWI susceptibility-weighted imaging, T t-statistic, F F-Statistic, VIF variance inflation factor
