TABLE 4.
Ridge linear regression on outcome variable THI score.
| Variable | Regression coefficient | Scaled coefficient | Standard error (scaled) | t-Value (scaled) | p-Value |
| Tonsil GMV | –0.000 | –0.802 | 4.833 | 0.166 | 0.869 |
| (P)FL-complex GMV | –0.006 | –5.536 | 5.222 | –1.060 | 0.292 |
| Total cerebellar cortical GMV | 0.000 | –0.069 | 4.485 | –0.015 | 0.988 |
| Age | –0.136 | –15.199 | 6.911 | –2.199 | 0.030* |
| Sex | 0.170 | 0.788 | 7.167 | 0.110 | 0.913 |
| PTA | –0.037 | –7.458 | 6.235 | –1.196 | 0.235 |
| TIV | –0.004 | –5.223 | 6.712 | –0.778 | 0.439 |
| Original study | |||||
| Langers (2014) | 2.684 | 10.664 | 6.089 | 1.751 | 0.083 |
| Amaral and Langers (2015) | –0.024 | –0.084 | 6.841 | –0.012 | 0.990 |
| Koops et al. (2020a) | 1.396 | 6.267 | 7.348 | 0.853 | 0.396 |
The automatically chosen ridge penalty was 1.615. R2 was merely 0.024, with a p-value of 0.275 for the entire model. Only age was significantly associated with THI score. No association was present between THI score and tonsil GMV, (P)FL-complex GMV and total cerebellar cortical GMV. Significant variables are marked with an asterisk (*).