Table 1.
Sample Characteristics.
| Feature | Training-Test data |
Training vs Test Kstest2-, p-values | Range | ||
|---|---|---|---|---|---|
| Training dataset | Test dataset | ||||
| Original sample (n = 96) | 71 | 25 | |||
| High quality imaging data (n = 88) | |||||
| Count | 64 | 24 | |||
| Sex (F/M) | 19/45 | 14/10 | 4.9, 0.02 | ||
| Freezing status (F/nF) | 28/36 | 10/14 | <0.01, 0.99 | ||
| Age | 69.0 (10) | 67.5 (17.75) | 0.27, 0.31 | 50 – 88 | |
| Disease duration | 5.9 (7.6) | 3.0 (3.0) | 0.34, 0.09 | 0.3 – 24.6 | |
| H&Y | 2 (0) | 2 (0.25) | 0.05, 0.9 | 1 – 4 | |
| LEDD | 678.5 | 450 | 0.37, 0.09 | 0 – 8680 | |
| MoCA | 25.1 (3.8) | 25.9 (3.6) | 0.15, 0.9 | 14 – 30 | |
| MDS-UPDRS III | 39.7 (14.2) | 39.5 (9.4) | 0.19, 0.77 | 13 – 72 | |
| MDS-UPDRS total | 67.5 (22.3) | 63.1 (17.7) | 0.18, 0.86 | 29 – 123 | |
Training dataset comes from Siemens Trio scanner. Test dataset comes from Siemens Prisma scanner. For differences in training and test set two sample Kolmogorov–Smirnov test was run for continuous variables and Chi squared test for sex and freezing status.