Table 5.
Analysis of the relationship between WMH volumes and age – scanner upgrade and retrospective data merging scenario – Summary of the one-way ANCOVA test and Johnson–Neyman (J-N) procedure.
| Scanner upgrade scenario |
Retrospective data merging scenario |
||||||
|---|---|---|---|---|---|---|---|
| Analysis option A | Analysis option B | Analysis option C | Analysis option D | Analysis option E | Analysis option B | Analysis option C | |
| One-way ANCOVA Slope | F(1, 709) = 0.043, p = 0.836 | F(1, 709) = 0.019, p = 0.892 | F(1, 709) = 8.358, p = 0.004 ** | F(1, 709) = 2.303, p = 0.129 | F(1, 709) = 0.774, p = 0.379 | F(1, 2994) = 10.705, p = 0.001 ** | F(1, 2994) = 8.334, p = 0.004 ** |
| One-way ANCOVA Intercept | F(1, 709) = 43.678, p < 0.001 *** | F(1, 709) = 4.772, p = 0.029 * | —————————– | F(1, 709) = 3.789, p < 0.052 | F(1, 709) = 0.025, p = 0.874 | —————————– | —————————– |
| J-N Region of non-significance (age interval in years) | —————————– | —————————– | 51.13–57.95 | —————————– | —————————– | N/A | 66.06–75.54 |
Options tested in our study are: (I) for the scanner upgrade scenario: (A) without BC, single-site training, FA included; (B) with BC, single-site training, FA included; (C) with BC, site-specific training, FA included; (D) with BC, mixed training, FA included; (E) with BC, mixed training, FA excluded; (II) for the retrospective data merging scenario: (A) with BC, site-specific training, FA excluded; (B) with BC, mixed training, FA excluded. The one-way ANCOVA test evaluated across-scanner differences (between WH1/WH2 or between WH1/WH2/UKB) characterising regression slope (interaction between age and scanner) and intercept at mean age (main effect of scanner) in the linear model relating WMH% to age. Results of the ANCOVA test are reported in terms of F(df)- and p-values: * (< 0.05), ** (< 0.01), *** (< 0.001). When the hypothesis of homogeneous regression slopes was not met, we used the Johnson–Neyman procedure to evaluate the across-scanner differences. Results for the Johnson–Neyman procedure are reported in terms of age intervals. Legend: N/A = Not Available (i.e. there was no age interval where WMH% were not different across scanners).