Skip to main content
. 2015 Apr 14;7:48. doi: 10.3389/fnagi.2015.00048

Table 4B.

Post hoc comparison results for the Individual Binary Threshold method.

Diagnostic tasks Tractography algorithm (I) Threshold (J) Threshold Mean difference (I-J) Sig. 95% confidence interval
Lower bound Upper bound
AD vs. MCI Tensor-RK2 0.05 0.30 -0.09325 0.023 -0.01800 -0.0065
0.40 -0.08961 0.035 -0.1763 -0.0029
Hough 0.05 0.25 -0.10897 0.034 -0.2142 -0.0038
MCI vs. NC Tensor-TL 0.05 0.35 -0.10285 0.014 0.0109 0.1948
PICo 0.05 0.15 0.11037 0.028 0.0056 0.2151
0.35 0.12817 0.004 0.0234 0.2329
0.40 0.12654 0.005 0.0218 0.2313

The “Sig.” column shows the SPSS adjusted p-value and only values below 0.05 are treated as nominally significant (Please refer to the footnote for detailed explanation). Only comparisons that passed Bonferroni correction are shown here. 95% confidence interval is on the mean difference (I-J). Using the Individual Binary Threshold as the feature extraction method, the AUCs from some tractography algorithms may be statistically affected by the threshold values chosen for specific diagnostic tasks.