Nothing Special   »   [go: up one dir, main page]
Skip to main content
. Author manuscript; available in PMC: 2019 Jun 21.
Published in final edited form as: Neuroimage. 2013 Mar 22;76:400–411. doi: 10.1016/j.neuroimage.2013.03.015

Table 5.

Skeletonized performance at the optimum parameter settings for each dataset, compared to registration performance of TBSS. Performance on the Rotterdam rescan data is computed using the registration settings determined to be optimal based on the Rotterdam baseline data. p-values listed are computed for paired t-tests, comparing the registration performance for all 30 subjects across both registration algorithms and between registration algorithms and TBSS.

Algorithm Rotterdam baseline Rotterdam rescan Oxford data
FNIRT 0.685 0.674 0.690
Elastix 0.689 0.677 0.686
TBSS 0.643 0.636 0.647
FNIRT–Elastix (p-value) − 0.004 (0.002) − 0.003 (0.01) 0.005 (<10−6)
FNIRT–TBSS (p-value) 0.04 (<10−6) 0.04 (<10−6) 0.04 (<10−6)
Elastix–TBSS (p-value) 0.05 (<10−6) 0.04 (<10−6) 0.04 (<10−6)