Table 3:

Interobserver reliability and precision of hippocampal FA and MD values (n = 20)

1st Investigator Mean ± SD (Range)2nd Investigator Mean ± SD (Range)Diff ± SDt valuerrsICCSEMCI 95%CVR95 (%)
FA left247 ± 26243 ± 285 ± 131.58.89.87.939.5±193.9%26 (10.8%)
(235–259)(179–283)
FA right228 ± 20230 ± 23−2 ± 14−0.63.78.86.8810.1±204.1%28 (12.2%)
(219–238)(198–286)
FA mean238 ± 16236 ± 201 ± 110.52.85.87.907.6±153.1%21 (8.9%)
(230–245)(195–279)
MD left782 ± 46776 ± 285 ± 28.85.81.82.8419.8±392.5%55 (7.0%)
(760–803)(728–821)
MD right769 ± 27766 ± 233 ± 18 (25)0.68.75.79.8512.9±251.7%36 (4.6%)
(756–782)(721–816)
MD mean775 ± 30)771 ± 23)4 ± 18 (27)1.02.81.90.8712.7±251.6%35 (.5%)
(761–790(733–817
  • Note.— FA indicates fractional anisotropy (×103); MD, mean diffusivity (μm2/s); Diff, difference of 1st assessment of first investigator (cf. Table 1, 1st column) and assessment of 2nd investigator ± SD of difference; t value, corresponding to paired t test, all P values >.10; r, Pearson correlation coefficient; rs, Spearman rank correlation coefficient; ICC, intraclass correlation coefficient; SEM, standard error of measurement based on within-subject variance; CI 95%, 95% confidence interval (±1.96 × SEM); CV = √(mean of within subject variance/mean squared); R95, repeatability according to Bland and Altman20: R95WSV = √2 × 1.96 × SEM; %, percentage of mean = R95/mean × 100.