Figure 3.

Inter-rater and test-retest reliability of manual infarct volumetry versus measurement with a macro. (a) Infarct volumetry by the conventional manual method (left) and with an ImageJ macro (right) plotted for each sample. The plot shows the disagreement between raters, which is similar across different infarct sizes. Outliers are identified by local-weighted regression (>2× the standard residual) in order to analyze the influence of single raters. (b) Test-retest reliability of the two different methods for infarct volumetry, analyzed with the Bland–Altman plot.