Bin,

If those different runs are scanned consecutively in the same session, people usually concatenate them and then run regression analysis. However if you have enough data in each of your 5 runs and if you are interested in comparing differences among various runs, you can run individual subject analysis on each run separately.

> As I understand the output of option '-fa' is the main significant common
> activity areas for the whole 5 runs. Is that correct?

Main effect of a factor is the average effect of all simple effects through all levels of that factor, expressed with the sum of squares of the factor by averaging the differences between different pairs of the factor effects, and a F-statistic called omnibus or overall test. The corresponding null hypothesis is that there is no difference among those 5 runs. If there is a main effect for the 5 runs, it means at least one run is significantly different from others.

>And what does the output of "-acontr 1 1 1 1 1 AllRun" mean?

The null hypothesis for this contrast is that the sum of all 5 simple effects is 0. It should be equavilent to '-amean 1 AllRun' in your case, and you will get a t statistic for it. Some people use this contrast to do double thresholding. For example, they argue that if a constrast (e.g. difference between run 1 and run 2) is significant only this sum of simple effects is significant.

>In addition, could you please let me know how to read and find the
> corresponding content to each Sub-brick in the final output bucket
> dataset ?

For each effect or contrast requested in the 3dANOVA2 script, you will find two subbriks in the ouput, one being the intensity, and the other statistic (t or F). You can use AFNI graphic viewer to navigate through those subbriks and exlpore various significance regions.

Hope this is helpful,
Gang