I have an event-related design. The 1st level included the canonical HRF and its time and dispersion derivatives, because we want to look if there are difference in response shape between 2 groups. At the 1st level analysis, I defined a t-contrast for each basis function (HRF, TD and DD). The 2nd level analysis employed an ANOVA 3 (basis functions) x 2 (groups). I did an F test defined as:

1  -1  0  0   0 0

0  0   1 -1   0 0

0  0   0 0    1 -1

(group1 - group2: HRF; group1 - group2: time der. and group1 - group2: dispersion der.). This tells me which voxels have different activity between the 2 groups with respect to either amplitude, either time, or dispersion. I plotted the results, but I found it difficult to understand/explain, as it's an F test. Moreover, as I understood from Friston's paper (Friston et al. 1998, Neuroimage, 7:30-40) this tells me how well the model fits or have I misunderstood the things?

My question is how can I look for differences in response shapes between 2 groups combining the 3 basis functions?

Any suggestion is more than welcome.

Thank you!

Kind regards,

Ramona