I am analyzing data from a  2x2 design. I have a significant difference between one factor across a wide range of brain regions. I believe that, in at least some regions, this difference is driven by an increase in the duration of neuronal firing. Thus, the HRF increases both in amplitude and duration. 

I have attached a plot of some HRFs. They were generated by taking the beta values for the 4 conditions, and multiplying them by the basis functions (HRF, derivative, and dispersion, all convolved with a 5 second boxcar using spm_Volterra). 

The HRFs for 2 conditions (both associated with one factor) appear to meet my hypothesis. The HRFs are larger in amplitude and show a wider FWHM, which would be consistent with an increased duration of neuronal firing being the reason for the difference between regions. 

Is there any way I can demonstrate this statistically? I though of comparing the dispersion beta between conditions, but I am not sure if this is appropriate or what it would mean. 

Any advice or comments are welcome : )

Colin