block averaging for varying block duration
Dear fNIRS experts,
I am currently doing fNIRS analysis via Homer3 and have accounted an issue whereby my blocks have varying timings. My task involves participants making button presses and hence, the faster they react the faster they complete each block. The blocks vary between 30s-50s and I have 9 blocks in total with 20 channels across 6 ROIs.
I have a few options in mind on how to analyse:
1. Block average using only 5s-30s whereby all participants will reach this duration.
2. Peak amplitude. Looking at the peak HbO amplitude for each block (i.e. some participants may have a peak HbO value at 12s in block 1, some maybe at 20s in block 2 etc..., then I will select this peak value at each block and average. So, I will then get a peak HbO value for each participant.
3. Peak average. Similar to number 2 above, but instead of just getting one peak value, I will include -5s to +5s before and after the peak value. So, averaging 10s within the peak HbO of the block.
Any help and suggestions on which of the above methods sounds reasonable are greatly appreciated! I am really grateful if anyone who sees this thread could share other analysis method with article references on why it is better and how should I do it.
I have a question about #2 in your analysis. I normally compare block averages between two different groups using the mean HRF. But I am interested in comparing peak amplitudes. How did you determine these values in Homer3?