Automated method for extracting response time of subject vocalizations in event-related fMRI experiments

Question

I have successfully extracted participants' overt word reading reaction time data from their responses within the MRI scanner. This is the function i have written in matlab http://www.mathworks.in/matlabcentral/fileexchange/38761-function-to-calculate-reaction-time but it is only working when there is only one voice response. It fails when audio file contain more than one human response.

Answer 1

Do I understand correctly that you identify the response like this:

Step 1: Calculate the average signal height in the set 

Step 2: Find the place where the signal is sufficiently higher than the average

In that case I suggest the following:

Start with a small subset (startpoint is the startpoint of the original set)

Do Step 1 and Step 2 for the subset

If nothing is found, increase the subset size and do Step 1 and Step 2 again.