> Under fmri specification of design, Microtime Resolution, and Microtime onset, we're supposed to specify the no. of time bins per scan - and not to change the default unless we have a long TR. The TR used in my study is 12s. What microtime resolution should I specify since my TR is long? Is there a method I can use to calculate it?
The *exact* microtime resolution (SPM.T) probably isn't critical; you
just want to sample frequently "enough". So if you assume that most
studies have a TR of 2 seconds or so, and SPM defaults to 16 time
bins, then you can use ~8 bins per second as a good estimate. So for
12 seconds, you could use 96 bins (or you could just round to 100).
The important thing with sparse imaging is to set the microtime onset
(SPM.T0) to the right bin. So let's say that it takes you 3 seconds
to acquire a volume (TA = 3 sec), and you acquire a volume every 12
seconds (TR = 12 sec). You would probably want to set SPM.T0 to be
around 1.5 seconds into your TR---that is, halfway through the actual
acquisition of the volume. You can figure out what bin this is by the
SPM.T0 = (TA/2)/TR * SPM.T
or in your case, with 100 time bins
SPM.T0 = round( (TA/2)/12 * 100)
Hope this helps!