Synchronization between stimulus refresh rate and TR


Are there any limitations you are aware of for using a TR which isn’t a multiple of the task presentation frame rate?
(for example: TR of 540 ms for a stimulus presented at frame rate of 60 Hz)


I think this is a good answer. Classically, many have advocated making sure that the inter-trial interval for event related-designs is not evenly divisible by the TR, yielding a more jittered sampling of the HRF. These ideas were popular before multi band, where TRs were quite long (e.g. 2-3.5s). The shorter TRs of the multi band era probably mean this is not a big effect. So I would consider the psychological effects (e.g. Bereitschaftspotential) for regularly spaced events versus jittered events.

Some analysis tools like the Siemens real-time fMRI require block designs and require the blocks are evenly divisible by the TR.

For nice guides on fMRI design see this PDF and this web page.

1 Like