Dear SPMers,
I am using SPM8 to analyse fMRI data for auditory experiments. I have been using "sparse sampling" or "sparse acquisitions" in the past, but have recently started using interleaved silent steady state (ISSS) imaging [NeuroImage 29 (2006) 774 – 782].
In a nutshell, ISSS is like "clustered acquisition", except that data is collected at regular intervals, even during the silent period. The slice-selection gradients are switched off during the silent period in order to keep the scanner quiet. As a result, ISSS does not have the varying T2 magnetisation effects that plagues clustered acquisition.
For my experiment, my MR TR is 2 s and my overall TR is 12 s. Within each 12 s period, the first 8 s are silent and the last 4 s are for acquisition of images. Because my MR TR is 2s, I acquire 4 "useless" images silently in the 8 s period and 2 useful images in the following 4 s; the cycle is then repeated for the duration of the session.
I am having trouble finding a way to analyse this data with SPM. I do have some questions:
1. What is the best way of setting up a design matrix? Should I simply discard the "useless" images acquired during the silent periods and proceed with the GLM as I would for a normal continuously sampled data? If not, how do I set the TR for the GLM model?
2. What is the best basis function type to use for ISSS? Should I use canonical HRF, canonical HRF with derivatives, or the FIR set? If I use the FIR set, what is a good window length and filter order?
I urgently need to know your thoughts and suggestions, SPMers, especially from those who have used ISSS before.
Thank you very much in advance,
Kwaku Akrofi, PhD
Postdoctoral Fellow
Department of Speech and Hearing Science
University of Illinois at Urbana-Champaign
[log in to unmask]
|