You can use the bayesian 1st level estimation (see screenshot) this corresponds to the method implemented in Penny et al 2005. I would of course suggest that you record pulse and respiration, and use nuisance variable regression instead, as described in Lund et al. 2006.

Penny, W., Kiebel, S. & Friston, K. Variational Bayesian inference for fMRI time series. NeuroImage 19, 727–741 (2003).

Lund, T. E., Madsen, K. H., Sidaros, K., Luo, W.-L. & Nichols, T. E. Non-white noise in fMRI: Does modelling have an impact? NeuroImage 29, 54–66 (2006).

Torben Ellegaard Lund
Associate Professor, PhD
Center of Functionally Integrative Neuroscience (CFIN)
Aarhus University
Aarhus University Hospital
Building 10G, 5th floor, room 31
Noerrebrogade 44
8000 Aarhus C
Denmark
Phone: +45 7846 4380
Fax: +45 7846 4400
http://www.cfin.au.dk
[log in to unmask]

Den 24/01/2014 kl. 11.13 skrev ashish sahib <[log in to unmask]>:

Dear SPM users

I am using the Multiband EPI sequence for the acquisition of my fmri data, and i have multiple sessions in which the TR ranges from 2s to 265 ms. But at short TR's the one-back autoregressive model will underestimate the autocorrelation in the fmri signal and may result in inflated t scores.

Is it possible to change the order of the autoregressive model.

I want to use the Short TR's to model single events of 500 ms duration. Is it possible to do such kind of analysis using spm.

Thanks
Ashish