 |
 |
Dear Hweeling and Martin Note that the value entered for TE via the GUI is not used – it defaults to 0.04 if I remember correctly. You can change this value by editing the script spm_gx_fmri.m. This is scheduled to be fixed for the next release. Best Peter From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Hwee Ling Lee Sent: 07 August 2018 14:18 To: [log in to unmask] Subject: Re: [SPM] Resting state DCM using multi-echo EPI sequence Dear Martin, Thank you very much for the prompt response! Cheers, Hweeling On Tue, 7 Aug 2018 at 15:03, Martin Havlicek <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear Hweeling, If you do simple averaging over time-courses acquired with different TEs then you could also specify only the average TE in the DCM.TE field. If you do the weighted sum as I have recommended then you could use the TE that is theoretically optimal for the magnetic field strength data (represented mainly by tissue parenchyma) were acquired with. This assumes that the optimal TE was covered within your TE range. Practically speaking, if you have acquired data with 3 T, and I assume that your voxels will be mixture of gray matter tissue but possible also some larger vessels, then you can pick for optimal TE anything between 30 to 45 ms. Hope this helps. Best, Martin On 07/08/2018 14:45, Hwee Ling Lee wrote: Dear all, Thanks for the suggestions and feedback. I have another question with regards to spectral DCM analyses using multi-echo EPI images. What should be the input for DCM.TE in the case of multiple echos? I assume that the variable only accepts one input numerical value? Cheers, Hweeling On Mon, 23 Jul 2018 at 22:27, Martin Havlicek <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear Hweeling, Although one could come up with more accurate and specific model of multi-echo BOLD data, reasonably good solution can be obtained by optimally combining different echo-times as suggested by Poser, A.B. and Norris D. (2009): Investigating the benefits of multi-echo EPI for fMRI at 7 T, NeuroImage. 1./ First you fit T2* (and S0) using mono-exponential model (S0*exp(-TE/T2*)) into your multi echo ROI data (in resting-states data it could be based on time and ROI average as you are trying to fit T2* for "baseline") 2./ Then you fitted T2* in Equation (1) from Poser and Norris (2009) to calculate weights. 3./ Optimally weighted BOLD signal is obtained as linear combination of multi-echo BOLD signals; e.g. BOLD(OPT) = w(1)*BOLD(TE1) + w(2)*BOLD(TE2) + w(3)*BOLD(TE3) + ... Hope this helps. Best regards, Martin On 23/07/2018 08:01, Adeel Razi wrote: Dear Hweeling I am not an expert on this but are these multi-echo sequences simultaneous acquisitions or are different runs? If they are simultaneous then averaging these out would be easiest/sensible(?) while if different sessions then concatenating them would be more appropriate. You can then use the standard pipeline for preproessing. However I would leave this to experts here having more experience with these acquisitions. Best wishes, Adeel On Fri, Jul 20, 2018 at 6:31 PM, Hwee Ling Lee <[log in to unmask]<mailto:[log in to unmask]>> wrote: Dear DCM experts, I was wondering if there would be an issue with using multi-echo EPI images to examine the intrinsic dynamics of resting state. Could someone also refer to me the preprocessing and analyses pipeline using multi-echo EPI images in SPM? I searched on google but I find limited information with regard to this. Thanks. Cheers, Hweeling