Thanks all for your helpful responses. Finally, I'll combine the 2 directions (_LR and _RL) in a fixed-effect 2nd level analysis step (as Matt suggests), after separate first level steps. I'll let you know asap if this solves my problem, or any other interested remarks.
Cheers, On Tue, Oct 1, 2013 at 9:52 PM, Harms, Michael <[email protected]> wrote: > > When you concatenate two runs, the subsequent analysis is likely > assuming that you are providing it a continuous time series, which it was > not. The assumption underlying the temporal autocorrelation estimate (and > temporal filtering, if applied) is that the data is a continuous time > series. Now, one might argue that the impact of ignoring the discontinuity > with concatenated data is quite possibly minimal -- I don't know if anyone > has empirically studied or modeled that. But why not just use the > "standard" approach of combining the 2 runs in a fixed-effects, 2nd level, > within subject analysis, so that you don't have to worry about the issue. > Users should know that is in fact what we have done to generate the > individual subject task fMRI data that was just released as part of the HCP > Q3 release. > > cheers, > -MH > > -- > Michael Harms, Ph.D. > ----------------------------------------------------------- > Conte Center for the Neuroscience of Mental Disorders > Washington University School of Medicine > Department of Psychiatry, Box 8134 > 660 South Euclid Ave. Tel: 314-747-6173 > St. Louis, MO 63110 Email: [email protected] > > From: <Glasser>, Matt Glasser <[email protected]> > Date: Tuesday, October 1, 2013 2:31 PM > To: "Harms, Michael" <[email protected]>, "Yan, Shulin" < > [email protected]>, Elvis Dohmatob <[email protected]>, " > [email protected]" <[email protected]> > > Subject: Re: [HCP-Users] Phase-encoding-direction-dependent (LR, RL) EVs > files, which ones to use in first level GLM ? > > That's what I thought, but can you explain exactly why Mike? > > Thanks, > > Matt. > > From: <Harms>, Michael <[email protected]> > Date: Tuesday, October 1, 2013 2:30 PM > To: "Yan, Shulin" <[email protected]>, Matt Glasser < > [email protected]>, Elvis Dohmatob <[email protected]>, " > [email protected]" <[email protected]> > Subject: Re: [HCP-Users] Phase-encoding-direction-dependent (LR, RL) EVs > files, which ones to use in first level GLM ? > > > Just to follow-up on what Matt said, concatenating the two runs would > NOT be the way that we would recommend to process/combine the two runs per > task. > > Best, > -MH > > -- > Michael Harms, Ph.D. > ----------------------------------------------------------- > Conte Center for the Neuroscience of Mental Disorders > Washington University School of Medicine > Department of Psychiatry, Box 8134 > 660 South Euclid Ave. Tel: 314-747-6173 > St. Louis, MO 63110 Email: [email protected] > > From: <Yan>, Shulin <[email protected]> > Date: Tuesday, October 1, 2013 2:25 PM > To: Matt Glasser <[email protected]>, Elvis Dohmatob < > [email protected]>, "[email protected]" < > [email protected]> > Subject: Re: [HCP-Users] Phase-encoding-direction-dependent (LR, RL) EVs > files, which ones to use in first level GLM ? > > Hey DED, > > I just did the work. I directly combined the two data say, _RL followed > _LR. Then, you remain the EVs file of _LR, and add (232*0.72s) to the time > in _RL's EVs file. > ------------------------------ > *From:*[email protected] [ > [email protected]] on behalf of Glasser, Matthew [ > [email protected]] > *Sent:* 01 October 2013 20:02 > *To:* Elvis Dohmatob; [email protected] > *Subject:* Re: [HCP-Users] Phase-encoding-direction-dependent (LR, RL) > EVs files, which ones to use in first level GLM ? > > I'm not sure I understand what you are trying to do. Are you sure you > don't just want to use the individual subject task fMRI data we've already > just released? > > I don't think there is any way you can combine the _LR and _RL data in a > first level analysis. You need to combine them in the level 2 analysis > (which uses flameo not film_gls). > > Peace, > > Matt. > > From: Elvis Dohmatob <[email protected]> > Date: Tuesday, October 1, 2013 1:58 PM > To: "[email protected]" <[email protected]> > Subject: [HCP-Users] Phase-encoding-direction-dependent (LR, RL) EVs > files, which ones to use in first level GLM ? > > Hi all, > > I'm trying to run first level (single-subject) GLM on task fMRI HCP > data. But I'm unsure on how to use the timing info. > > Example: subject 100307, MOTOR > > I have stimulus timing info (cue.txt , lf.txt , lh.txt, rf.txt, rh.txt, > Sync.txt, t.txt) in both > 100307/tfMRI_MOTOR/unprocessed/3T/tfMRI_MOTOR_LR/LINKED_DATA/EPRIME/EVs/ > and > 100307/tfMRI_MOTOR/unprocessed/3T/tfMRI_MOTOR_RL/LINKED_DATA/EPRIME/EVs/, > but the contents of the corresponding files in both directions vary (for > example, the contents of > 100307/tfMRI_MOTOR/unprocessed/3T/tfMRI_MOTOR_LR/LINKED_DATA/EPRIME/EVs/lf.txt > are different from > 100307/tfMRI_MOTOR/unprocessed/3T/tfMRI_MOTOR_RL/LINKED_DATA/EPRIME/EVs/lf.txt). > How am I to combine/merge these phase-encoding-direction-dependent EVs in > my GLM ? > > N.B.: During the 'topup' step of my preprocessing pipeline, I already > 'combined' the _LR and _RL fMRI data to produce a single 4D film > > Thanks in advance for your help. > -- > DED > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > > > ------------------------------ > > The materials in this message are private and may contain Protected > Healthcare Information or other information of a sensitive nature. If you > are not the intended recipient, be advised that any unauthorized use, > disclosure, copying or the taking of any action in reliance on the contents > of this information is strictly prohibited. If you have received this email > in error, please immediately notify the sender via telephone or return mail. > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > > > ------------------------------ > > The materials in this message are private and may contain Protected > Healthcare Information or other information of a sensitive nature. If you > are not the intended recipient, be advised that any unauthorized use, > disclosure, copying or the taking of any action in reliance on the contents > of this information is strictly prohibited. If you have received this email > in error, please immediately notify the sender via telephone or return mail. > -- DED _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
