Hi Doug,
I reduced the total time points, however, I found the eff and VRF value all became smaller as the time points decreased, and I still got some null events more than 7s. If I set the time points as 470s, I got most 5-7 null events, except the last 35s null event, and then its eff is 0.007, and VRFAvg is 1.7. Compared with 500 time points (eff 0.01, VFRAvg 2.58), 470s is bad design. From one paper, I learned that exponential distribution produced less variance compared with unity distribution, and maybe optseq2 likes exponential distribution more, so can I use the 470s design without the last 35s null events? Is the value of eff and VFRAvg OK?
...... ...... 819.0000 5 13.000 1.0000 diamond 832.0000 0 5.000 1.0000 NULL 837.0000 2 13.000 1.0000 star 850.0000 0 5.000 1.0000 NULL 855.0000 2 13.000 1.0000 star 868.0000 0 5.000 1.0000 NULL 873.0000 3 14.000 1.0000 star 887.0000 0 6.000 1.0000 NULL 893.0000 4 12.000 1.0000 diamond 905.0000 0 35.000 1.0000 NULL
On Tue, Oct 20, 2009 at 11:59 AM, Douglas N Greve < greve@nmr.mgh.harvard.edu> wrote:
I think this is partially a bug in optseq and partially a problem with your design. You have a total amount of stimulation time = 12*9+13*8+14*8+12*9+13*8+14*8 = 648 sec. Your run lasts 500*2 = 1000 sec, so the total amount of fixation time is 1000-648 = 352 sec. You have 8*4+9*2 = 50 stimuli, so the average time between stimuli (ie, average null) will be 352/50 = 7.04s. Yet you have constrained optseq to give you nulls between 5 and 7 sec. You need to relax this constraint or reduce the total number of time points.
doug
杨宏宇 wrote:
Hi list,
I am using optseq2 to design my fMRI study. These are my scripts:
optseq2 --ntp 500 --tr 2 --psdwin 0 40 1 --ev star 12 9 --ev star 13 8 --ev star 14 8 --ev diamond 12 9 --ev diamond 13 8 --ev diamond 14 8 --tnullmin 5 --tnullmax 7 --nkeep 3 --o ex.event12.13.14s --nsearch 5000
I have 2 events, which last for 12,13, and 14s. I found that the schedule documents I got had many null events more than 7s, however, I have set the maximam as 7s. In the sum logfile, I found that my script was recorded correctly, so could you please tell me the reason, and how to solve the problem?
Many thanks, Hongyu Division on Addiction, Psychiatry Department, UT Southwestern Medical Center
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
Hi Doug,
I figure out why this happened and how to solve the problem. Because My trials always lasted for 12,13,14s, and optseq2 believed that the baseline should last about 12-14s, which means null time should be as many as one task condition or the average of the task conditions. I set my trials as 5,6,7s, revised the time points, then I got very good sequence. I think now I need to do is to change the trial duration back and recode the time line.
On Thu, Oct 22, 2009 at 10:31 AM, 宏宇 yanghy720@gmail.com wrote:
Hi Doug,
I reduced the total time points, however, I found the eff and VRF value all became smaller as the time points decreased, and I still got some null events more than 7s. If I set the time points as 470s, I got most 5-7 null events, except the last 35s null event, and then its eff is 0.007, and VRFAvg is 1.7. Compared with 500 time points (eff 0.01, VFRAvg 2.58), 470s is bad design. From one paper, I learned that exponential distribution produced less variance compared with unity distribution, and maybe optseq2 likes exponential distribution more, so can I use the 470s design without the last 35s null events? Is the value of eff and VFRAvg OK?
...... ...... 819.0000 5 13.000 1.0000 diamond 832.0000 0 5.000 1.0000 NULL 837.0000 2 13.000 1.0000 star 850.0000 0 5.000 1.0000 NULL 855.0000 2 13.000 1.0000 star 868.0000 0 5.000 1.0000 NULL 873.0000 3 14.000 1.0000 star 887.0000 0 6.000 1.0000 NULL 893.0000 4 12.000 1.0000 diamond 905.0000 0 35.000 1.0000 NULL
On Tue, Oct 20, 2009 at 11:59 AM, Douglas N Greve < greve@nmr.mgh.harvard.edu> wrote:
I think this is partially a bug in optseq and partially a problem with your design. You have a total amount of stimulation time = 12*9+13*8+14*8+12*9+13*8+14*8 = 648 sec. Your run lasts 500*2 = 1000 sec, so the total amount of fixation time is 1000-648 = 352 sec. You have 8*4+9*2 = 50 stimuli, so the average time between stimuli (ie, average null) will be 352/50 = 7.04s. Yet you have constrained optseq to give you nulls between 5 and 7 sec. You need to relax this constraint or reduce the total number of time points.
doug
杨宏宇 wrote:
Hi list,
I am using optseq2 to design my fMRI study. These are my scripts:
optseq2 --ntp 500 --tr 2 --psdwin 0 40 1 --ev star 12 9 --ev star 13 8 --ev star 14 8 --ev diamond 12 9 --ev diamond 13 8 --ev diamond 14 8 --tnullmin 5 --tnullmax 7 --nkeep 3 --o ex.event12.13.14s --nsearch 5000
I have 2 events, which last for 12,13, and 14s. I found that the schedule documents I got had many null events more than 7s, however, I have set the maximam as 7s. In the sum logfile, I found that my script was recorded correctly, so could you please tell me the reason, and how to solve the problem?
Many thanks, Hongyu Division on Addiction, Psychiatry Department, UT Southwestern Medical Center
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
I think your constraint of 5-7 null time is too tight for optseq, and you may still have too much time in your run to make 5-7 work. It's hard to say whether the VRFAvg is good enough or not. You are using sub-TR estimation during the optimization. If you are going to assume a shape to the HRF, then it's probably fine. If you are going to use an FIR, then you'd need about 7 runs or so.
doug
杨宏宇 wrote:
Hi Doug,
I reduced the total time points, however, I found the eff and VRF value all became smaller as the time points decreased, and I still got some null events more than 7s. If I set the time points as 470s, I got most 5-7 null events, except the last 35s null event, and then its eff is 0.007, and VRFAvg is 1.7. Compared with 500 time points (eff 0.01, VFRAvg 2.58), 470s is bad design. From one paper, I learned that exponential distribution produced less variance compared with unity distribution, and maybe optseq2 likes exponential distribution more, so can I use the 470s design without the last 35s null events? Is the value of eff and VFRAvg OK?
...... ...... 819.0000 5 13.000 1.0000 diamond 832.0000 0 5.000 1.0000 NULL 837.0000 2 13.000 1.0000 star 850.0000 0 5.000 1.0000 NULL 855.0000 2 13.000 1.0000 star 868.0000 0 5.000 1.0000 NULL 873.0000 3 14.000 1.0000 star 887.0000 0 6.000 1.0000 NULL 893.0000 4 12.000 1.0000 diamond 905.0000 0 35.000 1.0000 NULL
On Tue, Oct 20, 2009 at 11:59 AM, Douglas N Greve <greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu>> wrote: I think this is partially a bug in optseq and partially a problem with your design. You have a total amount of stimulation time = 12*9+13*8+14*8+12*9+13*8+14*8 = 648 sec. Your run lasts 500*2 = 1000 sec, so the total amount of fixation time is 1000-648 = 352 sec. You have 8*4+9*2 = 50 stimuli, so the average time between stimuli (ie, average null) will be 352/50 = 7.04s. Yet you have constrained optseq to give you nulls between 5 and 7 sec. You need to relax this constraint or reduce the total number of time points. doug 杨宏宇 wrote: Hi list, I am using optseq2 to design my fMRI study. These are my scripts: optseq2 --ntp 500 --tr 2 --psdwin 0 40 1 --ev star 12 9 --ev star 13 8 --ev star 14 8 --ev diamond 12 9 --ev diamond 13 8 --ev diamond 14 8 --tnullmin 5 --tnullmax 7 --nkeep 3 --o ex.event12.13.14s --nsearch 5000 I have 2 events, which last for 12,13, and 14s. I found that the schedule documents I got had many null events more than 7s, however, I have set the maximam as 7s. In the sum logfile, I found that my script was recorded correctly, so could you please tell me the reason, and how to solve the problem? Many thanks, Hongyu Division on Addiction, Psychiatry Department, UT Southwestern Medical Center ------------------------------------------------------------------------ _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu> Phone Number: 617-724-2358 Fax: 617-726-7422 In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
杨宏宇