Hi:
To investigate the influence of the overt speech during fMRI, We adopted event-related design and Chinese character naming task. Two different presentation approach were used, irregular interval between stimulus(average of 7s,the longest ones were 21.5s, the shortest ones were 1.5s); regular interval between stimulus(14s).
To analyse the functional data, we simulated the head motion curve (MC1.1D for positive headmotion & MC2.1D for negtive headmotion) by Waver command line, and assess head motion parameters (vol.1D) automaticly by 3dvolreg command line.Then we deconvolved the functional data and corrected the head motion by making the simulated head motion curve and assessed motion parameters as base line in 3dDeconvolve command line.
Command line showing as follow:
3dDeconvolve -input dataset+orig -nfirst 9 -nlast 148 -polort 1 -rmsmin 1.0 -progress 1000 -num_stimts 9
-stim_file 1 task-related.1D -stim_label 1 task -stim_minlag 1 0
-stim_maxlag 1 6
-stim_file 2 'vol.1D[0]' -stim_label 2 roll -stim_base 2
-stim_file 3 'vol.1D[1]' -stim_label 3 pitch -stim_base 3
-stim_file 4 'vol.1D[2]' -stim_label 4 yaw -stim_base 4
-stim_file 5 'vol.1D[3]' -stim_label 5 dl -stim_base 5
-stim_file 6 'vol.1D[4]' -stim_label 6 ds -stim_base 6
-stim_file 7 'vol.1D[5]' -stim_label 7 dp -stim_base 7
-stim_file 8 MC1.1D -stim_label 8 MC1 -stim_minlag 8 0 -stim_maxlag 8 0
-stim_base 8
-stim_file 9 MC2.1D -stim_label 9 MC2 -stim_minlag 9 0 -stim_maxlag 9 0
-stim_base 9
-iresp 1 task.irf
-tshift
-fitts task-fitts
-fout
-bucket task.buck
BUT, to our surprise, the irf curve showed that the headmotion signal change rate was 25~50 times higher than that of task-related BOLD signal.
When we used the assessed headmotion parameters only(get rid of the simulated headmotion curve)in the 3dDeconvolve command line, the irf curve showed that the headmotion signal change rate was a half of the BOLD signal chage rate or even less.
Why? and how to account for this condition?
THANKS a lot!