Hi Afni Experts
I learn how to use the command "3dDeconvolve" for "1.4.2 Estimation of the System Impulse Response Function" based on the instrument "Deconvolution Analysis of FMRI Time Series Data". When I implement the Example 1.4.2.1, the results I got are different from the results in the intrument. I confuse and want to know the resean.
h(t) = {0 5 10 5 2}
f(t) = {0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0}
z(t) = 100 + t + h(t)*f(t)
= {100 101 102 108 114 110 108 107 108 114 120 116 114 113 114 120 126 122 120 119}
I used the command (same as the instrument)
3dDeconvolve -input1D z.1D -num_stimts 1 -stim_file 1 f.1D -stim_label 1 ''f'' -stim_maxlag 1 4
But the results are:
++ ==========================================================
++ ----- Signal+Baseline matrix condition [X] (16x7): 4.80619 ++ VERY GOOD ++
++ ----- Signal-only matrix condition [X] (16x5): 1 ++ VERY GOOD ++
++ ----- Baseline-only matrix condition [X] (16x2): 1.52392 ++ VERY GOOD ++
++ ----- polort-only matrix condition [X] (16x2): 1.52392 ++ VERY GOOD ++
++ +++++ Matrix inverse average error = 4.53943e-16 ++ VERY GOOD ++
++ Matrix setup time = 0.01 s
++ current memory malloc-ated = 17260 bytes (about 17 thousand)
++ Calculations starting; elapsed time=0.130
Results for Voxel #0:
Baseline:
P_0 coef = 109.5000 P_0 t-st = 1000.0000 p-value = 0.0000e+00
P_1 coef = 9.5000 P_1 t-st = 1000.0000 p-value = 0.0000e+00
Stimulus: f
h[ 0] coef = 0.0000 h[ 0] t-st = 0.0000 p-value = 1.0000e+00
h[ 1] coef = 5.0000 h[ 1] t-st = 1000.0000 p-value = 0.0000e+00
h[ 2] coef = 10.0000 h[ 2] t-st = 1000.0000 p-value = 0.0000e+00
h[ 3] coef = 5.0000 h[ 3] t-st = 1000.0000 p-value = 0.0000e+00
h[ 4] coef = 2.0000 h[ 4] t-st = 1000.0000 p-value = 0.0000e+00
R^2 = 1.0000 F[ 5, 9] = 1000.0000 p-value = 0.0000e+00
Full Model:
MSE = 0.0000
R^2 = 1.0000 F[ 5, 9] = 1000.0000 p-value = 0.0000e+00
++ Calculations finished; elapsed time=0.130
++ #Flops=5073 Average Dot Product=5.9661
++ ==========================================================
P_0 coef = 109.5 and P_1 coef = 9.5 are different from the results shown in instrument t^0 coef = 100.0000 and t^1 coef = 1.0000.
Would you like to help me to find out the problem? Why is there difference?
Thank you very much!
Xin