hello,
i am having trouble in outputting an iresp from a 3dDeconvolve on a 1D timeseries. here's my command:
3dDeconvolve \
-input1D ic06_ctrl_ts_task.1D \
-TR_1D 2.35 \
-x1D run1.csplin \
-polort A \
-basis_normall 1 \
-num_stimts 2 \
-stim_times 1 tstim_wo.1D "CSPLIN(0,14.1,7)" -stim_label 1 wo \
-stim_times 2 tstim_nw.1D "CSPLIN(0,14.1,7)" -stim_label 2 nw \
-TR_times 1 \
-iresp 1 wo.irf \
-iresp 2 nw.irf \
-bucket statbuck
and here's what i get:
++ 3dDeconvolve: AFNI version=AFNI_2007_05_29_1644 (Jan 10 2008) [32-bit]
++ Authored by: B. Douglas Ward, et al.
++ 1D TR is 2.350 seconds
++ Imaging duration=1222.0 s; Automatic polort=9
++ -stim_times using TR=2.35 s for stimulus timing conversion
++ -stim_times using TR=1 s for any -iresp output datasets
++ -stim_times 1 using LOCAL times
++ -stim_times 2 using LOCAL times
++ Wrote matrix values to file run1.csplin.xmat.1D
++ Signal+Baseline matrix condition [X] (520x24): 3.29911 ++ VERY GOOD ++
++ Signal-only matrix condition [X] (520x14): 2.42428 ++ VERY GOOD ++
++ Baseline-only matrix condition [X] (520x10): 1.05734 ++ VERY GOOD ++
++ -polort-only matrix condition [X] (520x10): 1.05734 ++ VERY GOOD ++
++ Matrix inverse average error = 1.54722e-16 ++ VERY GOOD ++
++ Matrix setup time = 0.08 s
++ current memory allocated = 843122 bytes (about 843 thousand)
++ Calculations starting; elapsed time=0.113
Results for Voxel #0:
Baseline:
P_0 coef = -0.0202 P_0 t-st = -13.7541 p-value = 1.0764e-36
P_1 coef = 0.0018 P_1 t-st = 1.0963 p-value = 2.7348e-01
P_2 coef = -0.0096 P_2 t-st = -4.3766 p-value = 1.4706e-05
P_3 coef = -0.0047 P_3 t-st = -1.8652 p-value = 6.2747e-02
P_4 coef = -0.0281 P_4 t-st = -9.8898 p-value = 3.6029e-21
P_5 coef = 0.0123 P_5 t-st = 3.8872 p-value = 1.1517e-04
P_6 coef = 0.0152 P_6 t-st = 4.4116 p-value = 1.2593e-05
P_7 coef = 0.0177 P_7 t-st = 4.8379 p-value = 1.7540e-06
P_8 coef = 0.0151 P_8 t-st = 3.7782 p-value = 1.7715e-04
P_9 coef = -0.0031 P_9 t-st = -0.7527 p-value = 4.5199e-01
Stimulus: wo
h[ 0] coef = -0.0123 h[ 0] t-st = -1.8236 p-value = 6.8812e-02
h[ 1] coef = 0.0166 h[ 1] t-st = 4.2147 p-value = 2.9729e-05
h[ 2] coef = 0.0718 h[ 2] t-st = 19.0466 p-value = 4.1469e-61
h[ 3] coef = 0.0412 h[ 3] t-st = 10.8932 p-value = 6.3855e-25
h[ 4] coef = -0.0054 h[ 4] t-st = -1.4201 p-value = 1.5621e-01
h[ 5] coef = -0.0169 h[ 5] t-st = -4.3711 p-value = 1.5066e-05
h[ 6] coef = -0.0155 h[ 6] t-st = -3.4319 p-value = 6.4948e-04
R^2 = 0.5722 F[ 7,496] = 94.7784 p-value = 2.5967e-87
Stimulus: nw
h[ 0] coef = -0.0054 h[ 0] t-st = -0.8207 p-value = 4.1221e-01
h[ 1] coef = 0.0187 h[ 1] t-st = 4.8245 p-value = 1.8707e-06
h[ 2] coef = 0.0795 h[ 2] t-st = 20.4338 p-value = 8.7532e-68
h[ 3] coef = 0.0510 h[ 3] t-st = 13.1246 p-value = 5.4894e-34
h[ 4] coef = -0.0017 h[ 4] t-st = -0.4452 p-value = 6.5636e-01
h[ 5] coef = -0.0186 h[ 5] t-st = -4.6607 p-value = 4.0548e-06
h[ 6] coef = -0.0160 h[ 6] t-st = -3.4261 p-value = 6.6322e-04
R^2 = 0.6599 F[ 7,496] = 137.4698 p-value = 7.4019e-112
Full Model:
MSE = 0.0005
R^2 = 0.7457 F[14,496] = 103.9011 p-value = 2.0034e-137
++ Calculations finished; elapsed time=0.114
** FATAL ERROR: Can't open dataset '(null)'
I am particularly puzzled by:
1) the (fatal) error
2) the lack of the iresp 1D files wo.irf.1D and nw.irf.1D: shouldn't they be there?
3) the fact that statbucket.1D contains the estimated coefficients for the various regressors in a *single* concatenated column:
> cat statbuck.1D
-0.020153
0.001842
-0.009572
-0.004719
-0.028132
0.012292
0.015243
0.017698
0.015124
-0.003095
-0.012294
0.016635
0.071804
0.041240
-0.005420
-0.016933
-0.015536
-0.005426
0.018702
0.079484
0.050962
-0.001731
-0.018612
-0.015967
4) the above estimated coefficients have *not* been spline-interpolated onto the finer temporal grid of 1 sec, as specified by the CSPLIN model and the option "-TR_times 1"
From the help and the other documentation for 3dDeconvolve, I thought that I would get the 1D files for the iresp. Of course it's not a big deal to extract the values from either the terminal output or from the statbucket.1D file, but i am just checking whether this is the intended behavior of the program or I am missing something.
once more, many thanks for any help
giuseppe
