How does Tentzero force the fitted curve to be continuous?

Hi,

I used tentzero for a block design experiment (TR=1.5s). The first time I choose the block length 12s and tentzero(0,12,9). And second time I make the time parameter longer and used tentzero(0,24,17), with a 12s added (that’s actually rest time between blocks).

Interestingly I found the fit result for the 24s result curve is larger than the 12s one at every fitted value, which surprises me since I was assuming the first 9 points value shouldn’t really change, if tent function only average the HRF responses.

On the 3dDeconvolve help document it said those zero functions will force the fitted curve to be continuous, and to me this seems to be the reason why the two curve values are so different. But I can’t find any more information how tentzero actually force the fitted curve to be continuous. Would you please provide some information on that?

Thanks,
Lingyan

Lingyan,

On the 3dDeconvolve help document it said those zero functions will force the fitted curve to be continuous

TENTzero places a zero value at the fist tent (usually the onset of each trial) and the last tent as well. In contrast, TENT does not have such a constraint.

Interestingly I found the fit result for the 24s result curve is larger than the 12s one at every fitted value, which surprises me since
I was assuming the first 9 points value shouldn’t really change, if tent function only average the HRF responses.

It is hard to guess what is going on. You can add option -fitts to both tentzero(0,12,9) and tentzero(0,24,17), and visualize the results. See if you can gain more insights about your confusion through the model diagnosis.

Hi Gang,

Thanks a lot! I’ll do the check.

Meanwhile, is there any document saying how exactly does the tentzero works, like it’s original code? I can’t get it out with tentzero -help.

Lingyan

is there any document saying how exactly does the tentzero works

You can find the following in “3dDeconvolve -help”:

You can also use ‘TENTzero’ and ‘CSPLINzero’, which means to eliminate the first and last basis functions from each set. The effect of these omissions is to force the deconvolved HRF to be zero at t=b and t=c (to start and and end at zero response).