Dear SPMers,

 

 I'm trying to do GLM with one more regressor column of button press to rull out any effect of motor activity during the task.

 But I realize that SPM treats the regressor that is convolved  with less number of button press onsets is not unquely specified. 

 For instance, subjects are supposed to press the button 6 times per each of 4 runs, but in some cases some subjects happened to press it more than 6 in any runs which ends up giving 

 e.g., 6,,6,7,6 onset values per each run respectively.  

 

 After specifying, SPM treats the 1,2, and 4 run has the column of not unquely specificed beta value in fact they are not.

 As a result, SPM shows three long black line in the design matrix, which makes me concerened of wrongly modeling the study.

 I'm wondering if there is any way to trick SPM so that it thinks those button columns are correctly specified with the given number of onset times. 

 If that message of "not uniquely" is harmless, I'm fine with leaving them as it is. 

 Any advice would be greatly appreciated. 

 

 Thanks in advance, 

 YSL