Censoring error in 3dTproject

AFNI Message Board
1

Dear AFNI experts,

I am using 3dTproject to do nuisance regression to my data. We are using the -censor option to mitigate the effect of outliers time points. Sometimes, I have got the following error:

ERROR_message "run #%d has only %d points after censoring need at least %d;

After a few tests I found that the program shows this error when the censor file have more than 25.9 % of time points to censor. Does the program has a established percent limit of time point to censor? In my case I have this error every time the outliers exceeds 25.9%.
I was looking at the code in GitHub just to understand what is happening and I found that there is a conditional which counts un-censored points in each run.


  if( nbl > 1 ){
     int aa,bb, nnk , nerr=0 ;
STATUS("count un-censored points in each run") ;
     for( tt=0 ; tt < nbl ; tt++ ){
       aa = bla[tt] ; bb = blb[tt] ;
       for( nnk=0,jj=aa ; jj <= bb ; jj++ ) if( tp->censar[jj] != 0.0f ) nnk++ ;
       if( nnk < MIN_RUN ){
         ERROR_message(
           "run #%d has only %d points after censoring (need at least %d)",
           tt+1,nnk,MIN_RUN) ;
         nerr++ ;
       }
     }
     if( nerr > 0 ) ERROR_exit("Cannot continue with such over-censoring!") ;
   }

My question is: Is 25.9 the maximum percentage allowed to censor time points? If so, is there some way to increase it to be less restrictive?

Kind regards,
Karel

2

Hi-

Can you please provide a bit more details:

  • what is your 3dTproject command? or are you using this via afni_proc.py?
  • how many time points are you starting with?
  • what is the actual output message in the terminal (you are only outputting code blocks of variables, not the output numbers).

MIN_RUN is set to 9 in the code, I believe-- are you getting down to less than 9 degrees of freedom? How many is your number of “points after censoring”? (You don’t put the actual output, just the code involved).

–pt