Dear experts,

I have recently started using the conn toolbox (13i in Matlab 2011b) and keep running into this message:

Warning: Returning NaN for out of range arguments 
> In spm_Fcdf at 108
  In conn_glm at 88
  In conn_process at 2590
  In conn_process at 26
  In conn at 2275 
Warning: Returning NaN for out of range arguments 
> In spm_Fcdf at 108
  In conn_glm at 88
  In conn_process at 2590
  In conn_process at 26
  In conn at 2275 
Warning: Returning NaN for out of range arguments 
> In spm_Fcdf at 108
  In conn_glm at 88
  In conn_process at 2590
  In conn_process at 26
  In conn at 2275 
Warning: Returning NaN for out of range arguments 
> In spm_Fcdf at 108
  In conn_glm at 88
  In conn_process at 2590
  In conn_process at 26
  In conn at 2275 

etc.

I'm getting this at the second-level results. My set-up is 1 subject, 2 conditions, and resting-state ROIs (rsREL) from the conn utils (which I hope coregistered to the structural and functional images in the preprocessing done in conn).

I'm not seeing any results, and some beta, F, and all p-unc and p-FDR are displayed as NaN. Hope you can help!

Many thanks,
Merina

Dear Merina,

The second-level analyses are meant for testing hypotheses about the population, they use random-effect analyses and they require data from multiple subjects to give meaningful statistics (with one subject the degrees of freedom is zero or below and no stats are computed). If you wish to obtain statistics about the functional connectivity from a single subject, you might want to try the previous version of the toolbox (conn 12p) which implemented fixed-effect analyses, and/or check this thread in the forum (http://www.nitrc.org/forum/forum.php?thread_id=2696&forum_id=1144) for other possible options.

Best
Alfonso
 
Originally posted by Merina Su:

Dear Experts,

I am also suffering from the same problem as merina suffered. I am using 2 subjects and 8 conditions, even then same "Warning messages" are displaying.
Can you please tell what's the minimum number of subjects that should be taken into account for 2nd level analysis ?

I am unable to specify "Between Condition Contrast" because of the warning messages.

Any help would be greatly appreciated.

Thanks a lot !!
Regards,
Sahil