Hi,
You cannot enter an empty or constant regressor for your parametric modulation. You’d have to remove those subjects that performed at floor/ceiling.
HTH
Best,
Leyla
Sent from my iPhone
> On 19 Aug 2022, at 17:20, Victor Parey <[log in to unmask]> wrote:
>
>
> Hello SPM listserv members,
>
> I am running into an issue with my parametric modulator which I cannot seem to resolve. For the first level analysis, I am hoping to include a parametric modulator to examine how a behavioral measure modulates trial-by-trial brain activity. However, for some (but not all) subjects, I received the error stating that the Contrast Manager failed due to error in contrast specification. When I examined the behavioral measure I used as the parametric modulator, I noticed that the ones that produced error were ones whose behavioral measure was very "flat" or even constant (i.e., little or no variability). This is due to the fact that these subjects either performed at floor or ceiling levels. Could it be that SPM cannot handle a constant parametric modulator?
>
> There were however some exceptions where subjects with little variability in the behavioral measure did not trigger the contrast error.
>
> Would anyone with knowledge on parametric modulators be inclined to give their opinion on the source of error I'm getting? Is there a way to get around this so I can build the contrast for every subject?
>
> Thanks in advance!
>
> Victor
>
>
|