Why is nilearn's first-level output discrete?

Hi, I am currently fitting heriarchical models to the condition dependent beta estimates (or rather the z-scores) of my ROIs. I was supprised when I saw that these values are actually discrete. As far as I remember the beta estimates in SPM are on a continous scale.
Can someone give some insight here?
Best, Marius

Dear maalaria,
Indeed, the values should not be discrete —or there is something pathological with the model or the data. If you can provide a snippet of code (or even better, code+data), we can try and figure out what’s going on.

Thanks Bertrand for you quick reply. It’s funny… I just compiled a minimal working example and the outcome is that values are continous now… I will try to figure out what’s going on and come back to you.
Best, Marius

Hey, ashamed I must admitt that I have found a code snippet where I round the values to integers manually…:rofl: I vaguely rember now that I once tried something requiring integers and apparantly have forgotten to put a warning into my code.
Have nice weekend,