[Neuroimaging] one sample t-test in nilearn: is there an automatic z-transform involved?

[Ze Wang]

My images from each individual have non-negative values across the 
brain, I would assume that t-value (or the z-value) for the one-sample 
t-test will be all greater than 0.  But I found negative values.  Does 
anyone know is  there a z-transform in nilearn before doing the 
one-sample t?

my code is simple:
design_matrix=pd.DataFrame(np.hstack( (np.ones( (cova.shape[0],1)), 
cova) ), columns=colname )
             model = SecondLevelModel(smoothing_fwhm=8.0, 
mask_img=brainmask)
             print(design_matrix)
             #print(imgs)
             model.fit(imgs.tolist(), design_matrix=design_matrix)
             omaps = model.compute_contrast('intercept', 
output_type='all')

The contrast for the intercept was taken as the one-sample t results. 
and the resulting z-map have negative values.

Thanks
Ze
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.python.org/pipermail/neuroimaging/attachments/20210826/32b88e05/attachment.html>