search

LIMO-EEG-Toolbox / limo_tools

Hierarchical Linear Modelling for MEEG data
https://limo-eeg-toolbox.github.io/limo_meeg/
Other
58 stars 28 forks source link

[BUG] regression Yr.mat inverted subjects #153

Closed sarbula closed 1 year ago

sarbula commented 1 year ago

After doing the regression analysis, the Yr.mat contains the data, but the order of the subjects is inverted (the first subject becomes the last). This seems to affect the results since the covariate is not inverted.

CPernet commented 1 year ago

thanks for that -- tried to optimize speed but messed up ordering :-( should be solved now

sarbula commented 1 year ago

Just to check, this bug was present since the early versions of LIMO? We have some results from the analyses we ran few years ago, but now with this fix they completely changed.. Thanks

CPernet commented 1 year ago

mmh looking at the history looks like the bug is there since 2020 ; only affects regression but yeah pretty big ! now if the subject order is reversed, but not the regressor isn't it surprising you got significant results ?

sarbula commented 1 year ago

we got significant results both then and now, but in different time windows and electrodes... we figured out something was wrong when we started plotting directly betas... thanks