Christoph Christmann ha scritto:
> hi,
>
> reanalizing old data I noticed that there is apparently a difference
> to SPM99 in the calculated degrees of freedom in a first level
> analysis. I have a data set of 71 time points with a single regressor
> (simple block design). No matter what preprocessing steps I used
> (smoothing etc.), I get DF=69 in SPM99 but DF=66 in SPM5 while using
> the same parameter configuration except the fact that with SPM99 I
> applied a hrf-based low pass filter.
>
> I have no idea why SPM5 decreases the DF and would be happy if anyone
> had an explanation.
>
> TIA
>
> christoph
>
>
|