Dear Bart,
Try doing highpass and lowpass separately instead of bandpass.
Bandpass is not a good thing to do for such a wide band although it is
used in SPM tutorial.
Vladimir
On 9 Apr 2010, at 17:50, Bart Machilsen wrote:
> Dear all,
>
> when using spm_eeg_filter to bandpass-filter continuous EEG data,
> the filtered data values are all NaN.
>
> Details:
> SPM8
> preprocessing: raw data re-referenced to the average of the
> mastoids, no further preprocessing
> filter.type = 'butterworth';
> filter.band = 'bandpass';
> filter.PHz = [0.1 30];
>
> In spm_filtfilt at line 83, my filtered data "y" contain Inf values.
> This causes NaN output on line 84.
>
> Is something wrong with my input data? How do I avoid the Inf values
> after filtering?
>
> Thanking you in advance,
>
> Bart Machilsen
|