Possible problem in bandpass filtering

[farnazdelavari]

I have been trying to generate CVR maps for simple resting state fMRI. The filtered average signal which is an output from the toolbox (petco2hrf) looked a bit strange to me. When checking the spectral density, it seems to us that the filtering might have failed.

Expected Behavior

I would expect such time-course:

and a high density within the filtered 0.02-0.04 range of the filtered time course.

Actual Behavior

The toolbox time-course is as below

with the spectral density that is different :

Steps to Reproduce the Problem

If the tc is the time-course filtered and stored in "avg_func_masked_petco2hrf.1D" and TR is the sampling rate of the MRI in seconds.
then this code in Matlab should generate the spectral density as stated.
N=lenghth(tc);
fs=1/TR;
xdft=fft(tc);
xdft = xdft(1:N/2+1);
psdx = (1/(fsN)) * abs(xdft).^2;
psdx(2:end-1) = 2psdx(2:end-1);
freq = 0:fs/N:fs/2;
figure
plot(freq,pow2db(psdx),'r')

Specifications

- phys2cvr version: 0.16.0


- resting-state fMRI, TR=2.4 s, number of volumes = 200, band used = 0.02-0.04.

Possible solution

[smoia]

@farnazdelavari can you check the new version against your matlab code please?