Hi guys,
I m having a problem with matlab and i need your help.
I have two output signals (let's say B and C) which are recordings of the
same physiological process, at symmetrical parts of the body, and the input
signal A.
To see if there is noise in my recordings, i m subtracting B and C , and
calculating the magnitude squared coherence of DV=C-B to A. As expected the
coherence is very near zero.
For practice, I was asked to do a least squares linear regression passing
through the origin, so as to transform B to be similar to C B=slope*C (i did
this in matlab with the command : slope=A\B; , NB=slope*C).
Theoretically, if we subtract the transformed B signal (NB) from C (DV2=C-NB),
and take DV2's magnitude squared coherence to A, the coherence should be
lower than before.
However, what i m getting in matlab is that despite DV2 being smaller than DV,
coherence after the transformation increases.
Am i doing something wrong? I would really appreciate your input on this.
Thanx in advance