How to correctly doing baseline correction? about neurokit.py

[DominiqueMakowski]

For Phasic EDA I'm not sure it actually makes sense to apply baseline correction. For HRV, I would rather correct for a window immediately preceeding your event. Then, the type of correction depends mostly on your hypotheses.

For example, I presented emotionally negative pictures to participants. Based on the litterature, I expected a higher heart rate deceleration upon stimulus presentation (let's say, for 5 s). So I took the minimum of the heart rate in these 5 s and I subtracted that from the mean of a [-1, 0] window. To quantify the deceleration amplitude, so to speak.

But again, I'm not sure there are any gold standarts, you should follow the litterature unless you've got a good (ie, justified) reason for doing something else 😄

[Eichhof]

Thank you for your answer. What do you mean by the [-1,0] window? So you had a 5s window beginning at the stimulus onset and over this window you searched the minimum and subtracted this minimum from the mean over this window?

[DominiqueMakowski]

@Eichhof I substracted the minimum of the 0-5s window of the mean of the 1-second-long window preceeding stimulus onset.

[Eichhof]

Now I've got it. ;) Did you take the mean of the heart rate over 1 seconds (I don't know how that mean could be calculated over such a small period)?

[DominiqueMakowski]

NeuroKit computes by default a third order spline interpolation of the RR intervals to get a continuous signal of heart rate. You could even take the value of this heart rate signal upon stimulus onset. But otherwise, you can compute your own heart rate using R peaks (in that case you'll need indeed a larger window)

[Eichhof]

Makes sense. :)