[FieldTrip] automatic artifact rejection

[De Meyer, Kris]

Hi all,

Looking at the documentation page for automatic artifact rejection<http://www.fieldtriptoolbox.org/tutorial/automatic_artifact_rejection/>, it contains a detail that comes across to me as odd.

At end of Section III. Z-transforming the filtered data and averaging it over channels, it says that the averaging is done by summing the z-transformed values for each time-point over all channels, and then dividing them by the square root of the number of channels.

I don't understand what the square root is doing in that formula. In my case I have 14 channels I want to average over for the artifact rejection, and if all of them had a z-value of 1, and my threshold was 3 or 3.5 (sensible values, given they are z scores) then my final averaged value is higher than my threshold only because of the summation!

To me this doesn't make much sense because it means that, rather than being able to set a constant threshold regardless of how many channels one averages over, it now means we need to use a different threshold for different numbers of channels.

What am I missing?

Best,

Kris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20200411/85534394/attachment.htm>