[Eeglablist] PREP pipeline & bad channel interpolation

[Andrew Corcoran]

Hi there

I've been experimenting with Nima's PREP pipeline recently and am trying to
get clear on exactly which channels are interpolated in the final output.

>From reading the paper it sounds like putative bad/outlier channels are
first identified relative to a first approximation of the robust reference
and subsequently interpolated. The 'true' robust reference is then
calculated. There then follows a final round of channel evaluation based on
thresholds set according to the robust reference estimate.

It seems to me that we should only want to interpolate (or exclude)
channels that fail the final set of threshold tests. However, it's not
clear to me whether those channels that were interpolated in the initial
phase of testing remain interpolated, or are added back into the data to be
evaluated from scratch in relation to the final reference signal (this is
the impression I got from the paper, but looking at the output summary
reports and EEGlab structures I'm not confident). I'm also not sure whether
additional channels that fail the final test phase are themselves
interpolated in the output data set.

Can anyone clarify this for me ?

Regards
Andrew
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20180726/94a72e34/attachment.html>