thanks a lot rob

my reason against average reference is that when doing connectivity analyses based on model prediction, if the target channel already contains by construction info from the other channels then I will introduce artificial influences.

I see your point, I will try to recover the original reference from the recordings.

also, I compared the results obtained with FASTER with the manual ICA-based rejection with the "good" reference, and they look very much the same, so I could possibly try like that

On Mon, Feb 25, 2013 at 5:35 PM, Robert Whelan <> wrote:
Hi Daniele,

It's not a good idea to skip the average reference step -- this is
because the pattern of artifacts will somewhat become a function of
the reference channel. This is why FASTER uses the average for
artifact rejection -- and EEGLAB will force a re-reference to the
average anyway for the ICA step.

However, you can always re-reference to any electrode(s) in EEGLAB
once pre-processing is over. So, my advice would be to let the
artifact rejection run with the average reference and then use EEGLAB
to re-reference to your desired reference. Or do you have a strong
motivation for not using the average ref?

All the best,

On Sun, Feb 24, 2013 at 10:33 AM, Daniele Marinazzo
<> wrote:
> Hi,
> I would like to skip the averaged reference step, but I don't have the
> original reference channel in my dataset. so, if possible, I would like to
> completely skip the reference part, leaving the original reference of the
> input dataset. is it possible to do this?
> thanks
> daniele
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> _______________________________________________
> Faster-eeg-list mailing list