[Eeglablist] Average reference questions.

[Robert Lawson]

How about wieghted average as in Lemos and Fisch or Hjorth laplacian 
reference as far as reducing this error?

Robert Oostenveld wrote:

> On 26 Jul 2004, at 21:27, Arnaud Delorme wrote:
> 
>>> 2. Which advantages and disadvantages average references method has?
>>>
>> I am not an expert on this topic. I would say that the advantage of 
>> using average reference is that you recover the signal from the 
>> reference channel (thus you have one extra channel even if it is 
>> linearly dependent on the activity of the other channels). Also, 
>> average reference is necessary for dipole fitting.
> 
> 
> In principle average referencing is not required for dipole fitting, but 
> the DIPFIT implementation in EEGLAB does require average reference data. 
> The advantage for average referencing in dipole fitting, as compared to 
> using a common reference, is that it does not put excessive weight on an 
> arbitrary single channel (i.e. the reference channel). Let me try to 
> explain:
> 
> During dipole fitting, the sum-squared error over all channels is 
> minimized. If the forward model is inaccurate on the reference channel, 
> the model error in the reference channel spreads over all other channels 
> (since the potential on the other channels is computed with the model as 
> being the difference between the potential on that channel and the 
> specified reference channel). Using a common reference results in an 
> strong tendancy to minimize the (model) error on the reference channel, 
> since the potential on the reference channel is reflected in all other 
> channels.
> 
> Furthermore, since the reference channel is commonly on mastoid or ear 
> electrodes, the volume conduction model error in them is relatively 
> large (the area around the ears is poorly modelled using a spherical 
> model), hence the model error in the reference channel tends to be 
> arelatively large if using the ear or mastoid. By taking the average 
> over all channels as reference, the error in the model is assumed to 
> average out in the reference, thereby minimizing its influence in the 
> sum-squared error that is used as the goal function to be minimized 
> during dipole fitting.
> 
> best regards,
> Robert
> 
> ----------------------------------------------------------------------
> Robert Oostenveld, PhD
> Center for Sensory-Motor Interaction (SMI)
> Aalborg University, Denmark
> 
> and
> 
> F.C. Donders Centre for Cognitive Neuroimaging
> University Nijmegen
> P.O. Box 9101
> NL-6500 AH Nijmegen
> The Netherlands
> 
> Tel: +31 (0)24 3619695
> Fax: +31 (0)24 3610989
> 
> 
> ----------------------------------------------------------------------
> Robert Oostenveld, PhD
> Center for Sensory-Motor Interaction (SMI)
> Aalborg University, Denmark
> 
> and
> 
> F.C. Donders Centre for Cognitive Neuroimaging
> University Nijmegen
> P.O. Box 9101
> NL-6500 AH Nijmegen
> The Netherlands
> 
> Tel: +31 (0)24 3619695
> Fax: +31 (0)24 3610989
> 
> _______________________________________________
> Eeglablist mailing list Eeglablist at sccn.ucsd.edu
> Eeglablist page: http://www.sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsub at sccn.ucsd.edu
> To switch to digest mode, send an empty email to 
> eeglablist-digest at sccn.ucsd.edu
> 
>