Open czawora opened 6 years ago
Interesting question. Can you say a bit more about your recording setup? What type of electrodes are you using (tetrodes, microwire bundle, MEA,...?) and how distant are your electrode sites from each other? Do you already do differential recording (i.e. do you subtract off a local reference?)
We are using Utah arrays (http://blackrockmicro.com/electrode-types/utah-array/). It's a 96 channel grid array with 400 μm spacing between electrodes. All the electrodes are referenced to a low impedance wire ~1 cm from the array. Up to this point, we have been doing common average referencing as a preprocessing step in addition to filtering.
Between the differential recording and the CAR, I'd be surprised if there is much more common-mode noise you could expect whitening to reject. Pinging @magland for his thoughts on single-channel sorting.
If I intend to sort all my channels independently, does it still make sense to perform the whitening step on all the channels?
I'd say... maybe. Is there correlation between the channels that is not related to the signal of interest? If so, then probably it does make sense to remove that.
If I intend to sort all my channels independently, does it still make sense to perform the whitening step on all the channels?