[Eeglablist] reference for optimum number of channels

[Tarik S Bel-Bahar]

Hi Fran,  Quick notes below that may be useful for orienting on this
topic...

***Not sure if there is a perfect  data-driven publication about these
topics in last few years (e.g., effect of channel density on ICA quality
and dipfit/dipole estimation, perhaps its a publication opportunity!).

*****Nunez and Srinivasan's book and articles deal with these topics, as a
core citation, though I don't think they used ICA. The book is available
online in various formats, and considered a true watermark in the field.
There is plenty of recent work, fyi citations further below.

***Generally speaking, yes a rule of thumb is ~50 and up for valid robust
ICA, though one can also get good ICs from 19 and 32 channel data,
However, 64+ chan data is certainly considered more valid for any kind of
estiamtion of total brain dynamics, and for better ICA and source
localization. Note however, classic ICA-EEG articles in Science by the
EEGLAB team, and many others, use ~32 channels for good ICA.

***It is often thought that "increasingly better coverage of the whole
head/sphere" should lead to the "increasingly more accurate ICA and source
solutions". However, researchers have noted there is maximum necessary
density of channels (e.g., what's the difference in information gained from
having a sensor every 2 cm, every 1 cm, or every 1 mm. Technically, we know
little about ultra-dense EEG at this time, but existing work suggests that
more refined information may indeed be gained when going ultra-hd-EEG.

***Caveat: total head coverage doesn not automatically equal perfectly
valid sources, and there's problems with covering some parts of the head
(ie, face, neck, ventral electrodes, etc.. which have noise and other
issues). Also, ventral and deep sources are still considered quite
problematic for EEG source estimates, in general.

***A general caveat for dipfit dipoles: they may be most relevant and valid
for ERP rather than continuous data, though both ERP and resting data have
been extensively published on in terms of dipfit solutions.

***Even in 128+ chan studies, there are usually only ~10 to 20 ICs that are
clearly neural/valid. This is well-validated in the published ICA-EEG
literature, though not necessarily systematically examined.

***other features, such as data quality and length are also important, and
impact ICA quality and validity heavily.

******Some pubs that may be of interest:

Petrov, Y., Nador, J., Hughes, C., Tran, S., Yavuzcetin, O., & Sridhar, S.
(2014). Ultra-dense EEG sampling results in two-fold increase of functional
brain information. *Neuroimage*, *90*, 140-145.
Robinson, A. K., Venkatesh, P., Boring, M. J., Tarr, M. J., Grover, P., &
Behrmann, M. (2017). Very high density EEG elucidates spatiotemporal
aspects of early visual processing. *Scientific reports*, *7*(1), 1-11.
Pal, P., Theisen, D. L., Datko, M., van Lutterveld, R., Roy, A., Ruf, A., &
Brewer, J. A. (2019). From research to clinic: A sensor reduction method
for high-density EEG neurofeedback systems. *Clinical Neurophysiology*,
*130*(3), 352-358.
Liu, Q., Ganzetti, M., Wenderoth, N., & Mantini, D. (2018). Detecting
large-scale brain networks using EEG: impact of electrode density, head
modeling and source localization. *Frontiers in neuroinformatics*, *12*, 4.
Iivanainen, J., Mäkinen, A., Zetter, R., Stenroos, M., Ilmoniemi, R. J., &
Parkkonen, L. (2019). Sampling theory for spatial field sensing:
Application to electro-and magnetoencephalography. *arXiv preprint
arXiv:1912.05401*.
Hu, S., Lai, Y., Valdes-Sosa, P. A., Bringas-Vega, M. L., & Yao, D. (2018).
How do reference montage and electrodes setup affect the measured scalp EEG
potentials?. *Journal of neural engineering*, *15*(2), 026013.
Dattola, S., La Foresta, F., Bonanno, L., De Salvo, S., Mammone, N.,
Marino, S., & Morabito, F. C. (2020). Effect of Sensor Density on eLORETA
Source Localization Accuracy. In *Neural Approaches to Dynamics of Signal
Exchanges* (pp. 403-414). Springer, Singapore.
Justesen, A. B., Foged, M. T., Fabricius, M., Skaarup, C., Hamrouni, N.,
Martens, T., ... & Beniczky, S. (2019). Diagnostic yield of high-density
versus low-density EEG: The effect of spatial sampling, timing and duration
of recording. *Clinical Neurophysiology*, *130*(11), 2060-2064.
Song, J., Davey, C., Poulsen, C., Luu, P., Turovets, S., Anderson, E., ...
& Tucker, D. (2015). EEG source localization: sensor density and head
surface coverage. *Journal of neuroscience methods*, *256*, 9-21.
Greischar, L. L., Burghy, C. A., van Reekum, C. M., Jackson, D. C.,
Pizzagalli, D. A., Mueller, C., & Davidson, R. J. (2004). Effects of
electrode density and electrolyte spreading in dense array
electroencephalographic recording. *Clinical Neurophysiology*, *115*(3),
710-720.




****************************************