[pymvpa] Paper on searchlight radius?

[Nick Oosterhof]

On Mon, 8 Oct 2018 at 14:10, Alyson Saenz <neuro.alyson at gmail.com> wrote:

> Does anyone know a paper in which someone has explored the effect of
> different searchlight radius?
>
> I know that a bigger radius increases performance but also processing
> time, and that at some point you will get bad performances, but, has anyone
> systematically explored this?
>

In addition to Yarik's and Jo's great suggestions and observations, let me
add:

- In general, the optimal radius would depend on the spatial extent and
distributed nature of the effect of interest. Different neural mechanisms
may have a different spatial distribution / scale.

- the original paper by Kriegeskorte et al (2006): "For the
information-based analysis, we chose a searchlight radius of 4 mm, because
this radius yielded the best detection performance in the simulation". In
my experience, such small radii are not common these days; most studies
report larger radii, typically 6-12 mm .

- with shameless self-promotion, there is a paper [1] possibly useful for
surface-based approaches. Note that the analyses focused in M1 and S1
regions with discrimination of different digits (fingers); your mileage may
vary with different tasks or brain regions.

[1] Oosterhof, Nikolaas N., et al. "Surface-based information mapping
reveals crossmodal vision–action representations in human parietal and
occipitotemporal cortex." *Journal of Neurophysiology* 104.2 (2010):
1077-1089.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-exppsy-pymvpa/attachments/20181010/5737c55a/attachment-0001.html>