[Eeglablist] surrogate statistics in SIFT

[Makoto Miyakoshi]

Dear Tyler,

> it tells me that it can't be performed on more than two datasets. I then
had a quick look at the SIFT manual, and it does appear as though only two
datasets are examined, one for each condition. I'm wondering how multiple
subjects is handled in this instance, is it designed for only one subject?
or am I meant to merge the datasets from the multiple subjects?

I think there is some kind of gorup-level solution for SIFT by the original
developer Tim Mullen and his colleague Nima Bigdely-Shamlo, both my former
colleagues in SCCN. If you use Nima's Measure Projection Toolbox, I think
you can find 'connectivity measures' or something like that in the GUI
menu. I have never tried it myself though. It probably projects only total
information outflow and inflow (i.e., sum of columns and rows of the
connnectivity matrix; in this way, you can handle the data as if it were
ERSP.)

I have developed groupSIFT solution, which can fully resolve ROI x ROI x
time x freq (x subjects) which is still version alpha (i.e., limited to my
collaborators only). The groupSIFT is based on Nima's Network Projection,
which is a connectivity version of the Measure Projection.
https://sccn.ucsd.edu/wiki/GroupSIFT
If you are interested in trying it out, write me separately.

Makoto

On Tue, Aug 21, 2018 at 9:55 AM Clement Lee <cll008 at eng.ucsd.edu> wrote:

> Dear Tyler,
>
> That is indeed correct. For group level analysis with SIFT, see groupSIFT
> <https://sccn.ucsd.edu/wiki/GroupSIFT> (currently under development).
> I've cc'd Makoto, the main developer (and author of wiki page), here.
>
> Best,
>
> On Mon, Aug 20, 2018 at 11:35 PM Tyler Grummett <
> tyler.grummett at flinders.edu.au> wrote:
>
>> Dear EEGLAB,
>>
>>
>> I've been slowly making my way through the SIFT toolbox, and I am
>> currently at the stage where I am trying to perform surrogate statistics
>> (or any statistics between two conditions, 20 subjects in each). All steps
>> previous were performed on each subject individually. When I run the
>> surrogate statistics, it tells me that it can't be performed on more than
>> two datasets. I then had a quick look at the SIFT manual, and it does
>> appear as though only two datasets are examined, one for each condition.
>> I'm wondering how multiple subjects is handled in this instance, is it
>> designed for only one subject? or am I meant to merge the datasets from the
>> multiple subjects?
>>
>>
>> Thanking you in advance as always,
>>
>> Tyler
>>
>>
>> *************************
>>
>> *Tyler Grummett ( BBSc, BSc(Hons I))*
>> *PhD Candidate*
>> *Brain Signals Laboratory*
>> *Multimodal Recording Facility*
>> *Flinders University Tonsley Building*
>> *Rm 4.17*
>> *Ext 19573*
>>
>
>
> --
> Clement Lee
> Applications Programmer
> Swartz Center for Computational Neuroscience
> Institute for Neural Computation, UC San Diego
> 858-822-7533
>


-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20180821/c480e5fa/attachment.html>