Subject: | | Re: Two-Sample t-Test |
From: | | Hector Espinos <[log in to unmask]> |
Reply-To: | | [log in to unmask][log in to unmask]> wrote:
> Dear Daniele, > thanks for your answer. > My aim is to use time-varying partial directed coherence on EEG source > activity and I was wondering whether some source reconstruction techniques > may be more suited than others (e.g., I read from the SPM manual that its > reconstruction methods work by reducing the error at some peak or time > window of interest, not sure if that would be optimal for time-varying > connectivity). > > As for the inevitable mixing, I see your point, do you have references > where both negligible and dramatic effects on source connectivity are shown? > > Best Regards, > David > > > 2017-05-05 22:25 GMT+02:00 Daniele Marinazzo <[log in to unmask]> > : > >> Dear David >> >> it depends on the type of connectivity measure that you want to apply. >> If it's DCM, then you don't need to reconstruct the activity at the >> sources, since the transfer function is part of the DCM itself. >> Otherwise, there are several algorithms, and everyone has their favourite >> recipe. The two main families are the lcmv/beamformer and the loreta/wmne. >> If you want to stick with matlab environment, spm, eeglab and fieldtrip >> have all valid source reconstruction methods. >> SIFT is a eeglab toolbox specialized for connectivity at the source level >> https://sccn.ucsd.edu/wiki/SIFT >> Otherwise you may want to try MNE, in python https://www.martinos.or >> g/mne/stable/index.html >> >> p.s. I don't want to be a party pooper but there's no free lunch in >> source reconstruction connectivity. Unfortunately there's some inevitable >> mixing when the activity propagates to the scalp, that cannot be undone. >> These effects go from negligible to dramatic. >> >> best regards >> >> d. >> >> On Fri, May 5, 2017 at 7:29 PM, David Pascucci <[log in to unmask]> wrote: >> >>> Dear SPM experts, >>> are you aware of any recommended source reconstruction method (among >>> those in SPM, fieldtrip or others) >>> that may be more appropriate for time-varying directed connectivity >>> analysis on EEG source activity? >>> >>> Best Regards, >>> David >>> >>> >> > > > -- > --------------------- > David Pascucci > > Postdoctoral Fellow > University of Fribourg > Department of Psychology > Rue de Faucigny 2 > 1700 Fribourg > Switzerland >[log in to unmask] |
Date: | | Mon, 22 May 2017 17:45:30 +0200 |
Content-Type: | | text/plain |
Parts/Attachments: |
|
|
|
|
Dear Sascha,
I understand that your contrast is a t-contrast. T-constrast have a
direction, for this reason it is used when we compare between groups,
while F-Contrast is for showing an effect of interest.
If you have group 1 .vs. group 2 and put a t-constrast 1 -1 you are
supposing that group 1 is major that group 2. If, in the contrary you
créate a t-contrast -1 1 you are supposing that the effect of the second
group is major. In other words, if you have significances in the
contrast 1 -1, the meanign is that you have more activity in the group 1
than group 2. In the other case, the group 1 shows less activity that
the group 2. These meaning is more or less, it depends that the study
that you are doing.
I hope that I help you!
If I am wrong, please SPM users, tell me!
Best,
Héctor Espinós
INCLIVA Public Health Institute
El 2017-05-22 11:42, Sascha Frölich escribió:
> Dear all,
>
> I have a question regarding the two-sample t-test in SPM:
>
> When comparing group 1 vs group 2 (ie. contras 1 -1) I get a different
> result from when I use contrast -1 1 (group2 vs group 1, as far as I
> understood). How is this possible? Shouldn't the t-test by symmetric?
> What do the different results mean?
>
> Thanks!
>
> Best,
> Sascha
|
|
|
|