Hello, <br><br>I have written yesterday about troubles while running ft_timelockstatistics. As it happens now, this seems to be a compatibility issue between matlab 2011 and the current fieldtrip version (which you can solve temporarily by using older versions of both).<br>
<br>Best wishes, <br>Anna.<br><br><br><div class="gmail_quote">2011/8/30 <span dir="ltr"><<a href="mailto:fieldtrip-request@donders.ru.nl">fieldtrip-request@donders.ru.nl</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Send fieldtrip mailing list submissions to<br>
<a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
<br>
To subscribe or unsubscribe via the World Wide Web, visit<br>
<a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
or, via email, send a message with subject or body 'help' to<br>
<a href="mailto:fieldtrip-request@donders.ru.nl">fieldtrip-request@donders.ru.nl</a><br>
<br>
You can reach the person managing the list at<br>
<a href="mailto:fieldtrip-owner@donders.ru.nl">fieldtrip-owner@donders.ru.nl</a><br>
<br>
When replying, please edit your Subject line so it is more specific<br>
than "Re: Contents of fieldtrip digest..."<br>
<br>
<br>
Today's Topics:<br>
<br>
1. Call for Papers :: NIPS 2011 Workshop on Interpretable<br>
Decoding of Higher Cognitive States from Neural Data (Irina Simanova)<br>
2. Issues with neighbours in ft_timelockstatistics (Anna Lambrechts)<br>
3. Re: assessing significance in using ft_timelockanalysis<br>
results (Kanal Eliezer)<br>
<br>
<br>
----------------------------------------------------------------------<br>
<br>
Message: 1<br>
Date: Tue, 30 Aug 2011 14:22:31 +0200<br>
From: Irina Simanova <<a href="mailto:irina.simanova@mpi.nl">irina.simanova@mpi.nl</a>><br>
To: Email discussion list for the FieldTrip project<br>
<<a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a>><br>
Subject: [FieldTrip] Call for Papers :: NIPS 2011 Workshop on<br>
Interpretable Decoding of Higher Cognitive States from Neural Data<br>
Message-ID: <<a href="mailto:2E7CF2CC-C7AF-4937-935E-D0337E8445C8@mpi.nl">2E7CF2CC-C7AF-4937-935E-D0337E8445C8@mpi.nl</a>><br>
Content-Type: text/plain; charset="us-ascii"; Format="flowed";<br>
DelSp="yes"<br>
<br>
Interpretable Decoding of Higher Cognitive States from Neural Data<br>
<br>
NIPS 2011 Workshop, Dec 16 or 17, 2011, Granada, Spain<br>
<br>
(Please feel free to distribute the CFP to all the interested persons<br>
and groups.)<br>
<br>
Overview<br>
<br>
Over recent years, machine learning methods have become a crucial<br>
analytical tool in cognitive neuroscience (see reviews by Formisano et<br>
al., 2008; Pereira et al., 2009). Decoding techniques have<br>
dramatically increased the sensitivity of experiments, and so also the<br>
subtlety of cognitive questions that can be asked. At the same time<br>
the mental phenomena being studied are moving beyond lower-level<br>
perceptual and motor processes which are directly grounded in external<br>
measurable realities.<br>
<br>
Decoding higher cognition and interpreting the learned behaviour of<br>
the classifiers used pose unique challenges, as these psychological<br>
states are complex, fast-changing and often ill-defined. Contemporary<br>
machine learning methods deal well with the small numbers of cases,<br>
and high numbers of co-linear dimensions typical of neural data, and<br>
are generally optimized to maximize classification performance, rather<br>
than to enable meaningful interpretation of the features they learn<br>
from. And indeed recent work has succeeded to decode psychological<br>
phenomena including visual object recognition (e.g. Kriegeskorte et<br>
al., 2008; Connolly et al., 2011), perceptual interpretation of sounds<br>
(Staeren et al., 2009), lexical semantics (Mitchell et al., 2008;<br>
Siminova et al., 2010; Devereux et al., 2010; Murphy et al., 2011),<br>
decision making during game playing (Xiang et al., 2009) and the<br>
process of mental arithmetic (Anderson et al., 2008). But for the<br>
cognitive scientists who use these methods, the primary question is<br>
often not "how much" but rather "how" and "why" the patterns of neural<br>
activity identified by a machine learning algorithm encode particular<br>
cognitive processes.<br>
<br>
The aim of this workshop is therefore to 1) discuss the achievements<br>
and problems of the decoding of high-level cognitive states, and 2)<br>
explore the use of machine learning methodologies and other<br>
computational models that enable such cognitive interpretation of<br>
neural recordings of different modalities. Advances in this field<br>
require close collaboration between machine learning experts,<br>
neuroscientists and cognitive scientists. Thus, this workshop is<br>
highly interdisciplinary and will aim to attract submissions also from<br>
outside the existing NIPS community. By stimulating discussions among<br>
experts in the different fields, the workshop seeks to generate novel<br>
insights and new directions for research.<br>
<br>
Topics of interest<br>
<br>
The field requires techniques that are capable of taking advantage of<br>
spatially distributed patterns in the brain, that are separated in<br>
space but coordinated in their activity. Methods should also be<br>
sensitive to the fine-grained temporal patterns of multiple processes<br>
- which may proceed in a serial fashion, overlapping or in parallel<br>
with each-other, or in multiple passes with bidirectional information<br>
flows. Different recording modalities have distinctive advantages:<br>
fMRI provides very fine millimetre-level localisation in the brain but<br>
poor temporal resolution, while EEG and MEG have millisecond temporal<br>
resolution at the cost of spatial resolution. Ideally machine learning<br>
methods would be able to meaningfully combine complementary<br>
information from these different neuroimaging techniques (see e.g. De<br>
Martino et al., 2010). Moreover, as the processes underlying higher<br>
cognition are so complex, methods should be able to disentangle even<br>
tightly linked and confounded subprocesses. Finally, general use<br>
algorithms that could induce latent dimensions from neural data, and<br>
so reveal the "hidden" psychological states, would be a dramatic<br>
advance on current hypothesis-driven analytical paradigms. Originality<br>
of approach is encouraged and submissions on any related<br>
methodological approach are welcomed, such as:<br>
<br>
- Interpreting spatial and temporal location of selected features and<br>
their weights<br>
- Discovering "hidden" or "latent" cognitive representations<br>
- Disentangling confounded processes and representations<br>
- Comparing or combining data from recording modalities (e.g. fMRI,<br>
EEG, structural MRI, DTI, MEG, NIRS, EcOG, single cell recordings)<br>
- Fuzzy and partial classifications<br>
- Unaligned or incommensurate feature spaces and data representation<br>
<br>
As noted above, the complexity of higher cognition poses challenges.<br>
To take language comprehension as an example, speech is received at<br>
3-4 words; acoustic, semantic and syntactic processing can occur in<br>
parallel; and the form of underlying representations (sentence<br>
structures, conceptual descriptions) remains controversial. We welcome<br>
submissions dealing with any high-level cognitive functions that<br>
exhibit similar complexity, for instance:<br>
<br>
- Knowledge representation and concepts<br>
- Language and communication<br>
- Understanding visual and auditory experience<br>
- Memory and learning<br>
- Reasoning and problem solving<br>
- Decision making and executive control<br>
<br>
Submissions<br>
<br>
Authors are invited to submit full papers on original, unpublished<br>
work in the topic area of this workshop via the NIPS 2011 submission<br>
site at <a href="https://cmt.research.microsoft.com/NIPS2011/Default.aspx" target="_blank">https://cmt.research.microsoft.com/NIPS2011/Default.aspx</a>.<br>
Submissions should be formatted using the NIPS 2011 stylefiles, with<br>
blind review and not exceeding 8 pages plus an extra page for<br>
references. Author and submission information can be found at <a href="http://nips.cc/PaperInformation/AuthorSubmissionInstructions" target="_blank">http://nips.cc/PaperInformation/AuthorSubmissionInstructions</a><br>
. The stylefiles are available at <a href="http://nips.cc/PaperInformation/StyleFiles" target="_blank">http://nips.cc/PaperInformation/StyleFiles</a><br>
. Each submission will be reviewed at least by two members of the<br>
programme committee. Accepted papers will be published in the workshop<br>
proceedings. Dual submissions to the main NIPS 2011 conference and<br>
this workshop are allowed; if you submit to the main session, indicate<br>
this when you submit to the workshop. If your paper is accepted for<br>
the main session, you should withdraw your paper from the workshop<br>
upon notification by the main session.<br>
<br>
Important Dates<br>
<br>
- Aug 30, 2011: Call for papers<br>
- Sep 23, 2011: Deadline for submission of workshop papers<br>
- Oct 15, 2011: Notification of acceptance<br>
- Oct 31, 2011: Camera-ready papers due<br>
- Dec 16 or 17, 2011: Workshop date<br>
<br>
Links<br>
<br>
- NIPS 2011 website: <a href="http://nips.cc/Conferences/2011/" target="_blank">http://nips.cc/Conferences/2011/</a><br>
- Workshop website: <a href="https://sites.google.com/site/decodehighcogstate" target="_blank">https://sites.google.com/site/decodehighcogstate</a><br>
- Call for Papers: <a href="https://sites.google.com/site/decodehighcogstate/cfp/" target="_blank">https://sites.google.com/site/decodehighcogstate/cfp/</a><br>
<br>
Kind regards,<br>
The Workshop Organizers,<br>
Kai-min Kevin Chang, Anna Korhonen, Irina Simanova, Brian Murphy<br>
<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <<a href="http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110830/27b062df/attachment-0001.html" target="_blank">http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110830/27b062df/attachment-0001.html</a>><br>
<br>
------------------------------<br>
<br>
Message: 2<br>
Date: Tue, 30 Aug 2011 17:46:52 +0200<br>
From: Anna Lambrechts <<a href="mailto:anna.lambrechts@gmail.com">anna.lambrechts@gmail.com</a>><br>
To: <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
Subject: [FieldTrip] Issues with neighbours in ft_timelockstatistics<br>
Message-ID:<br>
<<a href="mailto:CALdyJ7HB%2BuaRs%2BmKa5s6ZxtZTbis-nNFz3EP7YQjZfamV2E3-g@mail.gmail.com">CALdyJ7HB+uaRs+mKa5s6ZxtZTbis-nNFz3EP7YQjZfamV2E3-g@mail.gmail.com</a>><br>
Content-Type: text/plain; charset="iso-8859-1"<br>
<br>
Dear fieldtrip community,<br>
<br>
Until about two weeks ago I was using a batch to process group comparison<br>
statistics on MEG data using ft_timelockstatistics. A week later, the same<br>
batch was not working anymore, even on the same data.<br>
<br>
The error I get is the following:<br>
<br>
*??? Improper index matrix reference.<br>
<br>
Error in ==> clusterstat>makechanneighbstructmat at 520<br>
[seld] = match_str(cfg.channel, cfg.neighbours(chan).label);<br>
<br>
Error in ==> clusterstat at 60<br>
channeighbstructmat = makechanneighbstructmat(cfg);<br>
<br>
Error in ==> statistics_montecarlo at 320<br>
[stat, cfg] = clusterstat(cfg, statrand, statobs,'issource',issource);<br>
<br>
Error in ==> statistics_wrapper at 290<br>
[stat, cfg] = statmethod(cfg, dat, design, 'issource',issource);<br>
<br>
Error in ==> ft_timelockstatistics at 124<br>
[stat, cfg] = statistics_wrapper(cfg, varargin{:});<br>
<br>
Error in ==> myfunction_stats at 59<br>
gpstat = ft_timelockstatistics(cfg, data1, data2);<br>
<br>
*It seems the error concerns the 'neighbours.mat' file reading, even though<br>
I am not sure about it. When I delete the two last lines in this file (to<br>
match the size of matrix wanted) the error disappears, but the processing is<br>
incomplete and false (or at least it does not match the results I had<br>
before).<br>
<br>
Has anyone encounter the same issue?<br>
Did you do any update recently that might explain this problem, and do you<br>
have any idea how to remedy to it?<br>
<br>
Best wishes,<br>
Anna.<br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <<a href="http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110830/f81c8aad/attachment-0001.html" target="_blank">http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20110830/f81c8aad/attachment-0001.html</a>><br>
<br>
------------------------------<br>
<br>
Message: 3<br>
Date: Tue, 30 Aug 2011 11:58:34 -0400<br>
From: Kanal Eliezer <<a href="mailto:ekanal@cmu.edu">ekanal@cmu.edu</a>><br>
To: Email discussion list for the FieldTrip project<br>
<<a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a>><br>
Subject: Re: [FieldTrip] assessing significance in using<br>
ft_timelockanalysis results<br>
Message-ID: <<a href="mailto:8CC6BA25-9A03-4630-B940-486BE2B9EE92@cmu.edu">8CC6BA25-9A03-4630-B940-486BE2B9EE92@cmu.edu</a>><br>
Content-Type: text/plain; charset=us-ascii<br>
<br>
Hello Eric -<br>
<br>
Thanks for the response. It looks like the fixed vs random analysis is exactly what I'm referring to. From what I understood, it looks like the difference really only shows up in the variance of the resultant distribution; with a random, the variance also takes into account the betwee-subjects variance. Is there a way to specify whether I want to do a fixed or random effects analysis in FieldTrip when I'm running ft_timelockgrandaverage or ft_freqgrandaverage? Thanks!<br>
<br>
Elli<br>
<br>
p.s. - In case anyone else is trying to figure this out, chapter 12 of Friston's book "Statistical Parametric Mapping" does an excellent job explaining the difference between fixed and random analyses, as well as how to implement it algorithmically.<br>
<br>
<br>
<br>
On Aug 26, 2011, at 3:00 PM, Eric Maris wrote:<br>
<br>
> Hi Kanal,<br>
><br>
>> The event related statistics tutorial<br>
>> (<a href="http://fieldtrip.fcdonders.nl/tutorial/eventrelatedstatistics" target="_blank">http://fieldtrip.fcdonders.nl/tutorial/eventrelatedstatistics</a>) talks<br>
> about<br>
>> assessing significance parametrically by running t-tests on pooled<br>
>> timelockanalysis data. My question is, does the fact that the averages<br>
> were<br>
>> created from N trials make a difference? If I'm condition A has twelve<br>
>> averages and condition B has another twelve, and each average contains 70<br>
>> trials, is there a way to "inform" the statistical test that the power in<br>
> this<br>
>> dataset is greater than 24? Is this only possible if I run the t-test<br>
> comparing<br>
>> each set of 840 (70*12) trials?<br>
>><br>
>> I'm also curious whether this is possible with non-parametric analyses, as<br>
>> well. Thanks -<br>
><br>
> In an analysis over subjects (called random-effects analysis in the fMRI<br>
> literature), "informing" the statistical test about the number of trials per<br>
> condition only makes sense if this number is different for the two<br>
> conditions. I propose that you have a look the fMRI papers that deal with<br>
> the issue of fixed-versus-random effect analyses. The conceptual issues<br>
> involved are the same in fMRI and electrophysiology.<br>
><br>
><br>
> Best,<br>
><br>
> Eric Maris<br>
><br>
><br>
><br>
><br>
>><br>
>> Elli Kanal<br>
>><br>
>><br>
>> --------------------<br>
>> Eliezer Kanal, Ph.D.<br>
>> Postdoctoral Fellow<br>
>> Center for the Neural Basis of Cognition<br>
>> Carnegie Mellon University<br>
>> 4400 Fifth Ave, Suite 110A<br>
>> Pittsburgh PA 15213<br>
>> P: <a href="tel:412-268-4115" value="+14122684115">412-268-4115</a><br>
>> F: <a href="tel:412-268-5060" value="+14122685060">412-268-5060</a><br>
>><br>
>><br>
>> _______________________________________________<br>
>> fieldtrip mailing list<br>
>> <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
>> <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
><br>
> _______________________________________________<br>
> fieldtrip mailing list<br>
> <a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
> <a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
<br>
<br>
<br>
<br>
------------------------------<br>
<br>
_______________________________________________<br>
fieldtrip mailing list<br>
<a href="mailto:fieldtrip@donders.ru.nl">fieldtrip@donders.ru.nl</a><br>
<a href="http://mailman.science.ru.nl/mailman/listinfo/fieldtrip" target="_blank">http://mailman.science.ru.nl/mailman/listinfo/fieldtrip</a><br>
<br>
End of fieldtrip Digest, Vol 9, Issue 30<br>
****************************************<br>
</blockquote></div><br>