From polomacnenad at gmail.com Sat Dec 1 15:02:14 2012 From: polomacnenad at gmail.com (Nenad Polomac) Date: Sat, 1 Dec 2012 15:02:14 +0100 Subject: [FieldTrip] design matrix for ft_freqstatistics; between subjects Message-ID: Hi everybody, I have some doubts concerning statistic in between subjects experiment. I couldn't find explanation for this case in tutorial. I have two groups of subjects. And I would like to compare this two groups for the same condition using ft_freqstatistics (cfg.correctm = 'cluster'). I have two time frequency grand average for each group (8 subjects in each group; cfg.keepindividual = 'yes';). I am not sure how a design matrix for ft_freqstatistics suppose to look like I suppose: design=[1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2] Is this correct? Thank you in advance! Nenad -------------- next part -------------- An HTML attachment was scrubbed... URL: From k.muesch at uke.uni-hamburg.de Sat Dec 1 16:15:31 2012 From: k.muesch at uke.uni-hamburg.de (=?iso-8859-1?Q?Kathrin_M=FCsch?=) Date: Sat, 1 Dec 2012 16:15:31 +0100 Subject: [FieldTrip] design matrix for ft_freqstatistics; between subjects In-Reply-To: References: Message-ID: Hi Nenad, It's like a between trials experiment within one subject (first example of the tutorial). Your design is correct. Remember to use independent samples statistics because you compare two independent groups (in your case subjects). cfg.statistic = 'indepsamplesT'; cfg. design = design; cfg.ivar = 1; Best, Kathrin Am 01.12.2012 um 15:02 schrieb Nenad Polomac: > Hi everybody, > > I have some doubts concerning statistic in between subjects experiment. I couldn't find explanation for this case in tutorial. > I have two groups of subjects. And I would like to compare this two groups for the same condition using ft_freqstatistics (cfg.correctm = 'cluster'). I have two time frequency grand average for each group (8 subjects in each group; cfg.keepindividual = 'yes';). I am not sure how a design matrix for ft_freqstatistics suppose to look like > > I suppose: design=[1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2] > Is this correct? > > Thank you in advance! > > Nenad > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -------------- next part -------------- -- Pflichtangaben gemäß Gesetz über elektronische Handelsregister und Genossenschaftsregister sowie das Unternehmensregister (EHUG): Universitätsklinikum Hamburg-Eppendorf; Körperschaft des öffentlichen Rechts; Gerichtsstand: Hamburg Vorstandsmitglieder: Prof. Dr. Martin Zeitz (Vorsitzender), Dr. Alexander Kirstein, Joachim Prölß, Prof. Dr. Dr. Uwe Koch-Gromus -------------- next part -------------- An HTML attachment was scrubbed... URL: From fanny.lachat at gmail.com Mon Dec 3 13:24:40 2012 From: fanny.lachat at gmail.com (Fanny) Date: Mon, 3 Dec 2012 13:24:40 +0100 Subject: [FieldTrip] MOrler wavelet nonconstant Q Message-ID: Dear Fieldtrip users, I am reading this paper : 1- Schneider, T.R., et al., Enhanced EEG gamma-band activity reflects multisensory semantic matching in visual-toauditory object priming, NeuroImage (2008), I am interesting in the methodology : using a *nonconstant Q wavelet*. However I am a little confused : " *Time-frequency analysis was performed for each channel by convolving the data with a complex Morlet wavelet w(t, f0) having a Gaussian shape in time (σt) and in frequency (σf) around the center frequency ( f0). Nonconstant wavelets with Q increasing from f0/σf =8.5 to 13 for frequencies from 30 to 100 Hz (step size 2 Hz) were used. * " - I don't understand what is *f*0 and σ*f* and thus , how can I find the right Q with my data .. I am trying to find the right Q for my analysis and it seems that a *nonconsant Q* is what I need. ( 4-80HZ, step size 1Hz, form 1.2 sec to 1.3sec , time step 0.002) finally, Does anyone else used also the nonconstant Q method ? Thank you in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: From jm.horschig at donders.ru.nl Mon Dec 3 13:55:16 2012 From: jm.horschig at donders.ru.nl (=?ISO-8859-1?Q?=22J=F6rn_M=2E_Horschig=22?=) Date: Mon, 03 Dec 2012 13:55:16 +0100 Subject: [FieldTrip] MOrler wavelet nonconstant Q In-Reply-To: References: Message-ID: <50BCA134.5060400@donders.ru.nl> Dear Fanny, I have no experience in using this nonconstant Q wavelet, but I hope I can shed some light on the terminology. > " > /Time--frequency analysis was performed for > each channel by convolving the data with a complex Morlet > wavelet w(t, f0) having a Gaussian shape in time (?t) and in > frequency (?f) around the center frequency ( f0). Nonconstant > wavelets with Q increasing from f0/?f =8.5 to 13 for frequencies > from 30 to 100 Hz (step size 2 Hz) were used. > / > " > > > - I don't understand what is /f/_0 and ?/_f / If you compute the frequency spectrum of any signal, you will see some peaks in the frequency spectrum. If you compute the frequency spectrum of a wavelet, the maximal peak is conventionally called center frequency. In Matlab, you can obtain the center frequency by the function 'centfrq'. Note that the center frequency also depends on the scale of your wavelet. Any peak can be wide or narrow, and just like in statistics you can describe this with the standard deviation. I am not in particular familiar with/?f /and the method you are describing, but I would guess that it is really just the standard deviation, given that /? /is conventionally used to denote the standard deviation. Judging from your questions, without offending, that you are probably new to the field of wavelets, you should probably start off with reading more general literature on wavelets. Maybe start off with Wikipedia: http://en.wikipedia.org/wiki/Wavelet http://en.wikipedia.org/wiki/Morlet_wavelet and then go to some more sophisticated literature ;) Best, Jörn -- Jörn M. Horschig PhD Student Donders Institute for Brain, Cognition and Behaviour Centre for Cognitive Neuroimaging Radboud University Nijmegen Neuronal Oscillations Group FieldTrip Development Team P.O. Box 9101 NL-6500 HB Nijmegen The Netherlands Contact: E-Mail: jm.horschig at donders.ru.nl Tel: +31-(0)24-36-68493 Web: http://www.ru.nl/donders Visiting address: Trigon, room 2.30 Kapittelweg 29 NL-6525 EN Nijmegen The Netherlands -------------- next part -------------- An HTML attachment was scrubbed... URL: From t.schneider at uke.uni-hamburg.de Mon Dec 3 16:12:07 2012 From: t.schneider at uke.uni-hamburg.de (Till Schneider) Date: Mon, 3 Dec 2012 16:12:07 +0100 Subject: [FieldTrip] MOrler wavelet nonconstant Q In-Reply-To: <50BCA134.5060400@donders.ru.nl> References: <50BCA134.5060400@donders.ru.nl> Message-ID: <50BCC147.9000009@uke.uni-hamburg.de> Dear Fanny, we used wavelets with linearly increasing Q. Q defines the temporal and spectral resolution of the wavelet. In fieldtrip you have to change the parameter cfg.width, in order to adjust Q for the wavelet analysis using ft_freqanalysis. You could use a linear increasing value for cfg.width dependent on the frequency of interest, e.g. cfg.width = linspace(8.5, 13, length(30:2:100)). Together with a frequency range from 30 to 100 Hz (cfg.foi = 30:2:100), this would yield wavelets with increasing Q across frequencies. cfg = []; cfg.channel = 'EEG'; cfg.method = 'wavelet'; cfg.width = linspace(8.5, 13, length(30:2:100)); cfg.output = 'pow'; cfg.foi = 30:2:100; cfg.toi = -0.5:0.05:1.5; TFRwavelet = ft_freqanalysis(cfg, data); Best, Till -- Till Schneider, PhD Cognitive and Clinical Neurophysiology Group Dept. of Neurophysiology and Pathophysiology University Medical Center Hamburg-Eppendorf Martinistr. 52 20246 Hamburg Germany phone +49-40-7410-53188 fax +49-40-7410-57126 www.uke.de/neurophysiologie Am 03.12.12 13:55, schrieb "Jörn M. Horschig": > Dear Fanny, > > I have no experience in using this nonconstant Q wavelet, but I hope I > can shed some light on the terminology. > >> " >> /Time--frequency analysis was performed for >> each channel by convolving the data with a complex Morlet >> wavelet w(t, f0) having a Gaussian shape in time (?t) and in >> frequency (?f) around the center frequency ( f0). Nonconstant >> wavelets with Q increasing from f0/?f =8.5 to 13 for frequencies >> from 30 to 100 Hz (step size 2 Hz) were used. >> / >> " >> >> >> - I don't understand what is /f/_0 and ?/_f / > > If you compute the frequency spectrum of any signal, you will see some > peaks in the frequency spectrum. If you compute the frequency spectrum > of a wavelet, the maximal peak is conventionally called center > frequency. In Matlab, you can obtain the center frequency by the > function 'centfrq'. Note that the center frequency also depends on the > scale of your wavelet. > > Any peak can be wide or narrow, and just like in statistics you can > describe this with the standard deviation. I am not in particular > familiar with/?f /and the method you are describing, but I would guess > that it is really just the standard deviation, given that /? /is > conventionally used to denote the standard deviation. > > Judging from your questions, without offending, that you are probably > new to the field of wavelets, you should probably start off with > reading more general literature on wavelets. Maybe start off with > Wikipedia: > http://en.wikipedia.org/wiki/Wavelet > http://en.wikipedia.org/wiki/Morlet_wavelet > > and then go to some more sophisticated literature ;) > > Best, > Jörn > > > -- > Jörn M. Horschig > PhD Student > Donders Institute for Brain, Cognition and Behaviour > Centre for Cognitive Neuroimaging > Radboud University Nijmegen > Neuronal Oscillations Group > FieldTrip Development Team > > P.O. Box 9101 > NL-6500 HB Nijmegen > The Netherlands > > Contact: > E-Mail:jm.horschig at donders.ru.nl > Tel: +31-(0)24-36-68493 > Web:http://www.ru.nl/donders > > Visiting address: > Trigon, room 2.30 > Kapittelweg 29 > NL-6525 EN Nijmegen > The Netherlands > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Pflichtangaben gemäß Gesetz über elektronische Handelsregister und Genossenschaftsregister sowie das Unternehmensregister (EHUG): Universitätsklinikum Hamburg-Eppendorf; Körperschaft des öffentlichen Rechts; Gerichtsstand: Hamburg Vorstandsmitglieder: Prof. Dr. Martin Zeitz (Vorsitzender), Dr. Alexander Kirstein, Joachim Prölß, Prof. Dr. Dr. Uwe Koch-Gromus -------------- next part -------------- An HTML attachment was scrubbed... URL: From f.roux at bcbl.eu Mon Dec 3 16:58:13 2012 From: f.roux at bcbl.eu (Frederic Roux) Date: Mon, 03 Dec 2012 16:58:13 +0100 (CET) Subject: [FieldTrip] log transform of power values Message-ID: <192a3684-103d-4207-ba2e-d2b88c2c9882@thalamus_p> Dear all, I am planning to run freq_statistics on single trial data, with the power data computed in the following way: cfg = []; cfg.method = 'sincos'; etc ... planar_data = ft_meplanar(cfg,raw_data); cfg = []; cfg.foi = 30:150; cfg.output = 'pow'; cfg.taper = 'dpss'; etc ... freq_planar = ft_freqanalysis(cfg,planar_data); freq = ft_combineplanar([],freq_planar); When I then look at the topographies cfg = []; cfg.grad = freq.grad; cfg.layout = ft_perpare_layout(cfg); cfg.xlim = [.1 .6]; cfg.ylim = [40 60]; cfg.parameter = 'powspctrm'; figure; ft_topoplotTFR(cfg,freq); I noticed that it makes a huge difference if the data are log transformed before the plotting, ie freq.powspctrm = 20*log10(freq.powspctrm); The data looks more 'clustered' when I do the log transform, whereas the 'raw' power values look scattered and like tiny islands. Now I am wondering if this could help with the statistical analysis, meaning that it would give me more power to detect spatio-temporal clusters. (Baseline correction is not really a good idea, as there seems to be important stuff going on in the baseline) Any (quick) help and advice would be highly appreciated. Fred Frédéric Roux Postdoctoral Researcher www.bcbl.eu From octavian.lie at gmail.com Mon Dec 3 20:48:23 2012 From: octavian.lie at gmail.com (octavian lie) Date: Mon, 3 Dec 2012 13:48:23 -0600 Subject: [FieldTrip] Source imaging Message-ID: Dear All, I am new to FT, and I would appreciate your imput to the following: 1. Can I used MNE derived meshes in FT? 2. Can I use OpenMeeg BEM in FT? 3. Can I use MUSIC, RAP MUSIC, sLORETA and swLORETA in FT? 4. Are there any methods implemented in FT to calculate regularization parameter for distributed source models? 5. Can I import desikan and destrieux freesurfer atlases as scouts in FT? 6. What is the timeline for implementing Simbio in FT? Thank you, Octavian. -------------- next part -------------- An HTML attachment was scrubbed... URL: From khaled.alkamha at gmail.com Tue Dec 4 12:47:33 2012 From: khaled.alkamha at gmail.com (Khaled Al-Kamha) Date: Tue, 4 Dec 2012 14:47:33 +0300 Subject: [FieldTrip] Requirements Need Message-ID: Hey,, I'm an under graduated student and i don't have too much experience in MATLAB or any advance subject that serve the process of EEG analysis, but i'm a quick learner and i have a good information from papers i read. I just want your advice about what the first things that i have to know to speed up the process? Thanks in advance -------------- next part -------------- An HTML attachment was scrubbed... URL: From poil.simonshlomo at gmail.com Tue Dec 4 13:15:27 2012 From: poil.simonshlomo at gmail.com (Simon-Shlomo Poil) Date: Tue, 4 Dec 2012 13:15:27 +0100 Subject: [FieldTrip] Requirements Need In-Reply-To: References: Message-ID: Dear Khaled, You could start reading the tutorial sites for the different EEG analysis tools: http://www.nbtwiki.net/doku.php?id=tutorial:start#.UL3ontez4_Q http://fieldtrip.fcdonders.nl/tutorial http://sccn.ucsd.edu/wiki/EEGLAB What to read there depends on what exactly you want to analyse? -Simon 2012/12/4 Khaled Al-Kamha : > Hey,, > > I'm an under graduated student and i don't have too much experience in > MATLAB or any advance subject that serve the process of EEG analysis, > but i'm a quick learner and i have a good information from papers i read. > > I just want your advice about what the first things that i have to know to > speed up the process? > > Thanks in advance > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- -- Simon-Shlomo Poil Center of MR-Research University Children’s Hospital Zurich Mobile number: +41 (0)76 399 5809 Office number: +41 (0)44 266 3129 Skype: poil.simonshlomo Webpage: http://www.poil.dk/s/ and http://www.nbtwiki.net and http://www.kispi.uzh.ch/Kinderspital/Medizin/mrzentrum_en.html From eelke.spaak at donders.ru.nl Tue Dec 4 13:17:22 2012 From: eelke.spaak at donders.ru.nl (Eelke Spaak) Date: Tue, 4 Dec 2012 13:17:22 +0100 Subject: [FieldTrip] Requirements Need In-Reply-To: References: Message-ID: Hi Khaled, Assuming you already have Matlab and FieldTrip installed, the best way to get started learning to use FieldTrip is by going through the tutorials, http://fieldtrip.fcdonders.nl/tutorial (starting with 'Introduction'). It is quite helpful to have some Matlab experience as well, for that Mathworks themselves provide some excellent getting started guides. Best, Eelke On 4 December 2012 12:47, Khaled Al-Kamha wrote: > Hey,, > > I'm an under graduated student and i don't have too much experience in > MATLAB or any advance subject that serve the process of EEG analysis, > but i'm a quick learner and i have a good information from papers i read. > > I just want your advice about what the first things that i have to know to > speed up the process? > > Thanks in advance > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From haiteng.jiang at gmail.com Tue Dec 4 15:37:07 2012 From: haiteng.jiang at gmail.com (Haiteng Jiang) Date: Tue, 4 Dec 2012 15:37:07 +0100 Subject: [FieldTrip] read significant brain areas after source statistic Message-ID: Hi FTers, I want to know the significant brain areas after I do the group source statistic while I am not familiar with the brain structure, so it is difficult for me to identify them by eye. Is there a easy and straightforward way to extract them out? Any comment is appreciated! Thanks in advance ! Best, Haiteng -- Haiteng Jiang PhD candidate Neuronal Oscillations Group Donders Institute for Brain, Cognition and Behaviour Centre for Cognitive Neuroimaging Radboud University Nijmegen Visiting address Room 2.32 Donders Centre for Cognitive Neuroimaging Kapittelweg 29 6525 EN Nijmegen the Netherlands Tel.: +31 (0)243668291 Web: https://sites.google.com/site/haitengjiang/ -------------- next part -------------- An HTML attachment was scrubbed... URL: From f.roux at bcbl.eu Tue Dec 4 16:23:23 2012 From: f.roux at bcbl.eu (Frederic Roux) Date: Tue, 04 Dec 2012 16:23:23 +0100 (CET) Subject: [FieldTrip] ft_sourceplot does not work for non-interpolated source data anymore Message-ID: <059af47e-5f00-4e8d-a5f4-448d6f4d27c4@thalamus_p> Dear list members, in previous fieldtrip releases it was always possible to do a sourceplot of both: the non-interpolated beamformer maps as well as the interpolated beamformer maps by doing %non-interpolated sourceplot cfg = []; cfg.funparameter = 'avg.pow'; cfg = []; cfg.method = 'ortho'; cfg.funparameter = 'avg.pow'; figure; ft_sourceplot(cfg,lcmv_data); and %interpolated sourceplot cfg = []; cfg.parameter = 'avg.pow'; int_lcmv = ft_sourceinterpolate(cfg,lcmv_data,MNI_template); cfg = []; cfg.method = 'ortho'; cfg.funparameter = 'avg.pow'; figure; ft_sourceplot(cfg,int_lcmv); I am using grids computed from warped anatomies and recently switched from an older version to release fieldtrip-20120812 and noticed that it seems to be no longer possible to plot the non-interpolated source data. However, this was particularly useful to pick voxels of interest and compute virtual channels based on their gird number. Is there a particular reason why this feature has disappeared? How can I get grid numbers from the MNI coordinates / voxel number that ft_sourceplot is giving me ? Best, Frederic Frédéric Roux Postdoctoral Researcher www.bcbl.eu From andrea.brovelli at univ-amu.fr Wed Dec 5 10:32:12 2012 From: andrea.brovelli at univ-amu.fr (andrea brovelli) Date: Wed, 05 Dec 2012 10:32:12 +0100 Subject: [FieldTrip] 642 nodes for source model based on cortical sheet Message-ID: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> Dear all, I am constructing a source model based on cortical sheet and I get always 642 nodes. I tried to track down this magic number and I have found that at lines 209 and 229 of ft_prepare_sourcemodel.m you fix it to: cfg.spheremesh = ft_getopt(cfg, 'spheremesh', 642); % FIXME move spheremesh to cfg.grid What was the rationale behind this ? Did you modify this in more recent versions of Fieldtrip or how could we have it in the cfg in the future ? Thanks a lot bye Andrea From eelke.spaak at donders.ru.nl Wed Dec 5 10:52:09 2012 From: eelke.spaak at donders.ru.nl (Eelke Spaak) Date: Wed, 5 Dec 2012 10:52:09 +0100 Subject: [FieldTrip] 642 nodes for source model based on cortical sheet In-Reply-To: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> References: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> Message-ID: Dear Andrea, To be honest, I have never used or worked on ft_prepare_sourcemodel, but I do know that the line you quote is a call to ft_getopt in which the last value is only the default. So, if cfg has a field called 'spheremesh', then that value will be used, rather than the default. Put in other words: cfg = []; cfg.spheremesh = 12; result = ft_getopt(cfg, 'spheremesh', 642) then result should be 12, and not 642. Have you tried specifying cfg.spheremesh = some value? If that does not work, the value is being changed somewhere else, not in the line you quote (which might be a bug, I don't know enough about ft_prepare_sourcemodel to make that judgement). Best, Eelke On 5 December 2012 10:32, andrea brovelli wrote: > > Dear all, > > I am constructing a source model based on cortical sheet and I get always > 642 nodes. > > I tried to track down this magic number and I have found that at lines 209 > and 229 of ft_prepare_sourcemodel.m you fix it to: > > cfg.spheremesh = ft_getopt(cfg, 'spheremesh', 642); % FIXME move > spheremesh to cfg.grid > > > What was the rationale behind this ? > > Did you modify this in more recent versions of Fieldtrip or how could we > have it in the cfg in the future ? > > Thanks a lot > > bye > > Andrea > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From patrick.jung at esi-frankfurt.de Wed Dec 5 14:57:45 2012 From: patrick.jung at esi-frankfurt.de (Jung, Patrick) Date: Wed, 5 Dec 2012 13:57:45 +0000 Subject: [FieldTrip] Partial artifact rejection Message-ID: <36E953F5321EB743AC6B338995A56D632519C8@UM-EXCDAG-A01.um.gwdg.de> Dear all, in the reference documentation it says "In case of partial artifact rejection, a minimum length of the resulting sub-trials can be specified." I'd like to reject the trial segment from the beginning of the artifact onwards to keep only the data before the artifact occurred. But I don't want to append subtrials. How can I do that in fieldtrip? Many thanks for your help! Best, Patrick -------------- next part -------------- An HTML attachment was scrubbed... URL: From andrea.brovelli at univ-amu.fr Wed Dec 5 16:56:49 2012 From: andrea.brovelli at univ-amu.fr (andrea brovelli) Date: Wed, 05 Dec 2012 16:56:49 +0100 Subject: [FieldTrip] 642 nodes for source model based on cortical sheet In-Reply-To: References: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> Message-ID: <20121205165649.18002h7nqqevt9z4@wmelperso1.univmed.fr> Thanks, it works by simply putting cfg.spheremesh = xxxx in the cfg of ft_prepare_leadfield. But was there as special reason why this was set to 642 ? thanks Andrea ----- Message de eelke.spaak at donders.ru.nl --------- Date : Wed, 5 Dec 2012 10:52:09 +0100 De : Eelke Spaak Répondre à : FieldTrip discussion list Objet : Re: [FieldTrip] 642 nodes for source model based on cortical sheet À : FieldTrip discussion list > Dear Andrea, > > To be honest, I have never used or worked on ft_prepare_sourcemodel, > but I do know that the line you quote is a call to ft_getopt in which > the last value is only the default. So, if cfg has a field called > 'spheremesh', then that value will be used, rather than the default. > > Put in other words: > > cfg = []; > cfg.spheremesh = 12; > result = ft_getopt(cfg, 'spheremesh', 642) > > then result should be 12, and not 642. > > Have you tried specifying cfg.spheremesh = some value? If that does > not work, the value is being changed somewhere else, not in the line > you quote (which might be a bug, I don't know enough about > ft_prepare_sourcemodel to make that judgement). > > Best, > Eelke > > On 5 December 2012 10:32, andrea brovelli > wrote: >> >> Dear all, >> >> I am constructing a source model based on cortical sheet and I get always >> 642 nodes. >> >> I tried to track down this magic number and I have found that at lines 209 >> and 229 of ft_prepare_sourcemodel.m you fix it to: >> >> cfg.spheremesh = ft_getopt(cfg, 'spheremesh', 642); % FIXME move >> spheremesh to cfg.grid >> >> >> What was the rationale behind this ? >> >> Did you modify this in more recent versions of Fieldtrip or how could we >> have it in the cfg in the future ? >> >> Thanks a lot >> >> bye >> >> Andrea >> _______________________________________________ >> fieldtrip mailing list >> fieldtrip at donders.ru.nl >> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > ----- Fin du message de eelke.spaak at donders.ru.nl ----- From Margit.Schoenherr at uk-erlangen.de Wed Dec 5 17:07:42 2012 From: Margit.Schoenherr at uk-erlangen.de (=?iso-8859-1?Q?Sch=F6nherr=2C_Margit?=) Date: Wed, 5 Dec 2012 17:07:42 +0100 Subject: [FieldTrip] problem with reading .edf+ files Message-ID: <71D8F8A56F37A947912FC86D4D9F00F477C17C494F@XMAIL1.medads.uk-erlangen.de> Hello, I want to analyze invasive EEG data and I have problems reading the data correctly. The data are continuous and I used BESA to convert them to .edf format. When I read them into matlab, however, the data are not continuous anymore, but segmented into very short trials. For example, a data set of 1 hour is split into 8229 trials of 0.43 s length. Even more confusing, the trial length differs across subjects. What is the difference between .edf and .edf+? BESA Research produces the latter, could this be the reason for the issue? I also tried to track down the place where the segmentation into trials is done, but I ended up at line 111 of read_edf.m (EDF.Dur = str2num(H1(245:252)) % 8 Byte # duration of data record in sec). I would be very glad if someone could tell me, how to read the data correctly with fieldtrip. Thanks! Best, Margit From tomh at kurage.nimh.nih.gov Wed Dec 5 17:27:39 2012 From: tomh at kurage.nimh.nih.gov (Tom Holroyd (NIH/NIMH) [E]) Date: Wed, 05 Dec 2012 11:27:39 -0500 Subject: [FieldTrip] 642 nodes for source model based on cortical sheet In-Reply-To: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> References: <20121205103212.16086y45q3whe5j4@wmelperso1.univmed.fr> Message-ID: <50BF75FB.5050106@kurage.nimh.nih.gov> I think the model in question consists of one dipole at each vertex of a recursive subdivision of some platonic solid, such as a triangular subdivision of an octahedron's faces, a very popular method for generating spherical meshes. Each of these will produce a characteristic number, and in particular for a triangulated icosahedron it's 2 + 10 * 2**(d*2), where d is the recursion depth, which for d = 3 turns out to be 642. So your "cortical sheet" is probably a triangulated icosahedron with spherical topology. It may not look like a sphere. The mesh is allowed to deform. Somewhere in the code you could maybe make a new spheremesh, for the next d of 4, you get 2562 nodes. If you want more. andrea brovelli wrote: > > Dear all, > > I am constructing a source model based on cortical sheet and I get > always 642 nodes. > > I tried to track down this magic number and I have found that at lines > 209 and 229 of ft_prepare_sourcemodel.m you fix it to: > > cfg.spheremesh = ft_getopt(cfg, 'spheremesh', 642); % FIXME move > spheremesh to cfg.grid > > > What was the rationale behind this ? > > Did you modify this in more recent versions of Fieldtrip or how could we > have it in the cfg in the future ? > > Thanks a lot > > bye > > Andrea > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- The white knight is talking backwards. From inieuwenhuis at berkeley.edu Wed Dec 5 19:08:52 2012 From: inieuwenhuis at berkeley.edu (Ingrid Nieuwenhuis) Date: Wed, 05 Dec 2012 10:08:52 -0800 Subject: [FieldTrip] Partial artifact rejection In-Reply-To: <36E953F5321EB743AC6B338995A56D632519C8@UM-EXCDAG-A01.um.gwdg.de> References: <36E953F5321EB743AC6B338995A56D632519C8@UM-EXCDAG-A01.um.gwdg.de> Message-ID: <50BF8DB4.1020906@berkeley.edu> Hi Patrick, What you can do is use partial artifact rejection by calling ft_rejectarttifact with cfg.artfctdef.reject = 'partial'. Then subsequently loop over the trials you get, and identify the ones (by looking at data.time{iTr}) that do not start at the time your trials normally start (this means it's a trial after an artifact). Finally, exclude those by using ft_redefinetrials. For instance, if your trials start with t=0, then after ft_rejectartifact you can do something like: ind_keep = zeros(length(data.trial),1); %allocate for iTrl = 1:length(data.trial) if data.trial{iTrl}(1) ~= 0 %this means it's a trial occurring AFTER an artifact ind_keep(iTrl) = false; else %this means it's a trial occurring BEFORE an artifact ind_keep(iTrl) = true; end end %only keep the trials before the artifact: cfg = []; cfg.trials = find(ind_keep); data = ft_redefinetrial(cfg, data); Hope this helps, Ingrid On 12/5/2012 5:57 AM, Jung, Patrick wrote: > > Dear all, > > in the reference documentation it says "In case of partial artifact > rejection, a minimum length of the > > resulting sub-trials can be specified." > > I'd like to reject the trial segment from the beginning of the > artifact onwards to keep only the data before the artifact occurred. > > But I don't want to append subtrials. > > How can I do that in fieldtrip? > > Many thanks for your help! > > Best, > Patrick > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Ingrid Nieuwenhuis PhD Postdoctoral Fellow Sleep and Neuroimaging Laboratory Department of Psychology University of California, Berkeley California 94720-1650 Tolman Hall, room 5305 -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick.jung at esi-frankfurt.de Thu Dec 6 09:54:57 2012 From: patrick.jung at esi-frankfurt.de (Jung, Patrick) Date: Thu, 6 Dec 2012 08:54:57 +0000 Subject: [FieldTrip] Partial artifact rejection In-Reply-To: <50BF8DB4.1020906@berkeley.edu> References: <36E953F5321EB743AC6B338995A56D632519C8@UM-EXCDAG-A01.um.gwdg.de> <50BF8DB4.1020906@berkeley.edu> Message-ID: <36E953F5321EB743AC6B338995A56D63251A44@UM-EXCDAG-A01.um.gwdg.de> Hi Ingrid, thanks! That definitely helps a lot! Cheers, Patrick From: fieldtrip-bounces at science.ru.nl [mailto:fieldtrip-bounces at science.ru.nl] On Behalf Of Ingrid Nieuwenhuis Sent: 05 December 2012 19:09 To: fieldtrip at science.ru.nl Subject: Re: [FieldTrip] Partial artifact rejection Hi Patrick, What you can do is use partial artifact rejection by calling ft_rejectarttifact with cfg.artfctdef.reject = 'partial'. Then subsequently loop over the trials you get, and identify the ones (by looking at data.time{iTr}) that do not start at the time your trials normally start (this means it's a trial after an artifact). Finally, exclude those by using ft_redefinetrials. For instance, if your trials start with t=0, then after ft_rejectartifact you can do something like: ind_keep = zeros(length(data.trial),1); %allocate for iTrl = 1:length(data.trial) if data.trial{iTrl}(1) ~= 0 %this means it's a trial occurring AFTER an artifact ind_keep(iTrl) = false; else %this means it's a trial occurring BEFORE an artifact ind_keep(iTrl) = true; end end %only keep the trials before the artifact: cfg = []; cfg.trials = find(ind_keep); data = ft_redefinetrial(cfg, data); Hope this helps, Ingrid On 12/5/2012 5:57 AM, Jung, Patrick wrote: Dear all, in the reference documentation it says "In case of partial artifact rejection, a minimum length of the resulting sub-trials can be specified." I'd like to reject the trial segment from the beginning of the artifact onwards to keep only the data before the artifact occurred. But I don't want to append subtrials. How can I do that in fieldtrip? Many thanks for your help! Best, Patrick _______________________________________________ fieldtrip mailing list fieldtrip at donders.ru.nl http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Ingrid Nieuwenhuis PhD Postdoctoral Fellow Sleep and Neuroimaging Laboratory Department of Psychology University of California, Berkeley California 94720-1650 Tolman Hall, room 5305 -------------- next part -------------- An HTML attachment was scrubbed... URL: From f.roux at bcbl.eu Thu Dec 6 10:05:56 2012 From: f.roux at bcbl.eu (Frederic Roux) Date: Thu, 06 Dec 2012 10:05:56 +0100 (CET) Subject: [FieldTrip] using the PALS B12 atlas with fieldtrip Message-ID: <1b7b74e7-7b97-49b0-aa33-ffe91bb12f10@thalamus_p> Dear all, I was wondering if anyone had some partical experience or an idea on how to use the PALS B12 atlas provided by the Surface Management System DataBase (Sumsd), to generate a template brain in fieldtrip which can then be used for source reconstruction. So far I am doing my source reconstruction based on individual MRs morphed onto the MNI template brain. But I'd be intersted in knowing if it would make a difference if I use a different template. I've seen publications were people have done this so I know that in principle it should be possible. I just haven't found a practical way to read in the atlas data with fieldtrip yet. The data can be found here: http://sumsdb.wustl.edu/sums/directory.do?id=6585200&dir_name=CARET_TUTORIAL_SEPT-06 If you have any thoughts or comments on this, I'd be happy if you could share them with me. Best, Fred Frédéric Roux Postdoctoral Researcher www.bcbl.eu From r.oostenveld at donders.ru.nl Thu Dec 6 22:44:54 2012 From: r.oostenveld at donders.ru.nl (Robert Oostenveld) Date: Thu, 6 Dec 2012 22:44:54 +0100 Subject: [FieldTrip] problem with reading .edf+ files In-Reply-To: <71D8F8A56F37A947912FC86D4D9F00F477C17C494F@XMAIL1.medads.uk-erlangen.de> References: <71D8F8A56F37A947912FC86D4D9F00F477C17C494F@XMAIL1.medads.uk-erlangen.de> Message-ID: <70267DB0-752C-42D7-B560-C4413588BEFD@donders.ru.nl> Hi Margit The EDF and derived formats (EDF+, BDF, GDF) write the data as short segments in the file. Usually the subsequent segments can be glued together to form a continuous representation, but that is not guaranteed by the fileformat. I cannot check at the moment, as I don' have an EDF file handy, but I would expect that if you specify cfg.continuous='yes' to ft_preprocessing that it will be interpreted as a single continuous recording, rather than as an epoched recording. best regards, Robert On 5 Dec 2012, at 17:07, Schönherr, Margit wrote: > Hello, > > I want to analyze invasive EEG data and I have problems reading the data correctly. The data are continuous and I used BESA to convert them to .edf format. When I read them into matlab, however, the data are not continuous anymore, but segmented into very short trials. For example, a data set of 1 hour is split into 8229 trials of 0.43 s length. Even more confusing, the trial length differs across subjects. > > What is the difference between .edf and .edf+? BESA Research produces the latter, could this be the reason for the issue? > I also tried to track down the place where the segmentation into trials is done, but I ended up at line 111 of read_edf.m (EDF.Dur = str2num(H1(245:252)) % 8 Byte # duration of data record in sec). > > I would be very glad if someone could tell me, how to read the data correctly with fieldtrip. Thanks! > > Best, > Margit > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From r.oostenveld at donders.ru.nl Thu Dec 6 22:58:37 2012 From: r.oostenveld at donders.ru.nl (Robert Oostenveld) Date: Thu, 6 Dec 2012 22:58:37 +0100 Subject: [FieldTrip] using the PALS B12 atlas with fieldtrip In-Reply-To: <1b7b74e7-7b97-49b0-aa33-ffe91bb12f10@thalamus_p> References: <1b7b74e7-7b97-49b0-aa33-ffe91bb12f10@thalamus_p> Message-ID: Hi Fred That is an intersting and timely question, as it is something that we are presently working on for the Human Connectome Project. So for some parts FieldTrip might already help you out, for some other parts there still might be some work to do. The idea is that you would start with FT_READ_ATLAS. That represents the output structure either in the format as described in FT_DATATYPE_PARCELLATION (if it is a surface description) or as FT_DATATYPE_SEGMENTATION (if it is a 3-D volumetric description). See the help of those functions. Since the atlas is for a template MRI, and the MEG source reconstruction needs to be done in individual coordinates (with the individual head model), you have to derive the spatial mapping from individual T1 to template/atlas T1, and then apply the inverse mapping on the template mesh. The individualized mesh can then be used as source space (i.e. into cfg.grid for FT_SOURCEANALYSIS). Once you have the source reconstruction on all vertices of of the mesh, you can call FT_SOURCEPARCELLATE (which is not in the release version yet, but a skelleton exists, see also http://bugzilla.fcdonders.nl/show_bug.cgi?id=1775) which collapses the estimated source data over all vertices within each parcel to a single parcel-estimate. We are considering other alternatives for the source reconstruction, but this is more or less the first approach that we want to enable. However, we are presently not working with an atlas sheet, but rather with individual parcellated cortical sheets obtained from Freesurfer. So the step with the warp from atlas/template space to individual head space is one that we don't have to do for the HCP. It would be nice if we could get an example script on the fieldtrip wiki that demonstrates this. I hope that these guidelines provide you with some starting points. If you see parts of functionality that are still missing in fieldtrip to get the full pipeline to work, please let us know (e.g. through bugzilla). cheers Robert On 6 Dec 2012, at 10:05, Frederic Roux wrote: > Dear all, > > I was wondering if anyone had some partical experience or > an idea on how to use the PALS B12 atlas provided by the Surface > Management System DataBase (Sumsd), to generate a template brain in > fieldtrip which can then be used for source reconstruction. > > So far I am doing my source reconstruction based on > individual MRs morphed onto the MNI template brain. > But I'd be intersted in knowing if it would make a > difference if I use a different template. > > I've seen publications were people have done this > so I know that in principle it should be possible. I just haven't > found a practical way to read in the atlas data with fieldtrip yet. > The data can be found here: > http://sumsdb.wustl.edu/sums/directory.do?id=6585200&dir_name=CARET_TUTORIAL_SEPT-06 > > If you have any thoughts or comments on this, I'd be > happy if you could share them with me. > > Best, > Fred > > > > Frédéric Roux > Postdoctoral Researcher > www.bcbl.eu > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From jdien07 at mac.com Fri Dec 7 05:27:22 2012 From: jdien07 at mac.com (Joseph Dien) Date: Thu, 06 Dec 2012 23:27:22 -0500 Subject: [FieldTrip] Maryland Neuroimaging Center Research Associate position Message-ID: See attached. -------------------------------------------------------------------------------- Joseph Dien, Senior Research Scientist University of Maryland E-mail: jdien07 at mac.com Phone: 301-226-8848 Fax: 301-226-8811 http://joedien.com// -------------- next part -------------- A non-text attachment was scrubbed... Name: MNC Research Associate Announcement.pdf Type: application/pdf Size: 90943 bytes Desc: not available URL: From Margit.Schoenherr at uk-erlangen.de Fri Dec 7 09:38:56 2012 From: Margit.Schoenherr at uk-erlangen.de (=?iso-8859-1?Q?Sch=F6nherr=2C_Margit?=) Date: Fri, 7 Dec 2012 09:38:56 +0100 Subject: [FieldTrip] problem with reading .edf+ files In-Reply-To: <70267DB0-752C-42D7-B560-C4413588BEFD@donders.ru.nl> References: <71D8F8A56F37A947912FC86D4D9F00F477C17C494F@XMAIL1.medads.uk-erlangen.de>, <70267DB0-752C-42D7-B560-C4413588BEFD@donders.ru.nl> Message-ID: <71D8F8A56F37A947912FC86D4D9F00F477C17C4950@XMAIL1.medads.uk-erlangen.de> Hello Robert, thank you for the easy solution, it works with cfg.continuous='yes'. Margit ________________________________________ Von: fieldtrip-bounces at science.ru.nl [fieldtrip-bounces at science.ru.nl] im Auftrag von Robert Oostenveld [r.oostenveld at donders.ru.nl] Gesendet: Donnerstag, 6. Dezember 2012 22:44 An: FieldTrip discussion list Betreff: Re: [FieldTrip] problem with reading .edf+ files Hi Margit The EDF and derived formats (EDF+, BDF, GDF) write the data as short segments in the file. Usually the subsequent segments can be glued together to form a continuous representation, but that is not guaranteed by the fileformat. I cannot check at the moment, as I don' have an EDF file handy, but I would expect that if you specify cfg.continuous='yes' to ft_preprocessing that it will be interpreted as a single continuous recording, rather than as an epoched recording. best regards, Robert On 5 Dec 2012, at 17:07, Schönherr, Margit wrote: > Hello, > > I want to analyze invasive EEG data and I have problems reading the data correctly. The data are continuous and I used BESA to convert them to .edf format. When I read them into matlab, however, the data are not continuous anymore, but segmented into very short trials. For example, a data set of 1 hour is split into 8229 trials of 0.43 s length. Even more confusing, the trial length differs across subjects. > > What is the difference between .edf and .edf+? BESA Research produces the latter, could this be the reason for the issue? > I also tried to track down the place where the segmentation into trials is done, but I ended up at line 111 of read_edf.m (EDF.Dur = str2num(H1(245:252)) % 8 Byte # duration of data record in sec). > > I would be very glad if someone could tell me, how to read the data correctly with fieldtrip. Thanks! > > Best, > Margit > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip _______________________________________________ fieldtrip mailing list fieldtrip at donders.ru.nl http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From polomacnenad at gmail.com Sun Dec 9 15:06:57 2012 From: polomacnenad at gmail.com (Nenad Polomac) Date: Sun, 9 Dec 2012 15:06:57 +0100 Subject: [FieldTrip] FT_REGRESSCONFOUND Message-ID: Hi, I would like to know does FT_REGRESSCONFOUND is able to correct planar gradient data for the head movement? All the best! Nenad -------------- next part -------------- An HTML attachment was scrubbed... URL: From akiko.ikkai at gmail.com Sun Dec 9 23:48:59 2012 From: akiko.ikkai at gmail.com (Akiko Ikkai) Date: Sun, 9 Dec 2012 17:48:59 -0500 Subject: [FieldTrip] classification group-level stats? Message-ID: Dear FieldTrip users, I'm running multivariate classification on my EEG data (subset of electrodes), following this page ( http://fieldtrip.fcdonders.nl/tutorial/multivariateanalysis?s[]=classification). I have a question regarding a group-level stats. I have classification results (accuracy and binomial p) for each of 16 subs' selected sensors (averaged across time and freq). I don't think I can simply run one-sample t-test on accuracy against .5 (chance) at group level, since at individual level, accuracy of .56 (for example) could be associated with p of .3 (for example), depending on number of trials (across subjects, number of trials that goes into SVM is more or less constant). This could be significantly higher than chance (.5) if I simply run one-sample t-test across subjects with just accuracy. Does anyone have suggestions on how to run group stats (comparing with chance-level) considering these factors? Thanks in advance! Akiko -- Akiko Ikkai, Ph.D. Postdoctoral Fellow Department of Psychological and Brain Sciences Johns Hopkins University Ames Hall, 3400 N. Charles St. Baltimore, MD 21218 -------------- next part -------------- An HTML attachment was scrubbed... URL: From a.stolk8 at gmail.com Mon Dec 10 08:55:00 2012 From: a.stolk8 at gmail.com (A.stolk) Date: Mon, 10 Dec 2012 08:55:00 +0100 Subject: [FieldTrip] Ft regressconfound Message-ID: Hi nenad Yes, you can use ft regressconfound on planar data. Planar data amplitude is also prone to distances of the head/sources to the sensor array. Ft regressconfound will deal with the variability of the signal due to head movements, on a trial by trial basis. Yours, arjen -------------- next part -------------- An HTML attachment was scrubbed... URL: From irina.simanova at mpi.nl Mon Dec 10 10:30:39 2012 From: irina.simanova at mpi.nl (Irina Simanova) Date: Mon, 10 Dec 2012 10:30:39 +0100 Subject: [FieldTrip] classification group-level stats? In-Reply-To: References: Message-ID: <3C87F388-7BA3-48BA-84FD-9A2C134625C7@mpi.nl> Dear Akiko, I think the t-test against 0.5 is still valid. The null hypothesis there will be that the classification accuracy across subjects has the mean of 0.5. If there is no class information in the data, you should see the accuracy of 0.44 as often as 0.56. Then the null hypothesis is not rejected. Many fMRI searchlight studies use this logic. Alternatively, you can use a binomial test. You first apply the individual-subject threshold (say, p less than 0.05), and then count in how many subjects out of your sample the classification is significant (say, 8 out of 10). You then test this number with the binomial test. So you compute the probability of this amount of "coin flips" to return 1, given that the probability of such an event is 0.05. Hope this helps, Best, Irina Op 09.12.2012 om 23:48 heeft Akiko Ikkai het volgende geschreven: > Dear FieldTrip users, > > I'm running multivariate classification on my EEG data (subset of electrodes), following this page (http://fieldtrip.fcdonders.nl/tutorial/multivariateanalysis?s[]=classification). I have a question regarding a group-level stats. > > I have classification results (accuracy and binomial p) for each of 16 subs' selected sensors (averaged across time and freq). I don't think I can simply run one-sample t-test on accuracy against .5 (chance) at group level, since at individual level, accuracy of .56 (for example) could be associated with p of .3 (for example), depending on number of trials (across subjects, number of trials that goes into SVM is more or less constant). This could be significantly higher than chance (.5) if I simply run one-sample t-test across subjects with just accuracy. > > Does anyone have suggestions on how to run group stats (comparing with chance-level) considering these factors? > > Thanks in advance! Akiko > > -- > Akiko Ikkai, Ph.D. > Postdoctoral Fellow > Department of Psychological and Brain Sciences > Johns Hopkins University > Ames Hall, 3400 N. Charles St. > Baltimore, MD 21218 > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -------------- next part -------------- An HTML attachment was scrubbed... URL: From marco.rotonda at gmail.com Mon Dec 10 11:26:17 2012 From: marco.rotonda at gmail.com (Marco Rotonda) Date: Mon, 10 Dec 2012 11:26:17 +0100 Subject: [FieldTrip] ROI selection in a GC analysis Message-ID: Dear fieldtrip experts, I'm sorry to ask again to the list the same question but I had no answer and I still have not solved my questions. I would like to ask you if there is an elegant solution in the selection/definition of the ROI for a Granger Causality analysis. Just to explain better my doubts. Let suppose i've the sources reconstructed in the temporal domain and/or in the frequencies domain. If now I wish to perform a GC analysis I should know the ROI or sources on which to perform it. As far as I've seen most of them are selected with an a priori knowledge. I would like to know if there is a way to define the ROI from the data, not from my (eventually with bias) knowledge. As far as I've searched the only method developed is the Full-brain auto-regressive modeling (FARM). http://www.ncbi.nlm.nih.gov/pubmed/21439388 Since there are already ft_mvaranalysis and ft_connectivityanalysis, do you think that is possible to integrate the FARM approach into those functions? I'm sorry but I know my limitations and I don't know how to do it. I hope this could be useful not for me only. Regards, Marco -------------- next part -------------- An HTML attachment was scrubbed... URL: From clara.scholl at gmail.com Tue Dec 11 15:22:24 2012 From: clara.scholl at gmail.com (Clara A. Scholl) Date: Tue, 11 Dec 2012 09:22:24 -0500 Subject: [FieldTrip] Using cfg.maskparameter with topoplotTFR Message-ID: How can I specify cfg.maskparameter to mask out some channels when plotting topoplots with ft_topoplotTFR? When I specify a vector the same length as the number of channels I'm plotting, I have an error with line 697 of ft_topoplotTFR and get "??? Argument to dynamic structure reference must evaluate to a valid field name." I can see that cfg.maskparameter should be a field of my data structure (datmask = data.(cfg.maskparameter)), but I don't understand what that would be. More generally, I want to plot scalp topographies with some channels masked out (made transparent, or set to zero/white where the colorbar is modified to be white at zero) BUT I don't want to actually set those values to zero, because doing so generates interpolation artifacts. What's the best strategy? The cluster plotting tutorials are very nice for showing the whole topography with a subset of channels highlighted, but I find that the presence of channel markers obscures the underlying topography. I'd really appreciate any guidance or references for how to mask out a part of the topography. Thanks, Clara From jm.horschig at donders.ru.nl Tue Dec 11 15:51:55 2012 From: jm.horschig at donders.ru.nl (=?ISO-8859-1?Q?=22J=F6rn_M=2E_Horschig=22?=) Date: Tue, 11 Dec 2012 15:51:55 +0100 Subject: [FieldTrip] Using cfg.maskparameter with topoplotTFR In-Reply-To: References: Message-ID: <50C7488B.9060804@donders.ru.nl> Dear Clara, My feeling is that you illspecified the mask parameter. cfg,maskparameter should indeed be a field that you have to add to your data. It should have an equal length than your number of channels and consist of 1s and 0s. It is not possible to partially mask (at the moment). We might think about adding some alpha blending some day, though. Also, please use the most recent FieldTrip version. Line 697 in the most recent version has nothing to do with masking. It might be a bug in an old version, which is fixed now.Hope it works for you with the most recent version for you, because for me it does ;) Your code should look something like this: data = ...; cfg = []; cfg.layout = ...; cfg.xlim = [5 10]; cfg.zlim = [0.08 0.2]; data.mask = ones(size(data.label)); data.mask(1:end/2) = 0; % I made this somethign random for me cfg.maskparameter = 'mask'; ft_topoplotER(cfg, data); Should you keep on having problems, feel free to ask. Best, Jörn PS: It's generally consider nice to say 'hello' before asking for help :) On 12/11/2012 3:22 PM, Clara A. Scholl wrote: > How can I specify cfg.maskparameter to mask out some channels when > plotting topoplots with ft_topoplotTFR? When I specify a vector the > same length as the number of channels I'm plotting, I have an error > with line 697 of ft_topoplotTFR and get "??? Argument to dynamic > structure reference must evaluate to a valid field name." I can see > that cfg.maskparameter should be a field of my data structure (datmask > = data.(cfg.maskparameter)), but I don't understand what that would > be. > > More generally, I want to plot scalp topographies with some channels > masked out (made transparent, or set to zero/white where the colorbar > is modified to be white at zero) BUT I don't want to actually set > those values to zero, because doing so generates interpolation > artifacts. What's the best strategy? The cluster plotting tutorials > are very nice for showing the whole topography with a subset of > channels highlighted, but I find that the presence of channel markers > obscures the underlying topography. I'd really appreciate any > guidance or references for how to mask out a part of the topography. > > Thanks, > Clara > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Jörn M. Horschig PhD Student Donders Institute for Brain, Cognition and Behaviour Centre for Cognitive Neuroimaging Radboud University Nijmegen Neuronal Oscillations Group FieldTrip Development Team P.O. Box 9101 NL-6500 HB Nijmegen The Netherlands Contact: E-Mail: jm.horschig at donders.ru.nl Tel: +31-(0)24-36-68493 Web: http://www.ru.nl/donders Visiting address: Trigon, room 2.30 Kapittelweg 29 NL-6525 EN Nijmegen The Netherlands -------------- next part -------------- An HTML attachment was scrubbed... URL: From clara.scholl at gmail.com Tue Dec 11 16:08:50 2012 From: clara.scholl at gmail.com (Clara A. Scholl) Date: Tue, 11 Dec 2012 10:08:50 -0500 Subject: [FieldTrip] Using cfg.maskparameter with topoplotTFR In-Reply-To: <50C7488B.9060804@donders.ru.nl> References: <50C7488B.9060804@donders.ru.nl> Message-ID: Dear Jörn, Thank you so much for the helpful clarification and for your very rapid reply. I misunderstood how to specify the mask appropriately -- in the data structure and not in the configuration structure -- and the mask works when I follow your directions. Thank you also for the reminder to use the most recent version of FieldTrip, I was looking at a version from Feb. 2012. Thanks again, Clara On Tue, Dec 11, 2012 at 9:51 AM, "Jörn M. Horschig" wrote: > Dear Clara, > > My feeling is that you illspecified the mask parameter. cfg,maskparameter > should indeed be a field that you have to add to your data. It should have > an equal length than your number of channels and consist of 1s and 0s. It is > not possible to partially mask (at the moment). We might think about adding > some alpha blending some day, though. > Also, please use the most recent FieldTrip version. Line 697 in the most > recent version has nothing to do with masking. It might be a bug in an old > version, which is fixed now.Hope it works for you with the most recent > version for you, because for me it does ;) > > Your code should look something like this: > data = ...; > cfg = []; > cfg.layout = ...; > cfg.xlim = [5 10]; > cfg.zlim = [0.08 0.2]; > > data.mask = ones(size(data.label)); > data.mask(1:end/2) = 0; % I made this somethign random for me > cfg.maskparameter = 'mask'; > > ft_topoplotER(cfg, data); > > Should you keep on having problems, feel free to ask. > > Best, > Jörn > > PS: It's generally consider nice to say 'hello' before asking for help :) > > > > On 12/11/2012 3:22 PM, Clara A. Scholl wrote: > > How can I specify cfg.maskparameter to mask out some channels when > plotting topoplots with ft_topoplotTFR? When I specify a vector the > same length as the number of channels I'm plotting, I have an error > with line 697 of ft_topoplotTFR and get "??? Argument to dynamic > structure reference must evaluate to a valid field name." I can see > that cfg.maskparameter should be a field of my data structure (datmask > = data.(cfg.maskparameter)), but I don't understand what that would > be. > > More generally, I want to plot scalp topographies with some channels > masked out (made transparent, or set to zero/white where the colorbar > is modified to be white at zero) BUT I don't want to actually set > those values to zero, because doing so generates interpolation > artifacts. What's the best strategy? The cluster plotting tutorials > are very nice for showing the whole topography with a subset of > channels highlighted, but I find that the presence of channel markers > obscures the underlying topography. I'd really appreciate any > guidance or references for how to mask out a part of the topography. > > Thanks, > Clara > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > > > > -- > Jörn M. Horschig > PhD Student > Donders Institute for Brain, Cognition and Behaviour > Centre for Cognitive Neuroimaging > Radboud University Nijmegen > Neuronal Oscillations Group > FieldTrip Development Team > > P.O. Box 9101 > NL-6500 HB Nijmegen > The Netherlands > > Contact: > E-Mail: jm.horschig at donders.ru.nl > Tel: +31-(0)24-36-68493 > Web: http://www.ru.nl/donders > > Visiting address: > Trigon, room 2.30 > Kapittelweg 29 > NL-6525 EN Nijmegen > The Netherlands > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From alexander.opitz at tuebingen.mpg.de Tue Dec 11 16:57:38 2012 From: alexander.opitz at tuebingen.mpg.de (Alexander Opitz) Date: Tue, 11 Dec 2012 16:57:38 +0100 Subject: [FieldTrip] inconsistent tutorial data? Message-ID: Hi, I am going through the tutorial visual artifact rejection and I am getting an error after dummy = ft_rejectvisual(cfg,dataFIC); ??? Error using ==> ft_chantype at 61 the input that was provided to this function cannot be deciphered Error in ==> channelposition at 66 sel = ft_chantype(sens, 'meg'); Error in ==> ft_datatype_sens at 106 [chanpos, chanori, chanlab] = channelposition(sens, 'channel', 'all'); Error in ==> ft_datatype_raw at 99 data.grad = ft_datatype_sens(data.grad); Error in ==> ft_checkdata at 212 data = ft_datatype_raw(data, 'hassampleinfo', hassampleinfo); Error in ==> ft_rejectvisual at 181 data = ft_checkdata(data, 'datatype', 'raw', 'feedback', 'yes', 'hassampleinfo', 'yes'); After commenting out the ft_checkdata it works but then I am running into problems using the channel wise display. It seems to me that the way the channel information is stored in the data structure has changed but has not been updated in the tutorial data. I am using the latest fieldtrip-20121210 version on Matlab R2011a, 64bit Ubuntu. Best, Alex From ian.gould at psy.ox.ac.uk Tue Dec 11 18:06:29 2012 From: ian.gould at psy.ox.ac.uk (Ian Gould) Date: Tue, 11 Dec 2012 17:06:29 +0000 Subject: [FieldTrip] error using ft_combineplanar() after spm2fieldtrip() Message-ID: I'm getting the following error message when trying to use ft_combineplaner(): >Error using ft_checkdata (line 406) >This function requires ctf151_planar, ctf275_planar, neuromag122, neuromag306, >bti248_planar, bti148_planar, itab153_planar, yokogawa160_planar, yokogawa64_planar or >yokogawa440_planar data as input, but you are giving neuromag306 data. Obviously, this error message doesn't make complete sense…. The problem seems to be due to the fact that my data were preprocessed using SPM and converted to fieldtrip format using spm2fieldtrip(). Spm2fieldtrip() seems to spit out data structures compatible with the older fieldtrip data format, which has the following subfields, and lacks fields like 'grad' which are present after preprocessing the data wholly using recent fieldtrip releases: avg: [338x1501 double] var: [338x1501 double] time: [1x1501 double] dof: [338x1501 double] label: {338x1 cell} dimord: 'chan_time' cfg: [1x1 struct] I can circumvent this error for now with the quick/easy hack of editing ft_checkdata line 387 to accept 'avg' data so it reads: >if isfield(data, 'grad') || isfield(data, 'elec') || isfield(data, 'avg') This results in ft_combineplanar() and subsequent data topo/muliplotting working fine, but I'm not sure whether the missing 'grad' data would be troublesome in other situations, in which case spm2fieldtrip(), ft_combineplanar() or ft_checkdata() might need tweaking? Any advice on whether I'm doing something wrong here and/or if this is readily fixable would be greatly appreciated! I'm using the latest fieldtrip release (10/12/12) and SPM8. Thanks, Ian ------------------- Ian Gould, PhD Brain & Cognition Laboratory Department of Experimental Psychology University of Oxford South Parks Road Oxford, OX1 3UD -------------- next part -------------- An HTML attachment was scrubbed... URL: From marco.rotonda at gmail.com Tue Dec 11 21:41:22 2012 From: marco.rotonda at gmail.com (Marco Rotonda) Date: Tue, 11 Dec 2012 21:41:22 +0100 Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: References: Message-ID: Dear fieldtrip experts, I'm sorry to ask again to the list the same question but I had no answer and I still have not solved my questions. I would like to ask you if there is an elegant solution in the selection/definition of the ROI for a Granger Causality analysis. Just to explain better my doubts. Let suppose i've the sources reconstructed in the temporal domain and/or in the frequencies domain. If now I wish to perform a GC analysis I should know the ROI or sources on which to perform it. As far as I've seen most of them are selected with an a priori knowledge. I would like to know if there is a way to define the ROI from the data, not from my (eventually with bias) knowledge. As far as I've searched the only method developed is the Full-brain auto-regressive modeling (FARM). http://www.ncbi.nlm.nih.gov/pubmed/21439388 Since there are already ft_mvaranalysis and ft_connectivityanalysis, do you think that is possible to integrate the FARM approach into those functions? I'm sorry but I know my limitations and I don't know how to do it. I hope this could be useful not for me only. Regards, Marco -------------- next part -------------- An HTML attachment was scrubbed... URL: From a.stolk at fcdonders.ru.nl Tue Dec 11 22:17:57 2012 From: a.stolk at fcdonders.ru.nl (Stolk, A.) Date: Tue, 11 Dec 2012 22:17:57 +0100 (CET) Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: Message-ID: <1188854412.105779.1355260677094.JavaMail.root@sculptor.zimbra.ru.nl> Dear Marco, Just a quick thought; would it be appropriate to select your ROI by using the cluster(s) that 'survived' non-parametric testing (i.e. using the .mask field that points to statistically significant voxels/sensors) ? Or perhaps based on previous findings in literature or by yourself? H ave a look at ft_volumelookup for that matter . Perhaps someone else has more experience facing the issue of ROI definition for granger purposes. With respect to your question regarding 'FARM'; I do not know whether it is on the menu or going to be. If you happen to have some lines of code for it, we'd be happy to implement it of course. Hope this was of any help, Arjen ----- Oorspronkelijk bericht ----- > Van: "Marco Rotonda" > Aan: "fieldtrip" > Verzonden: Dinsdag 11 december 2012 21:41:22 > Onderwerp: [FieldTrip] ROI selection in a GC analysis > Dear fieldtrip experts, > I'm sorry to ask again to the list the same question but I had no > answer and I still have not solved my questions. > I would like to ask you if there is an elegant solution in the > selection/definition of the ROI for a Granger Causality analysis. > Just to explain better my doubts. > Let suppose i've the sources reconstructed in the temporal domain > and/or in the frequencies domain. > If now I wish to perform a GC analysis I should know the ROI or > sources on which to perform it. > As far as I've seen most of them are selected with an a priori > knowledge. > I would like to know if there is a way to define the ROI from the > data, not from my (eventually with bias) knowledge. > As far as I've searched the only method developed is the Full-brain > auto-regressive modeling (FARM). > http://www.ncbi.nlm.nih.gov/pubmed/21439388 > Since there are already ft_mvaranalysis and ft_connectivityanalysis, > do > you think that is possible to integrate the FARM approach into those > functions? > I'm sorry but I know my limitations and I don't know how to do it. > I hope this could be useful not for me only. > Regards, > Marco > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -------------- next part -------------- An HTML attachment was scrubbed... URL: From jm.horschig at donders.ru.nl Wed Dec 12 09:47:37 2012 From: jm.horschig at donders.ru.nl (=?ISO-8859-1?Q?=22J=F6rn_M=2E_Horschig=22?=) Date: Wed, 12 Dec 2012 09:47:37 +0100 Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: References: Message-ID: <50C844A9.6080002@donders.ru.nl> Dear Marco, I once played around with that as well and tried different approaches. I think the cleanest approach is to define some region of interest based on anatomy and current literature and take the neighbouring peak voxel for each subject. In a similar vein as SPM is doing it, you can first spatially blur your source reconstructed activity if you like. Of course, instead of taking the initial ROI from literature, you can also do it as Arjen suggested take the ROI from some cluster analysis. Anyway, by making it subject-specific by taking the peak voxel per subject and blurring it, you will end up with a somewhat consistent region per subject and will be less sensitive to inter-subject variability. Unfortunately, you need to write most of this yourself. There are tal2mni and mni2tal functions on the internet for transforming between these two coordinate systems. Note that this is only a rough transformation and might not 100% correct. For blurring (or smoothing if you want to make it sound fancy) you can use spm_smooth. Note that this function is does work like the usual Matlab function in that it does not produce an output argument but rather changes the input argument inherently (in more technical terms: call-by-reference rather than call-by-value). Good luck! :) Best, Jörn On 12/11/2012 9:41 PM, Marco Rotonda wrote: > Dear fieldtrip experts, > > I'm sorry to ask again to the list the same question but I had no > answer and I still have not solved my questions. > > I would like to ask you if there is an elegant solution in the > selection/definition of the ROI for a Granger Causality analysis. > > Just to explain better my doubts. > Let suppose i've the sources reconstructed in the temporal domain > and/or in the frequencies domain. > If now I wish to perform a GC analysis I should know the ROI or > sources on which to perform it. > As far as I've seen most of them are selected with an a priori knowledge. > > I would like to know if there is a way to define the ROI from the > data, not from my (eventually with bias) knowledge. > > As far as I've searched the only method developed is the Full-brain > auto-regressive modeling (FARM). > http://www.ncbi.nlm.nih.gov/pubmed/21439388 > > Since there are already ft_mvaranalysis and ft_connectivityanalysis, do > you think that is possible to integrate the FARM approach into those > functions? > > I'm sorry but I know my limitations and I don't know how to do it. > I hope this could be useful not for me only. > > Regards, > > Marco > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Jörn M. Horschig PhD Student Donders Institute for Brain, Cognition and Behaviour Centre for Cognitive Neuroimaging Radboud University Nijmegen Neuronal Oscillations Group FieldTrip Development Team P.O. Box 9101 NL-6500 HB Nijmegen The Netherlands Contact: E-Mail: jm.horschig at donders.ru.nl Tel: +31-(0)24-36-68493 Web: http://www.ru.nl/donders Visiting address: Trigon, room 2.30 Kapittelweg 29 NL-6525 EN Nijmegen The Netherlands -------------- next part -------------- An HTML attachment was scrubbed... URL: From jm.horschig at donders.ru.nl Wed Dec 12 10:03:15 2012 From: jm.horschig at donders.ru.nl (=?ISO-8859-1?Q?=22J=F6rn_M=2E_Horschig=22?=) Date: Wed, 12 Dec 2012 10:03:15 +0100 Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: <50C844A9.6080002@donders.ru.nl> References: <50C844A9.6080002@donders.ru.nl> Message-ID: <50C84853.1050309@donders.ru.nl> On 12/12/2012 9:47 AM, "Jörn M. Horschig" wrote: > Note that this function is does work like the usual Matlab function in > that it does not produce an output argument but rather changes the > input argument inherently (in more technical terms: call-by-reference > rather than call-by-value). It should of course say it does *not* work like a usual Matlab function :) you can call it like this: / spm_smooth(srcmatrix, srcmatrix, 1); % call-by-reference , 2x2x2 gaussian kernel, 3D matrix as input /Again, note that srcmatrix will be changed afterwards, no way to revert it back. Best, Jörn -- Jörn M. Horschig PhD Student Donders Institute for Brain, Cognition and Behaviour Centre for Cognitive Neuroimaging Radboud University Nijmegen Neuronal Oscillations Group FieldTrip Development Team P.O. Box 9101 NL-6500 HB Nijmegen The Netherlands Contact: E-Mail: jm.horschig at donders.ru.nl Tel: +31-(0)24-36-68493 Web: http://www.ru.nl/donders Visiting address: Trigon, room 2.30 Kapittelweg 29 NL-6525 EN Nijmegen The Netherlands -------------- next part -------------- An HTML attachment was scrubbed... URL: From jan.schoffelen at donders.ru.nl Wed Dec 12 10:28:59 2012 From: jan.schoffelen at donders.ru.nl (jan-mathijs schoffelen) Date: Wed, 12 Dec 2012 10:28:59 +0100 Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: <50C84853.1050309@donders.ru.nl> References: <50C844A9.6080002@donders.ru.nl> <50C84853.1050309@donders.ru.nl> Message-ID: Hi Marco, You could also have a look at this paper. Investigating causality between interacting brain areas with multivariate autoregressive models of MEG sensor data. Michalareas G, Schoffelen JM, Paterson G, Gross J. Hum Brain Mapp. 2012 Feb 13. doi: 10.1002/hbm.21482. [Epub ahead of print] PMID: 22328419 [PubMed - as supplied by publisher] Best, Jan-Mathijs On Dec 12, 2012, at 10:03 AM, Jörn M. Horschig wrote: > On 12/12/2012 9:47 AM, "Jörn M. Horschig" wrote: >> Note that this function is does work like the usual Matlab function in that it does not produce an output argument but rather changes the input argument inherently (in more technical terms: call-by-reference rather than call-by-value). > > It should of course say it does *not* work like a usual Matlab function :) you can call it like this: > > spm_smooth(srcmatrix, srcmatrix, 1); % call-by-reference , 2x2x2 gaussian kernel, 3D matrix as input > > Again, note that srcmatrix will be changed afterwards, no way to revert it back. > > Best, > Jörn > > > > > -- > Jörn M. Horschig > PhD Student > Donders Institute for Brain, Cognition and Behaviour > Centre for Cognitive Neuroimaging > Radboud University Nijmegen > Neuronal Oscillations Group > FieldTrip Development Team > > P.O. Box 9101 > NL-6500 HB Nijmegen > The Netherlands > > Contact: > E-Mail: jm.horschig at donders.ru.nl > Tel: +31-(0)24-36-68493 > Web: http://www.ru.nl/donders > > Visiting address: > Trigon, room 2.30 > Kapittelweg 29 > NL-6525 EN Nijmegen > The Netherlands > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip Jan-Mathijs Schoffelen, MD PhD Donders Institute for Brain, Cognition and Behaviour, Centre for Cognitive Neuroimaging, Radboud University Nijmegen, The Netherlands Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands J.Schoffelen at donders.ru.nl Telephone: +31-24-3614793 -------------- next part -------------- An HTML attachment was scrubbed... URL: From sdnewman at indiana.edu Wed Dec 12 13:24:47 2012 From: sdnewman at indiana.edu (sdnewman at indiana.edu) Date: Wed, 12 Dec 2012 07:24:47 -0500 Subject: [FieldTrip] new to EEG and fieldtrip Message-ID: <20121212072447.g92kzgqds8so8gkc@webmail.iu.edu> Hello, I am completely new to EEG data analysis and to fieldtrip. I am using a 256 electrode EGI system and trying to simply read in data collected by someone else. First, how can you determine what format the data was saved in? There are no file extensions. Second, I did get my hands on a .raw file but had problems reading in the data. It ran into a memory error. Is there a problem using a windows machine to run this analysis? Thanks in advance for your help. Sharlene From smoratti at psi.ucm.es Wed Dec 12 14:23:15 2012 From: smoratti at psi.ucm.es (smoratti at psi.ucm.es) Date: Wed, 12 Dec 2012 14:23:15 +0100 Subject: [FieldTrip] new to EEG and fieldtrip In-Reply-To: <20121212072447.g92kzgqds8so8gkc@webmail.iu.edu> References: <20121212072447.g92kzgqds8so8gkc@webmail.iu.edu> Message-ID: Dear Sharlene, You will have to read epochs, so you don't run into memory problems. On the field trip webpage there is a great tutorial that describes step by step how to define trials, read data and preprocess them. You may adapt it to the EGI data format. See http://fieldtrip.fcdonders.nl/dataformat to see what data format FT can read (EGI is among them). See also on this webpage what extensions your file needs. Best, Stephan ________________________________________________________ Stephan Moratti, PhD see also: http://web.me.com/smoratti/ Universidad Complutense de Madrid Facultad de Psicología Departamento de Psicología Básica I Campus de Somosaguas 28223 Pozuelo de Alarcón (Madrid) Spain and Center for Biomedical Technology Laboratory for Cognitive and Computational Neuroscience Parque Científico y Tecnológico de la Universidad Politecnica de Madrid Campus Montegancedo 28223 Pozuelo de Alarcón (Madrid) Spain email: smoratti at psi.ucm.es Tel.: +34 679219982 El 12/12/2012, a las 13:24, sdnewman at indiana.edu escribió: > Hello, > I am completely new to EEG data analysis and to fieldtrip. I am using a 256 electrode EGI system and trying to simply read in data collected by someone else. First, how can you determine what format the data was saved in? There are no file extensions. Second, I did get my hands on a .raw file but had problems reading in the data. It ran into a memory error. Is there a problem using a windows machine to run this analysis? Thanks in advance for your help. > > Sharlene > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -------------- next part -------------- An HTML attachment was scrubbed... URL: From b.osuagwu.1 at research.gla.ac.uk Thu Dec 13 00:39:34 2012 From: b.osuagwu.1 at research.gla.ac.uk (Bethel Osuagwu) Date: Wed, 12 Dec 2012 23:39:34 +0000 Subject: [FieldTrip] MEG-EMG coherence Message-ID: <7AB2A399CAFFA442A6262BF9939AC33B5BDE8CD003@CMS07.campus.gla.ac.uk> Dear all, I am new to FieldTrip. Is FieldTrip limited to MEG-EMG coherence computation? I want to use FieldTrip to compute EEG-EMG coherence. The EEG and EMG data was recorded with Biosig. What procedure do I follow to obtain the coherence. Any help will be highly appreciated. Thank you very much Bethel From zrhzrh at mail.ustc.edu.cn Fri Dec 14 04:15:26 2012 From: zrhzrh at mail.ustc.edu.cn (Rui-huai Zhang) Date: Fri, 14 Dec 2012 11:15:26 +0800 (CST) Subject: [FieldTrip] How to topoplot some custom values? Message-ID: <31452433.1869421355454926264.JavaMail.coremail@mailweb> Hi, All! I wanna know how to make a topoplot that the value of each channel is made by a custom function? In another word: I got a real value of each channel, and how could I make a topographic map? Thanks! -------------- next part -------------- An HTML attachment was scrubbed... URL: From Hanneke.vanDijk at med.uni-duesseldorf.de Fri Dec 14 07:05:58 2012 From: Hanneke.vanDijk at med.uni-duesseldorf.de (Hanneke van Dijk) Date: Fri, 14 Dec 2012 07:05:58 +0100 Subject: [FieldTrip] How to topoplot some custom values? In-Reply-To: <31452433.1869421355454926264.JavaMail.coremail@mailweb> References: <31452433.1869421355454926264.JavaMail.coremail@mailweb> Message-ID: Hi Zhang, You could add a field to an existing structure variable produced by either ft_timelockanalysis or freqanalysis. Let's say you call it "real". This field should has to contain at least a 2d matrix with the order channel_time. You could also make a dummy variable that looks like this. When calling ft_topoplotER or ft_topoplotTFR you should then use cfg.zparam = 'real'. Alternatively you can replace the data in the .pospctrm (for freq data) or .avg field ( for timelockdata ) with your data and use the default cfg.zparam when calling the plotting function. Hope this helps! Hanneke Op 14 dec. 2012 04:16 schreef "Rui-huai Zhang" het volgende: > > > Hi, All! > > I wanna know how to make a topoplot that the value of each channel is > made by a custom function? In another word: I got a real value of each > channel, and how could I make a topographic map? > > Thanks! > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -------------- next part -------------- An HTML attachment was scrubbed... URL: From akolchin at indiana.edu Fri Dec 14 14:31:19 2012 From: akolchin at indiana.edu (Artemy Kolchinsky) Date: Fri, 14 Dec 2012 13:31:19 +0000 Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) Message-ID: Hi all, I have a question about the cluster-based statistical testing in Fieldtrip, and also as described in Maris & Oostenveld's "Nonparametric statistical testing of EEG-and MEG-data". As far as I understand, the 'maxsum' statistic (as implemented in https://code.google.com/p/fieldtrip/source/browse/trunk/private/clusterstat.m) does the following: 1) Thresholds the t-statistic image (I just refer to t-statistics here for simplicity, I realize it can be other things also) above a 'non-corrected' threshold 2) Sums the t-statistics in each contiguous cluster (the 'cluster-mass') of values that exceed the threshold 3) Does a random-permutation-based null distribution of the 'maximum cluster-mass' statistic to test the significance of actually observed cluster-masses However, from reading about cluster-mass statistics in other places (such as fMRI-based literature, e.g. Bullmore's http://www.ai.mit.edu/events/talks/fMRI/papers/permutation_tests2.pdf [see Section G] , recent work on Random Field Theory of cluster masses http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2739659/ , etc.), each cluster's mass is actually taken to be the sum of the statistics *above* the non-corrected threshold (ie, the 'sum of suprathreshold statistics'). In other words, let's say our t-statistic for electrode i is t_i and the non-corrected threshold is c_i . Then, if I am correct, Fieldtrip does: \sum_i t_i where t_i > c_i whereas other literature suggests \sum_i (t_i - c_i) where t_i > c_i It seems like these would be testing different things --- for example, I suspect Fieldtrip's method, versus 'Bullmore's method', would disadvantage clusters that have small spatial support. To further complicate matters, from scanning Fieldtrip's code it seems that using the 'wcm' clusterstatistic option with wcm_weight = 1 would implement suprathreshold summing, but this option is not documented. Basically, my question is if there is a particular reason for these discrepancies, and also if there is any opinion on which cluster-based test 'works better' (for lack of a better criterion). Thanks greatly for any help, -a -------------- next part -------------- An HTML attachment was scrubbed... URL: From zrhzrh at mail.ustc.edu.cn Fri Dec 14 15:10:29 2012 From: zrhzrh at mail.ustc.edu.cn (Rui-huai Zhang) Date: Fri, 14 Dec 2012 22:10:29 +0800 (CST) Subject: [FieldTrip] How to topoplot some custom values? Message-ID: <9737488.1930371355494229598.JavaMail.coremail@mailweb> Hello Hanneke, Thank you for your reply! I'll have a try! Rui-Huai Zhang > Hi Zhang, > You could add a field to an existing structure variable produced by either ft_timelockanalysis or freqanalysis. Let's say you call it "real". This field should has to contain at least a 2d matrix with the order channel_time. You could also make a dummy variable that looks like this. > When calling ft_topoplotER or ft_topoplotTFR you should then use cfg.zparam = 'real'. > Alternatively you can replace the data in the .pospctrm (for freq data) or .avg field ( for timelockdata ) with your data and use the default cfg.zparam when calling the plotting function. > Hope this helps! > Hanneke >  > Op 14 dec. 2012 04:16 schreef "Rui-huai Zhang" het volgende: >  > > > > > > Hi, All! > > > > > > I wanna know how to make a topoplot that the value of each channel is made by a custom function? In another word: I got a real value of each channel, and how could I make a topographic map? > > > > > > Thanks! > > _______________________________________________ > > fieldtrip mailing list > > fieldtrip at donders.ru.nl > > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > > From russgport at gmail.com Fri Dec 14 21:28:58 2012 From: russgport at gmail.com (Russell G Port) Date: Fri, 14 Dec 2012 15:28:58 -0500 Subject: [FieldTrip] using fieldtrip to analyze CED data where each channel corresponds to a different channel Message-ID: Hi All, I very new to fieldtrip (as in 2 days), and I was just wondering what peoples would do if it were them. Basically I have spike2 files, where each channel corresponds to a separate animal. The problem comes into play because we have multiple sessions, each with a different drug. Currently the channels are labelled 'animalnumber_genotype'. To get the data to append the different days I would use ft_appenddata. Though I have a feeling because the channels have the same name on different days, and the order of drug is randomized, it would be not so smooth. I think the best way for me to get around this is changing the channel to 'animalnumber_genotype_drug'. Is this correct? I would do this because then when I appended them I would have a channel per animal per drug. I would then use a design matrix for a paired comparison on the statistics? Does this sound feasible? There may be the chance that I am screwing up some of the analysis since fieldtrip thinks these are just channels and not separate subjects, so I wanted to check. Thanks Russ Port -------------- next part -------------- An HTML attachment was scrubbed... URL: From m.chait at ucl.ac.uk Mon Dec 17 12:57:50 2012 From: m.chait at ucl.ac.uk (Chait, Maria) Date: Mon, 17 Dec 2012 11:57:50 +0000 Subject: [FieldTrip] PhD Position at UCL Ear Institute Message-ID: <3BA3DF582C0B7542AE0CB625F0119AB816FD867E@AMSPRD0104MB100.eurprd01.prod.exchangelabs.com> Dear Colleagues, I am writing to point your attention to PhD studentship (available for an immediate start) opening at the UCL Ear Institute and would be grateful if you could distribute the advert to relevant members of your institution. A 3 year PhD studentship in auditory cognitive neuroscience is available as part of a research collaboration between the UCL Ear Institute (London, UK) and NTT Communication Science Labs (Nippon Telegraph and Telephone corporation, Atsugi, Japan). The student will be based at the UCL Ear Institute and supervised by Dr. Maria Chait. They will also be working with Prof. Makio Kashino and Dr. Shigeto Furukawa (NTT) and the research program will involve experiments conducted both at UCL and in Japan, requiring occasional travel. The project will use psychophysics, eye tracking, autonomic response measures and MEG functional brain imaging to investigate which features of sound are perceptually salient. Namely, those sounds that automatically capture attention in a busy scene, even when listeners' initial perceptual focus is elsewhere. The UCL Ear Institute provides state-of-the-art research facilities across a wide range of disciplines and is one of the foremost centres for hearing-related research within Europe. Key Requirements The position is only available for UK/EU passport holders. Applicants should hold a 1St class, or upper 2nd bachelor's degree (masters degree an advantage) in an engineering or scientific subject. Previous experience with auditory research, functional brain imaging, neuroscience and/or acoustics is desirable. For an informal discussion, or to submit an application please contact Dr. Maria Chait (m.chait at ucl.ac.uk). Applicants should submit a supporting statement, a CV, and the details of two recommenders. Application review begins December 2012 and will continue until the position is filled. The studentship includes fees and a yearly stipend (about £16000; tax free). Maria Chait PhD m.chait at ucl.ac.uk Senior Lecturer UCL Ear Institute 332 Gray's Inn Road London WC1X 8EE -------------- next part -------------- An HTML attachment was scrubbed... URL: From e.maris at psych.ru.nl Mon Dec 17 16:55:06 2012 From: e.maris at psych.ru.nl (Eric Maris) Date: Mon, 17 Dec 2012 16:55:06 +0100 (CET) Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) In-Reply-To: References: Message-ID: <00b201cddc6e$e042cd40$a0c867c0$@maris@psych.ru.nl> Dear Artemy, Keep in mind that in the "Bullmore-style" cluster-mass statistic \sum_i (t_i - c_i) where t_i > c_i the thresholds c_i are constants (i.e., independent of the data). Importantly, these constants also enter in the permutation distribution that is used to evaluated the significance of the maximum cluster-mass statistic, to the effect that the Bullmore-style and the Fieldtrip-style permutation distributions are shifted versions of each other. As a result, the p-values that roll out of the two approaches are identical. Best, Eric Maris From: Artemy Kolchinsky [mailto:akolchin at indiana.edu] Sent: vrijdag 14 december 2012 14:31 To: fieldtrip at science.ru.nl Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) Hi all, I have a question about the cluster-based statistical testing in Fieldtrip, and also as described in Maris & Oostenveld's "Nonparametric statistical testing of EEG-and MEG-data". As far as I understand, the 'maxsum' statistic (as implemented in https://code.google.com/p/fieldtrip/source/browse/trunk/private/clustersta t.m) does the following: 1) Thresholds the t-statistic image (I just refer to t-statistics here for simplicity, I realize it can be other things also) above a 'non-corrected' threshold 2) Sums the t-statistics in each contiguous cluster (the 'cluster-mass') of values that exceed the threshold 3) Does a random-permutation-based null distribution of the 'maximum cluster-mass' statistic to test the significance of actually observed cluster-masses However, from reading about cluster-mass statistics in other places (such as fMRI-based literature, e.g. Bullmore's http://www.ai.mit.edu/events/talks/fMRI/papers/permutation_tests2.pdf [see Section G] , recent work on Random Field Theory of cluster masses http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2739659/ , etc.), each cluster's mass is actually taken to be the sum of the statistics *above* the non-corrected threshold (ie, the 'sum of suprathreshold statistics'). In other words, let's say our t-statistic for electrode i is t_i and the non-corrected threshold is c_i . Then, if I am correct, Fieldtrip does: \sum_i t_i where t_i > c_i whereas other literature suggests \sum_i (t_i - c_i) where t_i > c_i It seems like these would be testing different things --- for example, I suspect Fieldtrip's method, versus 'Bullmore's method', would disadvantage clusters that have small spatial support. To further complicate matters, from scanning Fieldtrip's code it seems that using the 'wcm' clusterstatistic option with wcm_weight = 1 would implement suprathreshold summing, but this option is not documented. Basically, my question is if there is a particular reason for these discrepancies, and also if there is any opinion on which cluster-based test 'works better' (for lack of a better criterion). Thanks greatly for any help, -a -------------- next part -------------- An HTML attachment was scrubbed... URL: From clme at tf.uni-kiel.de Mon Dec 17 17:22:23 2012 From: clme at tf.uni-kiel.de (clme at tf.uni-kiel.de) Date: Mon, 17 Dec 2012 17:22:23 +0100 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) Message-ID: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> Hello ft-users, Today I tried to plot planar gradients for neuromag306 data. When calling avgFICplanar = ft_megplanar(cfg, avgFIC); I get the following error: ??? Error using ==> ft_checkdata at 406 This function requires ctf151, ctf275, bti148, bti248, itab153, yokogawa160 or yokogawa64 data as input, but you are giving neuromag306 data. Error in ==> ft_megplanar at 106 data = ft_checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes', 'hassampleinfo', 'yes', 'ismeg', 'yes', 'senstype', {'ctf151', 'ctf275', 'bti148', 'bti248', 'itab153', 'yokogawa160', 'yokoga Error in ==> reading_fif_planar at 39 avgFCplanar = ft_megplanar(cfg, avgFC); When manually adding ",'neuromag306'" to the senstype in line 106 and commenting out line 297 erverything seems to work as it should. Is this a bug or am I doing something wrong? From wallisgj at gmail.com Mon Dec 17 17:37:09 2012 From: wallisgj at gmail.com (George Wallis) Date: Mon, 17 Dec 2012 16:37:09 +0000 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) In-Reply-To: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> References: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> Message-ID: <50CF4A35.1050404@gmail.com> Hi, The electa machine has 2 physical planar gradiometers at each sensor location, which are oriented orthogonally to one another (it also has a magnetometer at each sensor location). I believe ft_megplanar synthesises planar gradiometer data from, for example, data collected using axial gradiometers (I'm not sure - we've got an electa scanner, so I haven't used it!). But essentially, it's creating virtual sensors of a type the neuromag scanner already has in the hardware. Instead, you might want to take the two planar gradiometer sets, and combine them, so that for each sensor location you have a single gradiometer value (you'll still have the magnetometer data in addition). ft_combineplanar does this: it takes the signal from each gradiometer at each sensor location, and combines them into a single scalar signal. You can tweak the method it uses to do this, but in the time domain I think it just finds the cartesian sum (sqrt( s1^2 + s2^2) where s1 and s2 are the signals from each sensor). best wishes, George On 17/12/2012 16:22, clme at tf.uni-kiel.de wrote: > Hello ft-users, > > Today I tried to plot planar gradients for neuromag306 data. > > When calling avgFICplanar = ft_megplanar(cfg, avgFIC); > > I get the following error: > ??? Error using ==> ft_checkdata at 406 > This function requires ctf151, ctf275, bti148, bti248, itab153, > yokogawa160 or yokogawa64 data as input, but you are giving > neuromag306 data. > > Error in ==> ft_megplanar at 106 > data = ft_checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes', > 'hassampleinfo', 'yes', 'ismeg', 'yes', > 'senstype', {'ctf151', 'ctf275', 'bti148', 'bti248', 'itab153', > 'yokogawa160', 'yokoga > Error in ==> reading_fif_planar at 39 > avgFCplanar = ft_megplanar(cfg, avgFC); > > When manually adding ",'neuromag306'" to the senstype in line 106 and > commenting out line 297 erverything seems to work as it should. > > Is this a bug or am I doing something wrong? > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From maess at cbs.mpg.de Mon Dec 17 17:39:25 2012 From: maess at cbs.mpg.de (Burkhard Maess) Date: Mon, 17 Dec 2012 17:39:25 +0100 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) In-Reply-To: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> References: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> Message-ID: <50CF4ABD.3060701@cbs.mpg.de> Dear clme, you try to compute planar gradiometer signals for a system which measures planar gradiometer signals. This is why fieldtrip tries to stop you from doing it. So, instead of computing such signals you simply select the channels which have names according to this pattern: 'MEG...[23]' Channel names ending with an 2 or 3 are planar gradiometer channels. all the best, Burkhard clme at tf.uni-kiel.de wrote: > Hello ft-users, > > Today I tried to plot planar gradients for neuromag306 data. > > When calling avgFICplanar = ft_megplanar(cfg, avgFIC); > > I get the following error: > ??? Error using ==> ft_checkdata at 406 > This function requires ctf151, ctf275, bti148, bti248, itab153, > yokogawa160 or yokogawa64 data as input, but you are giving > neuromag306 data. > > Error in ==> ft_megplanar at 106 > data = ft_checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes', > 'hassampleinfo', 'yes', 'ismeg', 'yes', > 'senstype', {'ctf151', 'ctf275', 'bti148', 'bti248', 'itab153', > 'yokogawa160', 'yokoga > Error in ==> reading_fif_planar at 39 > avgFCplanar = ft_megplanar(cfg, avgFC); > > When manually adding ",'neuromag306'" to the senstype in line 106 and > commenting out line 297 erverything seems to work as it should. > > Is this a bug or am I doing something wrong? > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- ----- Dr. Burkhard Maess Max Planck Institute for Human Cognitive and Brain Sciences Stephanstr. 1a, P.O. Box 500355, D-04303 Leipzig Aussenstelle Bennewitz, phone/fax: +49(3425)8875-2526/-2511 mail: maess 'at' cbs.mpg.de, http://www.cbs.mpg.de From eelke.spaak at donders.ru.nl Mon Dec 17 17:42:11 2012 From: eelke.spaak at donders.ru.nl (Eelke Spaak) Date: Mon, 17 Dec 2012 17:42:11 +0100 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) In-Reply-To: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> References: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> Message-ID: Dear clme, ft_megplanar is designed to compute an approximation of planar magnetic gradient based on axial gradiometer data. This approximation does not make sense if the data you have is already reflecting the planar gradient, as is the case for neuromag306 data. Therefore, this is not a bug, but a feature. Best, Eelke On 17 December 2012 17:22, wrote: > Hello ft-users, > > Today I tried to plot planar gradients for neuromag306 data. > > When calling avgFICplanar = ft_megplanar(cfg, avgFIC); > > I get the following error: > ??? Error using ==> ft_checkdata at 406 > This function requires ctf151, ctf275, bti148, bti248, itab153, > yokogawa160 or yokogawa64 data as input, but you are giving > neuromag306 data. > > Error in ==> ft_megplanar at 106 > data = ft_checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes', > 'hassampleinfo', 'yes', 'ismeg', 'yes', > 'senstype', {'ctf151', 'ctf275', 'bti148', 'bti248', 'itab153', > 'yokogawa160', 'yokoga > Error in ==> reading_fif_planar at 39 > avgFCplanar = ft_megplanar(cfg, avgFC); > > When manually adding ",'neuromag306'" to the senstype in line 106 and > commenting out line 297 erverything seems to work as it should. > > Is this a bug or am I doing something wrong? > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From eelke.spaak at donders.ru.nl Mon Dec 17 17:44:17 2012 From: eelke.spaak at donders.ru.nl (Eelke Spaak) Date: Mon, 17 Dec 2012 17:44:17 +0100 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) In-Reply-To: References: <922b822de0425f73979a4d4c979be97e.squirrel@webmail.tf.uni-kiel.de> Message-ID: PS: You might want to take a look at ft_combineplanar, which *does* work for planar gradient data produced by hardware planar gradiometers (e.g. Neuromag systems). On 17 December 2012 17:42, Eelke Spaak wrote: > Dear clme, > > ft_megplanar is designed to compute an approximation of planar > magnetic gradient based on axial gradiometer data. This approximation > does not make sense if the data you have is already reflecting the > planar gradient, as is the case for neuromag306 data. Therefore, this > is not a bug, but a feature. > > Best, > Eelke > > On 17 December 2012 17:22, wrote: >> Hello ft-users, >> >> Today I tried to plot planar gradients for neuromag306 data. >> >> When calling avgFICplanar = ft_megplanar(cfg, avgFIC); >> >> I get the following error: >> ??? Error using ==> ft_checkdata at 406 >> This function requires ctf151, ctf275, bti148, bti248, itab153, >> yokogawa160 or yokogawa64 data as input, but you are giving >> neuromag306 data. >> >> Error in ==> ft_megplanar at 106 >> data = ft_checkdata(data, 'datatype', {'raw' 'freq'}, 'feedback', 'yes', >> 'hassampleinfo', 'yes', 'ismeg', 'yes', >> 'senstype', {'ctf151', 'ctf275', 'bti148', 'bti248', 'itab153', >> 'yokogawa160', 'yokoga >> Error in ==> reading_fif_planar at 39 >> avgFCplanar = ft_megplanar(cfg, avgFC); >> >> When manually adding ",'neuromag306'" to the senstype in line 106 and >> commenting out line 297 erverything seems to work as it should. >> >> Is this a bug or am I doing something wrong? >> >> >> _______________________________________________ >> fieldtrip mailing list >> fieldtrip at donders.ru.nl >> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From akolchin at indiana.edu Mon Dec 17 21:16:54 2012 From: akolchin at indiana.edu (Artemy Kolchinsky) Date: Mon, 17 Dec 2012 20:16:54 +0000 Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) In-Reply-To: <50cf40a8.a795320a.367a.ffff8b76SMTPIN_ADDED_BROKEN@mx.google.com> References: <50cf40a8.a795320a.367a.ffff8b76SMTPIN_ADDED_BROKEN@mx.google.com> Message-ID: Thank you greatly for the response! Keep in mind that in the “Bullmore-style” cluster-mass statistic**** > > ** ** > > \sum_i (t_i - c_i) where t_i > c_i**** > > ** ** > > the thresholds c_i are constants (i.e., independent of the data). > I think I see what you mean. In my case, t_i are t-statistics, and I use a p < 0.05 cutoff on both tails for the t distribution. So yes, here all the c_i would be the same for all voxels and independent of the data. > Importantly, these constants also enter in the permutation distribution > that is used to evaluated the significance of the maximum cluster-mass > statistic, to the effect that the Bullmore-style and the Fieldtrip-style > permutation distributions are shifted versions of each other. As a result, > the p-values that roll out of the two approaches are identical. > If I understand correctly, having the same resulting p-values could only be if the two methods assign the same rank-ordering to a given a set of clusters. But I don't think that is the case. Let's imagine that the t-statistic cutoff 'c' is equal to 1, and the data contains two suprathreshold clusters (let's say this is a spatial test and the clusters are composed of electrodes): - The first cluster has 10 electrodes, each one with a t-statistic equal to 1.1 - The second cluster has 2 electrodes, both with a t-statistic equal to 3 As I understand, Bullmore's method would assign cluster 1 a mass of 10*(1.1-1) = 1 and cluster 2 a mass of 2*(3-1)=4 , while your method would assign cluster 1 a mass of 10*1.1 = 11 and cluster 2 a mass of 2*3 = 6. Hence, given a null distribution, it should be possible to choose a cluster-based threshold that indicates as significant only cluster 1 under Bullmore's method, and only cluster 2 under yours. Thanks again, -Artemy -------------- next part -------------- An HTML attachment was scrubbed... URL: From clme at tf.uni-kiel.de Mon Dec 17 23:08:45 2012 From: clme at tf.uni-kiel.de (Claas Meckelnburg) Date: Mon, 17 Dec 2012 23:08:45 +0100 Subject: [FieldTrip] Bug in ft_megplanar? (Version 20121209) In-Reply-To: <50CF4ABD.3060701@cbs.mpg.de> References: <50CF4ABD.3060701@cbs.mpg.de> Message-ID: <50CF97ED.5060002@tf.uni-kiel.de> Hello Burkhard, just calling ft_combineplanar() worked like a charm. Sincerely, Claas Meckelnburg From bertram0611 at pku.edu.cn Tue Dec 18 02:55:43 2012 From: bertram0611 at pku.edu.cn (=?utf-8?B?6JSh5p6X?=) Date: Tue, 18 Dec 2012 09:55:43 +0800 (CST) Subject: [FieldTrip] =?gbk?q?about_when_to_filter?= Message-ID: <1139819751.565.1355795743084.JavaMail.root@bj-mail07.pku.edu.cn> Hi all, Now I have an urgent problem about how to filter EEG data recorded in NeuroScan. I don't know the distiction between raw data(ie. cnt file)and the data after ocular artifact reduction. I filtered these two different data, and then I got diffrent results. Especially, I found many vertical lines after filtering the data through ocular artifact reduction, but I didn't find vertical lines filtering raw data. Please give me some advice about how to filter offset. Thank you in advance. -- Lin Cai Department of Psychology, Peking University, Beijing 100871, P.R.China -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: SpamAssassinReport.txt URL: From r.vandermeij at donders.ru.nl Tue Dec 18 11:00:38 2012 From: r.vandermeij at donders.ru.nl (Roemer van der Meij) Date: Tue, 18 Dec 2012 11:00:38 +0100 Subject: [FieldTrip] about when to filter In-Reply-To: <1139819751.565.1355795743084.JavaMail.root@bj-mail07.pku.edu.cn> References: <1139819751.565.1355795743084.JavaMail.root@bj-mail07.pku.edu.cn> Message-ID: Hi Lin, Could you provide a little more information? How did you reduce your ocular artifacts? How did you filter your data? What kind of fieldtrip settings did you use, and what fieldtrip version were you using? Adding a picture of the artifacts you're referring to might also help. All the best, Roemer On Tue, Dec 18, 2012 at 2:55 AM, 蔡林 wrote: > Hi all, > > Now I have an urgent problem about how to filter EEG data recorded in > NeuroScan. I don't know the distiction between raw data(ie. cnt file)and > the data after ocular artifact reduction. I filtered these two different > data, and then I got diffrent results. Especially, I found many vertical > lines after filtering the data through ocular artifact reduction, but I > didn't find vertical lines filtering raw data. > Please give me some advice about how to filter offset. Thank you in > advance. > > -- > Lin Cai > Department of Psychology, Peking University, Beijing 100871, P.R.China > > Spam detection software, running on the system "kompost.science.ru.nl", > has > identified this incoming email as possible spam. The original message > has been attached to this so you can view it (if it isn't spam) or label > similar future email. If you have any questions, see > postmaster at science.ru.nl for details. > > Content preview: Hi all, Now I have an urgent problem about how to filter > EEG > data recorded in NeuroScan. I don't know the distiction between raw > data(ie. > cnt file)and the data after ocular artifact reduction. I filtered these > two > different data, and then I got diffrent results. Especially, I found > many > vertical lines after filtering the data through ocular artifact > reduction, > but I didn't find vertical lines filtering raw data. Please give me some > advice about how to filter offset. Thank you in advance. [...] > > Content analysis details: (6.2 points, 5.0 required) > > pts rule name description > ---- ---------------------- > -------------------------------------------------- > 3.2 CHARSET_FARAWAY_HEADER A foreign language charset used in headers > 3.0 BAYES_95 BODY: Bayesian spam probability is 95 to 99% > [score: 0.9800] > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- Roemer van der Meij M.Sc. PhD Candidate Donders Institute for Brain, Cognition and Behaviour Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3655932 E-mail: r.vandermeij at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From bertram0611 at pku.edu.cn Tue Dec 18 15:46:37 2012 From: bertram0611 at pku.edu.cn (=?utf-8?B?6JSh5p6X?=) Date: Tue, 18 Dec 2012 22:46:37 +0800 (CST) Subject: [FieldTrip] =?utf-8?b?5Zue5aSN77yaIFJlOiAgYWJvdXQgd2hlbiB0byBm?= =?utf-8?q?ilter?= In-Reply-To: Message-ID: <504028835.9897.1355841997211.JavaMail.root@bj-mail07.pku.edu.cn> Hi, Roemer Thank you! I preprocess the EEG raw data by NeuroScan instead of fieldtrip.I do not know when to filter the data. I want to filter the raw data at first, and then reduce the ocular artifacts, and then epoch...Am I right? Lin ----- 原始邮件 ----- 发件人: Roemer van der Meij 收件人: FieldTrip discussion list 已发送邮件: Tue, 18 Dec 2012 18:00:38 +0800 (CST) 主题: Re: [FieldTrip] about when to filter Hi Lin, Could you provide a little more information? How did you reduce your ocular artifacts? How did you filter your data? What kind of fieldtrip settings did you use, and what fieldtrip version were you using? Adding a picture of the artifacts you're referring to might also help. All the best, Roemer On Tue, Dec 18, 2012 at 2:55 AM, 蔡林 wrote: > Hi all, > > Now I have an urgent problem about how to filter EEG data recorded in > NeuroScan. I don't know the distiction between raw data(ie. cnt file)and > the data after ocular artifact reduction. I filtered these two different > data, and then I got diffrent results. Especially, I found many vertical > lines after filtering the data through ocular artifact reduction, but I > didn't find vertical lines filtering raw data. > Please give me some advice about how to filter offset. Thank you in > advance. > > -- > Lin Cai > Department of Psychology, Peking University, Beijing 100871, P.R.China > > Spam detection software, running on the system "kompost.science.ru.nl", > has > identified this incoming email as possible spam. The original message > has been attached to this so you can view it (if it isn't spam) or label > similar future email. If you have any questions, see > postmaster at science.ru.nl for details. > > Content preview: Hi all, Now I have an urgent problem about how to filter > EEG > data recorded in NeuroScan. I don't know the distiction between raw > data(ie. > cnt file)and the data after ocular artifact reduction. I filtered these > two > different data, and then I got diffrent results. Especially, I found > many > vertical lines after filtering the data through ocular artifact > reduction, > but I didn't find vertical lines filtering raw data. Please give me some > advice about how to filter offset. Thank you in advance. [...] > > Content analysis details: (6.2 points, 5.0 required) > > pts rule name description > ---- ---------------------- > -------------------------------------------------- > 3.2 CHARSET_FARAWAY_HEADER A foreign language charset used in headers > 3.0 BAYES_95 BODY: Bayesian spam probability is 95 to 99% > [score: 0.9800] > > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- Roemer van der Meij M.Sc. PhD Candidate Donders Institute for Brain, Cognition and Behaviour Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3655932 E-mail: r.vandermeij at donders.ru.nl -- Lin Cai Department of Psychology, Peking University, Beijing 100871, P.R.China From dave at davebritton.com Tue Dec 18 17:59:41 2012 From: dave at davebritton.com (Dave Britton) Date: Tue, 18 Dec 2012 11:59:41 -0500 Subject: [FieldTrip] processing ANT cnt files with FieldTrip Message-ID: <50D0A0FD.9020805@davebritton.com> I am trying to process EEG data in .cnt files obtained from the ANT system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have installed the Linux 64 bit version of FieldTrip. What is required to successfully process ANT eeg .cnt and .trg files? The libraries that came with the FieldTrip download did not work as it came from the download, and my investigations led me to what should be the latest version of the ANT import library. I have downloaded the latest ANT eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make install, successfully, but MATLAB cannot find the shared object file it should generate, even when I link the .so.3 file into the same directory as the MATLAB m-file and the MATLAB current directory. Initially the system was unable to read the cnt files, which is why I attempted to reinstall the anteepimort library. Is there more to this installation than the standard" configure, make, make install" process? Any clues would be appreciated! Dave Britton City College of New York Psychology Dept. Here is the error message I get: >> cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; >> cfg.hdr = ft_read_header(cfg.headerfile); Invalid MEX-file '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': libeep.so.3: cannot open shared object file: No such file or directory Error in ft_read_header (line 627) hdr = read_eep_cnt(filename, 1, 1); Previous efforts before recompiling libeep produced failures that seemed to indicate the read_eep_cnt routine was incorrectly assessing the cnt file size and obtaining garbage, which is why I attempted to rebuild the eepimport library: ===== error messages ==== >> clear; % required items for cfg: flank1 = 'FLANK1'; % markersA in tutorial walkthru flank2 = 'FLANK2'; target = 'TARGET'; cfg=[]; cfg.trialdef.prestim = 0.05; % 50 ms baseline cfg.trialdef.poststim = 0.5; % trials last 500 ms cfg.trialdef.eventtype = 'input'; cfg.trialdef.eventalue = target; cfg.dataset = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; cfg.triggerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.trg'; cfg.trialfun = 'trialfun_semattn'; cfg.padding = 2; cfg.dataformat = 'ns_cnt32'; % see faq about reading Neuroscan cnt files cfg.headerformat = 'ns_cnt32'; cfg.eventformat = 'ns_cnt32'; cfg.blockread = 1; % cfg = ft_definetrial(cfg); evaluating trialfunction 'trialfun_semattn' Warning: could not determine channel type from the CTF header > In ft_chantype at 213 In ft_read_header at 1711 In trialfun_semattn at 5 In ft_definetrial at 162 Warning: could not determine channel type from the CTF header > In ft_chantype at 213 In ft_read_header at 1711 In ft_read_event at 600 In trialfun_semattn at 6 In ft_definetrial at 162 found 407 events created 132 trials the call to "ft_definetrial" took 4 seconds and required the additional allocation of an estimated 13 MB Loading file /home/dave/semanticflanker/data/ANT/20121114_1013.cnt ... Reading data ..... Scaling data ..... Reading Event Table... Skipping event table (tag != 1,2,3 ; theoritically impossible) Warning: channel labels should not be empty, creating unique labels > In ft_read_header at 1684 In ft_preprocessing at 343 In semattn_main at 34 processing channel { '�RXqwh ��' ............(channel data is garbage).... ===== end of error messages=== From arno at cerco.ups-tlse.fr Tue Dec 18 18:38:50 2012 From: arno at cerco.ups-tlse.fr (Arnaud Delorme) Date: Tue, 18 Dec 2012 09:38:50 -0800 Subject: [FieldTrip] processing ANT cnt files with FieldTrip In-Reply-To: <50D0A0FD.9020805@davebritton.com> References: <50D0A0FD.9020805@davebritton.com> Message-ID: <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> Dear Dave, last time I have checked FileIO (the Fieldtrip data importer) will not read ANT trigger files by default (it should read the raw data though - Robert Oostenveld one of the main developer of Fieldtrip was interning at ANT about 10 years ago and wrote the Matlab import function for them). The anteepimport1.09 is a plugin for the EEGLAB software. It can import events (and then you can save the data in EEGLAB format and reload it in Fieldtrip). You cannot recompile the ANT project (their source files are proprietary and not included). Hope this helps. Best wishes, Arno On 18 Dec 2012, at 08:59, Dave Britton wrote: > I am trying to process EEG data in .cnt files obtained from the ANT system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have installed the Linux 64 bit version of FieldTrip. > > What is required to successfully process ANT eeg .cnt and .trg files? The libraries that came with the FieldTrip download did not work as it came from the download, and my investigations led me to what should be the latest version of the ANT import library. I have downloaded the latest ANT eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make install, successfully, but MATLAB cannot find the shared object file it should generate, even when I link the .so.3 file into the same directory as the MATLAB m-file and the MATLAB current directory. > > Initially the system was unable to read the cnt files, which is why I attempted to reinstall the anteepimort library. Is there more to this installation than the standard" configure, make, make install" process? Any clues would be appreciated! > > Dave Britton > City College of New York > Psychology Dept. > > Here is the error message I get: > > >> cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; > >> cfg.hdr = ft_read_header(cfg.headerfile); > Invalid MEX-file > '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': > libeep.so.3: cannot open shared object file: No such file or directory > > Error in ft_read_header (line 627) > hdr = read_eep_cnt(filename, 1, 1); > > Previous efforts before recompiling libeep produced failures that seemed to indicate the read_eep_cnt routine was incorrectly assessing the cnt file size and obtaining garbage, which is why I attempted to rebuild the eepimport library: > ===== error messages ==== > >> clear; > % required items for cfg: > flank1 = 'FLANK1'; % markersA in tutorial walkthru > flank2 = 'FLANK2'; > target = 'TARGET'; > cfg=[]; > cfg.trialdef.prestim = 0.05; % 50 ms baseline > cfg.trialdef.poststim = 0.5; % trials last 500 ms > cfg.trialdef.eventtype = 'input'; > cfg.trialdef.eventalue = target; > cfg.dataset = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; > cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; > cfg.triggerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.trg'; > cfg.trialfun = 'trialfun_semattn'; > cfg.padding = 2; > cfg.dataformat = 'ns_cnt32'; % see faq about reading Neuroscan cnt files > cfg.headerformat = 'ns_cnt32'; > cfg.eventformat = 'ns_cnt32'; > cfg.blockread = 1; > % > cfg = ft_definetrial(cfg); > evaluating trialfunction 'trialfun_semattn' > Warning: could not determine channel type from the CTF header > > In ft_chantype at 213 > In ft_read_header at 1711 > In trialfun_semattn at 5 > In ft_definetrial at 162 > Warning: could not determine channel type from the CTF header > > In ft_chantype at 213 > In ft_read_header at 1711 > In ft_read_event at 600 > In trialfun_semattn at 6 > In ft_definetrial at 162 > found 407 events > created 132 trials > the call to "ft_definetrial" took 4 seconds and required the additional allocation of an estimated 13 MB > Loading file /home/dave/semanticflanker/data/ANT/20121114_1013.cnt ... > Reading data ..... > Scaling data ..... > Reading Event Table... > Skipping event table (tag != 1,2,3 ; theoritically impossible) > Warning: channel labels should not be empty, creating unique labels > > In ft_read_header at 1684 > In ft_preprocessing at 343 > In semattn_main at 34 > processing channel { '�RXqwh ��' ............(channel data is garbage).... > > ===== end of error messages=== > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From dave at davebritton.com Tue Dec 18 20:18:05 2012 From: dave at davebritton.com (Dave Britton) Date: Tue, 18 Dec 2012 14:18:05 -0500 Subject: [FieldTrip] processing ANT cnt files with FieldTrip In-Reply-To: <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> References: <50D0A0FD.9020805@davebritton.com> <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> Message-ID: <50D0C16D.5060105@davebritton.com> Thanks, Arno, I've installed the ant-eep import files into the EEGLAB plugin directory, but EEGLAB gives me the same error message as I got with FieldTrip: EEGLAB error in function loadeep() at line 75 Invalid MEX-file '/home/dave/eeglab11_0_4_3b/plugins/anteeimport1.08/read_eep_cnt.mexa64': libeep.so.3:cannot open shared object file:No such file or directory The libeep.so.3 file is copied into the same directory as the read_eep_cnt.mexa64 file. Do you have any suggestions? -Dave On 12/18/2012 12:38 PM, Arnaud Delorme wrote: > Dear Dave, > > last time I have checked FileIO (the Fieldtrip data importer) will not read ANT trigger files by default (it should read the raw data though - Robert Oostenveld one of the main developer of Fieldtrip was interning at ANT about 10 years ago and wrote the Matlab import function for them). > The anteepimport1.09 is a plugin for the EEGLAB software. It can import events (and then you can save the data in EEGLAB format and reload it in Fieldtrip). You cannot recompile the ANT project (their source files are proprietary and not included). > > Hope this helps. > Best wishes, > > Arno > > On 18 Dec 2012, at 08:59, Dave Britton wrote: > >> I am trying to process EEG data in .cnt files obtained from the ANT system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have installed the Linux 64 bit version of FieldTrip. >> >> What is required to successfully process ANT eeg .cnt and .trg files? The libraries that came with the FieldTrip download did not work as it came from the download, and my investigations led me to what should be the latest version of the ANT import library. I have downloaded the latest ANT eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make install, successfully, but MATLAB cannot find the shared object file it should generate, even when I link the .so.3 file into the same directory as the MATLAB m-file and the MATLAB current directory. >> >> Initially the system was unable to read the cnt files, which is why I attempted to reinstall the anteepimort library. Is there more to this installation than the standard" configure, make, make install" process? Any clues would be appreciated! >> >> Dave Britton >> City College of New York >> Psychology Dept. >> >> Here is the error message I get: >> >>>> cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; >>>> cfg.hdr = ft_read_header(cfg.headerfile); >> Invalid MEX-file >> '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': >> libeep.so.3: cannot open shared object file: No such file or directory >> >> Error in ft_read_header (line 627) >> hdr = read_eep_cnt(filename, 1, 1); >> >> From berryv.dberg at gmail.com Tue Dec 18 20:56:12 2012 From: berryv.dberg at gmail.com (berry van den berg) Date: Tue, 18 Dec 2012 14:56:12 -0500 Subject: [FieldTrip] possible bugs in using mask in ft_topoplotTFR? Message-ID: Hi Fieldtrip experts, Im working on an EEG analysis in fieldtrip and I came across some possible bugs or perhaps I am doing something wrong.... I have a dataset with 28 subjects and my grandavg (including the 28 individuals) look like this: grandavg = individual: [28x64x1251 double] label: {64x1 cell} time: [1x1251 double] dimord: 'subj_chan_time' cfg: [1x1 struct] mask: [1x64 double] I first tried to use a mask in ft_topoplotER in the 11/11/2012 fieldtrip version but this did not work: input: cfg = []; cfg.xlim = [0.3 0.4]; cfg.baseline = [-.2 0]; cfg.layout = ft_prepare_layout(timelock_mat{1,1}) cfg.interactive='yes'; grandavg.mask = [ones(1, 60) zeros(1,4)]; %just an example cfg.maskparameter = 'mask'; ft_topoplotER(cfg,grandavg) Reference to non-existent field 'mask'. Error in ft_topoplotTFR (line 713) datmask = data.(cfg.maskparameter); it seems that around line 420 in the function data = ft_timelockanalysis(tmpcfg, data) it does not retrieve the mask parameter. To check if it was an old bug I downloaded the newest FT version: 16/12/2012: the mask parameter does not work in this version yet.. Moreover: topoplotTFR line 409-411 there is a section added compared to version 11/11/2012: %%%%%%% if isfield(cfg, 'parameter') && ~isfield(data, cfg.parameter) error('cfg.parameter=%s is not present in data structure', cfg.parameter); end %%%%%%% which checks whether the is present... but because I have all my individual subjects in the data I am waiting for line 425:data = ft_timelockanalysis(tmpcfg, data) to average my dataset. It crashes here (after removal of these lines it runs like normal) Error using ft_topoplotTFR (line 412) cfg.parameter=avg is not present in data structure Is there something I am missing? Thanks a lot! (also for providing this very good package!) Berry -------------- next part -------------- An HTML attachment was scrubbed... URL: From clara.scholl at gmail.com Tue Dec 18 21:04:27 2012 From: clara.scholl at gmail.com (Clara A. Scholl) Date: Tue, 18 Dec 2012 15:04:27 -0500 Subject: [FieldTrip] possible bugs in using mask in ft_topoplotTFR? In-Reply-To: References: Message-ID: Dear Berry, I found that the mask works when plotting topos directly from an average field (grandavg.avg), but not when plotting the individual data ( grandavg.individual -- which are averaged from the individual field, so effectively the same thing). Can you try averaging over the first dimension of grandavg.individual (grandavg.avg=squeeze(mean(grandavg.individual))) and removing the individual field? Does that work? Good luck, Clara On Tue, Dec 18, 2012 at 2:56 PM, berry van den berg wrote: > Hi Fieldtrip experts, > > Im working on an EEG analysis in fieldtrip and I came across some possible > bugs or perhaps I am doing something wrong.... I have a dataset with 28 > subjects and my grandavg (including the 28 individuals) look like this: > > grandavg = > > individual: [28x64x1251 double] > label: {64x1 cell} > time: [1x1251 double] > dimord: 'subj_chan_time' > cfg: [1x1 struct] > mask: [1x64 double] > > I first tried to use a mask in ft_topoplotER in the 11/11/2012 fieldtrip > version but this did not work: > > input: > > cfg = []; > cfg.xlim = [0.3 0.4]; > cfg.baseline = [-.2 0]; > cfg.layout = ft_prepare_layout(timelock_mat{1,1}) > cfg.interactive='yes'; > grandavg.mask = [ones(1, 60) zeros(1,4)]; %just an example > cfg.maskparameter = 'mask'; > ft_topoplotER(cfg,grandavg) > > Reference to non-existent field 'mask'. > Error in ft_topoplotTFR (line 713) > datmask = data.(cfg.maskparameter); > > it seems that around line 420 in the function data = > ft_timelockanalysis(tmpcfg, data) it does not retrieve the mask parameter. > > To check if it was an old bug I downloaded the newest FT version: > 16/12/2012: the mask parameter does not work in this version yet.. > > Moreover: > topoplotTFR line 409-411 there is a section added compared to version > 11/11/2012: > > %%%%%%% > if isfield(cfg, 'parameter') && ~isfield(data, cfg.parameter) > error('cfg.parameter=%s is not present in data structure', cfg.parameter); > end > %%%%%%% > > which checks whether the is present... but because I have all my individual > subjects in the data I am waiting for line 425:data = > ft_timelockanalysis(tmpcfg, data) to average my dataset. It crashes here > (after removal of these lines it runs like normal) > > Error using ft_topoplotTFR (line 412) > cfg.parameter=avg is not present in data structure > > Is there something I am missing? > > Thanks a lot! (also for providing this very good package!) > > Berry > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From berryv.dberg at gmail.com Tue Dec 18 21:38:00 2012 From: berryv.dberg at gmail.com (berry van den berg) Date: Tue, 18 Dec 2012 15:38:00 -0500 Subject: [FieldTrip] possible bugs in using mask in ft_topoplotTFR? In-Reply-To: References: Message-ID: Dear Clara, Thanks for the quick reply! That indeed does the job! Im new to fieldtrip so I dont know all the details but it seems to be the same as adding the following around line 428 in ft_topoplotTFR (fieldtrip 20121216)... tmpmask = data.(cfg.maskparameter); data = ft_timelockanalysis(tmpcfg, data); data.(cfg.maskparameter) = tmpmask; plus removing lines 411-413: % if isfield(cfg, 'parameter') && ~isfield(data, cfg.parameter) % error('cfg.parameter=%s is not present in data structure', cfg.parameter); % end Cheers! Berry 2012/12/18 Clara A. Scholl > Dear Berry, > > I found that the mask works when plotting topos directly from an > average field (grandavg.avg), but not when plotting the individual > data ( grandavg.individual -- which are averaged from the individual > field, so effectively the same thing). Can you try averaging over the > first dimension of grandavg.individual > (grandavg.avg=squeeze(mean(grandavg.individual))) and removing the > individual field? Does that work? > > Good luck, > Clara > > On Tue, Dec 18, 2012 at 2:56 PM, berry van den berg > wrote: > > Hi Fieldtrip experts, > > > > Im working on an EEG analysis in fieldtrip and I came across some > possible > > bugs or perhaps I am doing something wrong.... I have a dataset with 28 > > subjects and my grandavg (including the 28 individuals) look like this: > > > > grandavg = > > > > individual: [28x64x1251 double] > > label: {64x1 cell} > > time: [1x1251 double] > > dimord: 'subj_chan_time' > > cfg: [1x1 struct] > > mask: [1x64 double] > > > > I first tried to use a mask in ft_topoplotER in the 11/11/2012 > fieldtrip > > version but this did not work: > > > > input: > > > > cfg = []; > > cfg.xlim = [0.3 0.4]; > > cfg.baseline = [-.2 0]; > > cfg.layout = ft_prepare_layout(timelock_mat{1,1}) > > cfg.interactive='yes'; > > grandavg.mask = [ones(1, 60) zeros(1,4)]; %just an example > > cfg.maskparameter = 'mask'; > > ft_topoplotER(cfg,grandavg) > > > > Reference to non-existent field 'mask'. > > Error in ft_topoplotTFR (line 713) > > datmask = data.(cfg.maskparameter); > > > > it seems that around line 420 in the function data = > > ft_timelockanalysis(tmpcfg, data) it does not retrieve the mask > parameter. > > > > To check if it was an old bug I downloaded the newest FT version: > > 16/12/2012: the mask parameter does not work in this version yet.. > > > > Moreover: > > topoplotTFR line 409-411 there is a section added compared to version > > 11/11/2012: > > > > %%%%%%% > > if isfield(cfg, 'parameter') && ~isfield(data, cfg.parameter) > > error('cfg.parameter=%s is not present in data structure', > cfg.parameter); > > end > > %%%%%%% > > > > which checks whether the is present... but because I have all my > individual > > subjects in the data I am waiting for line 425:data = > > ft_timelockanalysis(tmpcfg, data) to average my dataset. It crashes here > > (after removal of these lines it runs like normal) > > > > Error using ft_topoplotTFR (line 412) > > cfg.parameter=avg is not present in data structure > > > > Is there something I am missing? > > > > Thanks a lot! (also for providing this very good package!) > > > > Berry > > > > _______________________________________________ > > fieldtrip mailing list > > fieldtrip at donders.ru.nl > > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- Berry van den Berg berryv.dberg at gmail.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From arno at cerco.ups-tlse.fr Tue Dec 18 21:44:51 2012 From: arno at cerco.ups-tlse.fr (Arnaud Delorme) Date: Tue, 18 Dec 2012 12:44:51 -0800 Subject: [FieldTrip] processing ANT cnt files with FieldTrip In-Reply-To: <50D0C16D.5060105@davebritton.com> References: <50D0A0FD.9020805@davebritton.com> <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> <50D0C16D.5060105@davebritton.com> Message-ID: <9DF8F864-24FD-4A96-AD24-32AB4D86211C@cerco.ups-tlse.fr> This is because ANT did not compile their program for your platform. There is nothing you can do about (except maybe contact ANT so they compile it for your machine). Import your data on a different machine. Then you can process it on your machine. Best, Arno -- Arnaud Delorme, PhD Centre de Recherche Cerveau et Cognition - UMR 5549 Pavillon Baudot, Hopital Purpan, BP 25202 31052 Toulouse Cedex, France On 18 Dec 2012, at 11:18, Dave Britton wrote: > Thanks, Arno, > I've installed the ant-eep import files into the EEGLAB plugin directory, but EEGLAB gives me the same error message as I got with FieldTrip: > > EEGLAB error in function loadeep() at line 75 > Invalid MEX-file '/home/dave/eeglab11_0_4_3b/plugins/anteeimport1.08/read_eep_cnt.mexa64': libeep.so.3:cannot open shared object file:No such file or directory > > The libeep.so.3 file is copied into the same directory as the read_eep_cnt.mexa64 file. > > Do you have any suggestions? > -Dave > > > On 12/18/2012 12:38 PM, Arnaud Delorme wrote: >> Dear Dave, >> >> last time I have checked FileIO (the Fieldtrip data importer) will not read ANT trigger files by default (it should read the raw data though - Robert Oostenveld one of the main developer of Fieldtrip was interning at ANT about 10 years ago and wrote the Matlab import function for them). >> The anteepimport1.09 is a plugin for the EEGLAB software. It can import events (and then you can save the data in EEGLAB format and reload it in Fieldtrip). You cannot recompile the ANT project (their source files are proprietary and not included). >> >> Hope this helps. >> Best wishes, >> >> Arno >> >> On 18 Dec 2012, at 08:59, Dave Britton wrote: >> >>> I am trying to process EEG data in .cnt files obtained from the ANT system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have installed the Linux 64 bit version of FieldTrip. >>> >>> What is required to successfully process ANT eeg .cnt and .trg files? The libraries that came with the FieldTrip download did not work as it came from the download, and my investigations led me to what should be the latest version of the ANT import library. I have downloaded the latest ANT eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make install, successfully, but MATLAB cannot find the shared object file it should generate, even when I link the .so.3 file into the same directory as the MATLAB m-file and the MATLAB current directory. >>> >>> Initially the system was unable to read the cnt files, which is why I attempted to reinstall the anteepimort library. Is there more to this installation than the standard" configure, make, make install" process? Any clues would be appreciated! >>> >>> Dave Britton >>> City College of New York >>> Psychology Dept. >>> >>> Here is the error message I get: >>> >>>>> cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; >>>>> cfg.hdr = ft_read_header(cfg.headerfile); >>> Invalid MEX-file >>> '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': >>> libeep.so.3: cannot open shared object file: No such file or directory >>> >>> Error in ft_read_header (line 627) >>> hdr = read_eep_cnt(filename, 1, 1); >>> >>> > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From r.vandermeij at donders.ru.nl Wed Dec 19 11:47:14 2012 From: r.vandermeij at donders.ru.nl (Roemer van der Meij) Date: Wed, 19 Dec 2012 11:47:14 +0100 Subject: [FieldTrip] about when to filter In-Reply-To: References: <1139819751.565.1355795743084.JavaMail.root@bj-mail07.pku.edu.cn> Message-ID: Hi Lin, That depends on what kind of filters you use, and how you remove your ocular artifacts. Obviously, if you are removing ocular artifacts by hand any filtering that will help you in detecting eye movements will be of course be of benefit. If you use a blind decomposition technique such as ICA to try and remove ocular artifacts, it's a bit more tricky. Some specific filters might help depending on what kind of other artifacts are present, but it's important for example not to smear out your eye movements. Before ones does ICA to remove artifacts it's a good idea to remove trials that are unfixable, like trials with jumps in them (i.e. hitting the ceiling of the amplifier). I can only help you further if you more precisely describe what you mean with filtering and ocular artifact removal. There are many many ways to go about these two. All the best, Roemer On Tue, Dec 18, 2012 at 11:00 AM, Roemer van der Meij < r.vandermeij at donders.ru.nl> wrote: > Hi Lin, > > Could you provide a little more information? How did you reduce your > ocular artifacts? How did you filter your data? What kind of fieldtrip > settings did you use, and what fieldtrip version were you using? Adding a > picture of the artifacts you're referring to might also help. > > All the best, > Roemer > > > On Tue, Dec 18, 2012 at 2:55 AM, 蔡林 wrote: > >> Hi all, >> >> Now I have an urgent problem about how to filter EEG data recorded in >> NeuroScan. I don't know the distiction between raw data(ie. cnt file)and >> the data after ocular artifact reduction. I filtered these two different >> data, and then I got diffrent results. Especially, I found many vertical >> lines after filtering the data through ocular artifact reduction, but I >> didn't find vertical lines filtering raw data. >> Please give me some advice about how to filter offset. Thank you in >> advance. >> >> -- >> Lin Cai >> Department of Psychology, Peking University, Beijing 100871, P.R.China >> >> Spam detection software, running on the system "kompost.science.ru.nl", >> has >> identified this incoming email as possible spam. The original message >> has been attached to this so you can view it (if it isn't spam) or label >> similar future email. If you have any questions, see >> postmaster at science.ru.nl for details. >> >> Content preview: Hi all, Now I have an urgent problem about how to >> filter EEG >> >> data recorded in NeuroScan. I don't know the distiction between raw >> data(ie. >> cnt file)and the data after ocular artifact reduction. I filtered >> these two >> different data, and then I got diffrent results. Especially, I found >> many >> vertical lines after filtering the data through ocular artifact >> reduction, >> but I didn't find vertical lines filtering raw data. Please give me >> some >> advice about how to filter offset. Thank you in advance. [...] >> >> Content analysis details: (6.2 points, 5.0 required) >> >> pts rule name description >> ---- ---------------------- >> -------------------------------------------------- >> 3.2 CHARSET_FARAWAY_HEADER A foreign language charset used in headers >> 3.0 BAYES_95 BODY: Bayesian spam probability is 95 to 99% >> [score: 0.9800] >> >> >> >> _______________________________________________ >> fieldtrip mailing list >> fieldtrip at donders.ru.nl >> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip >> > > > > -- > Roemer van der Meij M.Sc. > PhD Candidate > Donders Institute for Brain, Cognition and Behaviour > Centre for Cognition > P.O. Box 9104 > 6500 HE Nijmegen > The Netherlands > Tel: +31(0)24 3655932 > E-mail: r.vandermeij at donders.ru.nl > > -- Roemer van der Meij M.Sc. PhD Candidate Donders Institute for Brain, Cognition and Behaviour Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3655932 E-mail: r.vandermeij at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From julian.keil at gmail.com Wed Dec 19 12:30:21 2012 From: julian.keil at gmail.com (Julian Keil) Date: Wed, 19 Dec 2012 12:30:21 +0100 Subject: [FieldTrip] mtmconvol time window Message-ID: Dear all, I have a rather specific question on the representation of the time window in the mtmconvol freqanalysis. In the help-file, it says that cfg.toi is the time vector, on which the analysis window should be centered, however as far as I can see in the code for the ft_specest_mtmconvol, the time of interest (timeoi), the time-bin of interest (timeboi) as well as the requested time bin (reqtimeboi) (ft_specest_mtmconvol, fieldtrip-20120815, line 280 ff) seem to suggest that the time I set in cfg.toi are actually the starting time for the spectrum estimation. In other words, if I'm looking to the 10 Hz signal and I set the cfg.toi=0.15 and cfg.t_ftimwin=0.3 in order to get 3 cycles of the 10 Hz oscillation, will the spectrum be estimated for the interval of 0ms to 300ms or for the interval of 150ms to 450ms? Thanks a lot! Julian ******************** Dr. Julian Keil AG Multisensorische Integration Psychiatrische Universitätsklinik der Charité im St. Hedwig-Krankenhaus Große Hamburger Straße 5-11, Raum E 307 10115 Berlin Telefon: +49-30-2311-1879 Fax: +49-30-2311-2209 http://psy-ccm.charite.de/forschung/bildgebung/ag_multisensorische_integration -------------- next part -------------- An HTML attachment was scrubbed... URL: From r.vandermeij at donders.ru.nl Wed Dec 19 13:21:00 2012 From: r.vandermeij at donders.ru.nl (Roemer van der Meij) Date: Wed, 19 Dec 2012 13:21:00 +0100 Subject: [FieldTrip] mtmconvol time window In-Reply-To: References: Message-ID: Hi Julian, Your first description is the what it reflects. If you have a cfg.toi = 0.15, the window will be centered at 0.15, going from 0 to 0.3. If it helps, you can think of cfg.toi purely as a selection. The fourier decomposition//wavelet convolution will be calculated for all samples in each trial (with the window centered on each sample), the only thing cfg.toi does is select which of those to keep as output. Cfg.toi will never affect computation, only post-computation selection. To be more precise, in mtmconvol you are doing a wavelet convolution in the time-domain by multiplication in the frequency-domain. Timeoi and timeboi are used to select those bits of the resulting complex time-series (and they reflect cfg.toi). All the best, Roemer On Wed, Dec 19, 2012 at 12:30 PM, Julian Keil wrote: > Dear all, > > I have a rather specific question on the representation of the time window > in the mtmconvol freqanalysis. > In the help-file, it says that cfg.toi is the time vector, on which the > analysis window should be centered, however as far as I can see in the code > for the ft_specest_mtmconvol, the time of interest (timeoi), the time-bin > of interest (timeboi) as well as the requested time bin (reqtimeboi) > (ft_specest_mtmconvol, fieldtrip-20120815, line 280 ff) seem to suggest > that the time I set in cfg.toi are actually the starting time for the > spectrum estimation. > > In other words, if I'm looking to the 10 Hz signal and I set the > cfg.toi=0.15 and cfg.t_ftimwin=0.3 in order to get 3 cycles of the 10 Hz > oscillation, will the spectrum be estimated for the interval of 0ms to > 300ms or for the interval of 150ms to 450ms? > > Thanks a lot! > > Julian > > > ******************** > *Dr. Julian Keil* > * > * > AG Multisensorische Integration > Psychiatrische Universitätsklinik > der Charité im St. Hedwig-Krankenhaus > Große Hamburger Straße 5-11, Raum E 307 > 10115 Berlin > > Telefon: +49-30-2311-1879 > Fax: +49-30-2311-2209 > > http://psy-ccm.charite.de/forschung/bildgebung/ag_multisensorische_integration > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- Roemer van der Meij M.Sc. PhD Candidate Donders Institute for Brain, Cognition and Behaviour Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3655932 E-mail: r.vandermeij at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From polomacnenad at gmail.com Wed Dec 19 15:22:00 2012 From: polomacnenad at gmail.com (Nenad Polomac) Date: Wed, 19 Dec 2012 15:22:00 +0100 Subject: [FieldTrip] Fourier transformation for the quick overview Message-ID: Hi everybody, I would like to know is there a method with which I could have a quick overview on power-spectrum of data before preprocessing. So, just Fourier transformation of trials after ft_definetrial. I need that to see the line noise artifact in my data. Thanks in advance! Nenad -------------- next part -------------- An HTML attachment was scrubbed... URL: From r.vandermeij at donders.ru.nl Wed Dec 19 16:25:49 2012 From: r.vandermeij at donders.ru.nl (Roemer van der Meij) Date: Wed, 19 Dec 2012 16:25:49 +0100 Subject: [FieldTrip] Fourier transformation for the quick overview In-Reply-To: References: Message-ID: Hi Nenad, I'm glad you ask, because I implemented functionality just for that purpose! ;) Either before or after calling ft_definetrial, plot the data using ft_databrowser. Once inside the databrowser, select a piece of data by clicking+dragging a selection box around it. Then right-click the selected data, click on 'simplefft', and presto! A 'raw' fft of your data, perfect for detecting line-spectra (i.e. tapered with a 'boxcar' taper). Cheers, Roemer On Wed, Dec 19, 2012 at 3:22 PM, Nenad Polomac wrote: > Hi everybody, > I would like to know is there a method with which I could have a quick > overview on power-spectrum of data before preprocessing. So, just Fourier > transformation of trials after ft_definetrial. I need that to see the line > noise artifact in my data. > > Thanks in advance! > > Nenad > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -- Roemer van der Meij M.Sc. PhD Candidate Donders Institute for Brain, Cognition and Behaviour Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3655932 E-mail: r.vandermeij at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From taherahlansari at hotmail.com Wed Dec 19 16:50:27 2012 From: taherahlansari at hotmail.com (Tahereh L. Ansari) Date: Wed, 19 Dec 2012 15:50:27 +0000 Subject: [FieldTrip] RealTime fMRI - Serial-event Tool Message-ID: Hi,I'm trying to setup ft buffer to use with realtime fMRI but i'm not clear on how to use the 'serial-event tool'. My questions are (sorry if they seem basic, i'm a newcomer to fMRI): 1 - Is one of the purposes of the serial-event tool to stop the real-time analysis loop? if not, how is the buffer 'emptied'? (my matlab preprocessing script continues to read 'old' data after gui_streamer is paused - even after the matlab call: buffer('flush_hdr', 'buffer://localhost:1972) at the end of the scan 2 - Does the Siemens scanner output the serial event or does it ONLY output the TTL signal? if the latter, how is the TTL signal (from the scanner) fed into the synchronisation module? Our Siemens 3T generates a TTL pulse and this is transmitted to the presentation computer through the parallel port. Can we not use this setup with the FT buffer? 3 - Can we achieve what the serial-event tool is doing using parallel port (with or without the serial-event tool)? 4 - If a serial port setup is necessary, can you suggest a (hardware) setup (preferably with a schematic)? Many many thanks for your help!! L -------------- next part -------------- An HTML attachment was scrubbed... URL: From polomacnenad at gmail.com Wed Dec 19 16:54:25 2012 From: polomacnenad at gmail.com (Nenad Polomac) Date: Wed, 19 Dec 2012 16:54:25 +0100 Subject: [FieldTrip] Fourier transformation for the quick overview Message-ID: Dear Roemer, Thank you very much! This is a very useful tool! Best! Nenad -------------- next part -------------- An HTML attachment was scrubbed... URL: From farrugia at cbs.mpg.de Wed Dec 19 17:00:45 2012 From: farrugia at cbs.mpg.de (Nicolas Farrugia) Date: Wed, 19 Dec 2012 17:00:45 +0100 (CET) Subject: [FieldTrip] Testing ITC differences using diff_itc in a first and second level analysis In-Reply-To: <2101549396.5901.1355931977821.JavaMail.root@zimbra> Message-ID: <402219618.6060.1355932845195.JavaMail.root@zimbra> Dear all, I m trying to analyse differences in Inter-Trial Coherence between conditions in a group of subjects. I figured out with this paper "Nonparametric statistical testing of coherence differences" by Maris, Schoffelen & Fries 2007, that this can be done reliably using non-parametric tests with Montecarlo permutations and it s implemented in Fieldtrip under the statfun diff_itc in the montecarlo method. I m now using the cluster based tests in Fieldtrip on my complex TF data using the following parameters (I extracted only the relevant ones for the discussion) : cfg.method = 'montecarlo'; cfg.correctm = 'cluster'; cfg.clusteralpha = 0.05; cfg.clusterstatistic = 'maxsum'; cfg.minnbchan = 2; cfg.tail = 0; cfg.clustertail = 0; cfg.alpha = 0.025; cfg.numrandomization = 500; cfg.statistic = 'diff_itc'; cfg.complex = 'absdiff'; cfg.parameter = 'fourierspctrm'; cfg.avgoverfreq = 'yes'; This also need a clustercritval : so my first question will be : -> How to choose the clustercritval in the case of this test of ITC differences ? I ve been trying different things but I m not sure. Then, this constitutes my first level analysis, so for each subject I have cluster statistics, their time/spatial course and their significance. >From here how can I go towards for a second level analysis ( similarly as in SPM) ? Is there already building blocks for that in Fieldtrip ? Should one try to find a cluster center / time window and use the cluster statistics, and use the resulting values from all subjects in a non-parametric test ? Thanks for your help, Nicolas ---------------------------------------------------------------------------- ----- Dr. Nicolas Farrugia Post-doc Research Group "Subcortical contributions to comprehension" Department of Neuropsychology Max Planck Institute for Human Cognitive and Brain Sciences Stephanstraße 1A, 04103 Leipzig, Germany Phone : +49 (0) 176 848 90 393 http://nicolasfarrugia.fr From andreas.wutz-1 at unitn.it Wed Dec 19 17:44:18 2012 From: andreas.wutz-1 at unitn.it (Wutz, Andreas) Date: Wed, 19 Dec 2012 17:44:18 +0100 Subject: [FieldTrip] Testing ITC differences using diff_itc in a first and second level analysis In-Reply-To: <402219618.6060.1355932845195.JavaMail.root@zimbra> References: <2101549396.5901.1355931977821.JavaMail.root@zimbra>, <402219618.6060.1355932845195.JavaMail.root@zimbra> Message-ID: Dear Nicolas, I cannot help you on your specific question, but I can show you how I usually compare two conditions within subjects using the ITC. You would need the CircStatToolbox (is open source), though. After you computed your complex TF, you can get the corresponding angles in radians using: data.radspctrm=angle(data.fourierspctrm); you get the ITC for each time-frequency-sensor point by calculating the resultant vector length over the trials: data.ITC=squeeze(circ_r(data.radspctrm, [],[],1)); Now, I simply compute the differences in ITC between the two conditions using cluster-based nonparametric statistics as implemented in FT with cfg.parameter = 'ITC'; You may have to add one line of code, allowing ft_freqstatistics to operate with the field .ITC. I see no reason why a dependent samples t-test should be different for ordinary powerspectra or ITC-spectra. I hope this helps. Best, Andreas Andreas Wutz PhD Student CIMeC - Center for Mind/Brain Sciences Università degli studi di Trento ________________________________________ Von: fieldtrip-bounces at science.ru.nl [fieldtrip-bounces at science.ru.nl] im Auftrag von Nicolas Farrugia [farrugia at cbs.mpg.de] Gesendet: Mittwoch, 19. Dezember 2012 17:00 An: fieldtrip at science.ru.nl Betreff: [FieldTrip] Testing ITC differences using diff_itc in a first and second level analysis Dear all, I m trying to analyse differences in Inter-Trial Coherence between conditions in a group of subjects. I figured out with this paper "Nonparametric statistical testing of coherence differences" by Maris, Schoffelen & Fries 2007, that this can be done reliably using non-parametric tests with Montecarlo permutations and it s implemented in Fieldtrip under the statfun diff_itc in the montecarlo method. I m now using the cluster based tests in Fieldtrip on my complex TF data using the following parameters (I extracted only the relevant ones for the discussion) : cfg.method = 'montecarlo'; cfg.correctm = 'cluster'; cfg.clusteralpha = 0.05; cfg.clusterstatistic = 'maxsum'; cfg.minnbchan = 2; cfg.tail = 0; cfg.clustertail = 0; cfg.alpha = 0.025; cfg.numrandomization = 500; cfg.statistic = 'diff_itc'; cfg.complex = 'absdiff'; cfg.parameter = 'fourierspctrm'; cfg.avgoverfreq = 'yes'; This also need a clustercritval : so my first question will be : -> How to choose the clustercritval in the case of this test of ITC differences ? I ve been trying different things but I m not sure. Then, this constitutes my first level analysis, so for each subject I have cluster statistics, their time/spatial course and their significance. >From here how can I go towards for a second level analysis ( similarly as in SPM) ? Is there already building blocks for that in Fieldtrip ? Should one try to find a cluster center / time window and use the cluster statistics, and use the resulting values from all subjects in a non-parametric test ? Thanks for your help, Nicolas ---------------------------------------------------------------------------- ----- Dr. Nicolas Farrugia Post-doc Research Group "Subcortical contributions to comprehension" Department of Neuropsychology Max Planck Institute for Human Cognitive and Brain Sciences Stephanstraße 1A, 04103 Leipzig, Germany Phone : +49 (0) 176 848 90 393 http://nicolasfarrugia.fr _______________________________________________ fieldtrip mailing list fieldtrip at donders.ru.nl http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From russgport at gmail.com Wed Dec 19 18:10:09 2012 From: russgport at gmail.com (Russell G Port) Date: Wed, 19 Dec 2012 12:10:09 -0500 Subject: [FieldTrip] ITC Message-ID: Hi All, Looking through the mailing list I found the following way to calculate ITC: tmpdat = freq.fourierspctrm; tmpdat = tmpdat = tmpdat./abs(tmpdat); ; % this will normalise each trial for its amplitude; itc = abs(mean(tmpdat)); % this will give the itc Is there a good way to graph this data so that I can visualize this? Thanks Russ Port -------------- next part -------------- An HTML attachment was scrubbed... URL: From marco.rotonda at gmail.com Wed Dec 19 19:48:16 2012 From: marco.rotonda at gmail.com (Marco Rotonda) Date: Wed, 19 Dec 2012 19:48:16 +0100 Subject: [FieldTrip] ROI selection in a GC analysis In-Reply-To: References: <50C844A9.6080002@donders.ru.nl> <50C84853.1050309@donders.ru.nl> Message-ID: First of all I would like to sorry to have double posted my request but before to post the send one I have checked and i had not seen my first post, quite strange. then... well Jan, thanks for the suggestion of that article... is exactly what i was looking for, I mean the aim of that article is on the same way of my thoughts Anyway... if Arjen is going to implement FARM... well that's could be quite interesting ;) Thanks to all Marco -------------- next part -------------- An HTML attachment was scrubbed... URL: From l.roijendijk at donders.ru.nl Thu Dec 20 18:31:38 2012 From: l.roijendijk at donders.ru.nl (Linsey Roijendijk) Date: Thu, 20 Dec 2012 18:31:38 +0100 Subject: [FieldTrip] ft_channelrepair with channel absent from data In-Reply-To: <50B33A53.2000701@donders.ru.nl> References: <50B23369.5020105@fcdonders.ru.nl> <50B33A53.2000701@donders.ru.nl> Message-ID: Hi all, I have also been trying to reinterpolate MEG channels that were never recorded with the nearest neighbour interpolation method, however when calling the method (in the same way Tom also used it) the function seems to do something, however it did not interpolate the missing sensors. There were still 273 sensors instead of 275. Do I need to give along a special cfg.grad or do something else? Thanks in advance, Linsey On Nov 26, 2012, at 10:45 AM, Jörn M. Horschig wrote:q > Hey Tom, > >> >> I guess I have several questions all at once: >> 1) Is it actually possible to reinterpolate MEG channels that were never recorded - eg because the channels were deactivated when the acquisition took place? > > Should work, but only with the SSI (=spherical spline interpolation) method. For the nearest neighbour interpolation, the function constructs a list of neuighbouring and constructs a weighted average. The function, however, wants to weigh each sensor according to its distance from the sensor to be reconstructed. That information can only be found in the grad structure, but in the grad structure there are only sensors which were recorded, thus the distance from a missing sensor cannot be measured because its position is missing in the grad structure. SSI should work though, cause it relies on a different rationale. > > >> 2) What might be causing the error I'm seeing? > > Since you have 301 channels rather than 275, my suspicion is that you include reference gradiometers. This should not fail in principle, though... maybe I can drop by your office to check it out? Got some time right now? > >> 3) Where is ft_fetch_sens? > > It's a private function, can be found in FieldTrip/private. You have to cd to that directory in order for Matlab to find it (the idea behind private function is that there are only visible for higher-level function and 'hidden' for all other functions). > > Best, > Jörn > > >> >> Thanks for help with any of the above! >> Best, >> Tom >> -- >> Tom Marshall, MSc. >> Neuronal Oscillations Group, Donders Centre for Cognitive Neuroimaging >> tel: +31(0)243668487 >> email: t.marshall at fcdonders.ru.nl >> postal: PO Box 9101, 6500HB, Nijmegen, The Netherlands >> visiting: Kapittelweg 29, 6525EN, Nijmegen, The Netherlands >> >> >> _______________________________________________ >> fieldtrip mailing list >> fieldtrip at donders.ru.nl >> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > > > -- > Jörn M. Horschig > PhD Student > Donders Institute for Brain, Cognition and Behaviour > Centre for Cognitive Neuroimaging > Radboud University Nijmegen > Neuronal Oscillations Group > FieldTrip Development Team > > P.O. Box 9101 > NL-6500 HB Nijmegen > The Netherlands > > Contact: > E-Mail: jm.horschig at donders.ru.nl > Tel: +31-(0)24-36-68493 > Web: http://www.ru.nl/donders > > Visiting address: > Trigon, room 2.30 > Kapittelweg 29 > NL-6525 EN Nijmegen > The Netherlands > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Linsey Roijendijk MSc. Donders Institute for Brain, Cognition and Behaviour Centre for Neuroscience dept. of Biophysics Radboud University Nijmegen Montessorilaan 3, Room B.02.34 P.O.Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3612616 L.Roijendijk at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From l.roijendijk at donders.ru.nl Thu Dec 20 18:44:31 2012 From: l.roijendijk at donders.ru.nl (Linsey Roijendijk) Date: Thu, 20 Dec 2012 18:44:31 +0100 Subject: [FieldTrip] ft_channelrepair with channel absent from data In-Reply-To: References: <50B23369.5020105@fcdonders.ru.nl> <50B33A53.2000701@donders.ru.nl> Message-ID: <6B298E85-A270-4D51-9219-F74A1C4B4AEE@donders.ru.nl> Correction: I made a mistake in my e-mail, I meant the spherical spline interpolation method, because that's the one Jorn said should work. On Dec 20, 2012, at 6:31 PM, Linsey Roijendijk wrote: > Hi all, > > I have also been trying to reinterpolate MEG channels that were never recorded with the nearest neighbour interpolation method, however when calling the method (in the same way Tom also used it) the function seems to do something, however it did not interpolate the missing sensors. > There were still 273 sensors instead of 275. Do I need to give along a special cfg.grad or do something else? > > Thanks in advance, > > Linsey > > > On Nov 26, 2012, at 10:45 AM, Jörn M. Horschig wrote:q > >> Hey Tom, >> >>> >>> I guess I have several questions all at once: >>> 1) Is it actually possible to reinterpolate MEG channels that were never recorded - eg because the channels were deactivated when the acquisition took place? >> >> Should work, but only with the SSI (=spherical spline interpolation) method. For the nearest neighbour interpolation, the function constructs a list of neuighbouring and constructs a weighted average. The function, however, wants to weigh each sensor according to its distance from the sensor to be reconstructed. That information can only be found in the grad structure, but in the grad structure there are only sensors which were recorded, thus the distance from a missing sensor cannot be measured because its position is missing in the grad structure. SSI should work though, cause it relies on a different rationale. >> >> >>> 2) What might be causing the error I'm seeing? >> >> Since you have 301 channels rather than 275, my suspicion is that you include reference gradiometers. This should not fail in principle, though... maybe I can drop by your office to check it out? Got some time right now? >> >>> 3) Where is ft_fetch_sens? >> >> It's a private function, can be found in FieldTrip/private. You have to cd to that directory in order for Matlab to find it (the idea behind private function is that there are only visible for higher-level function and 'hidden' for all other functions). >> >> Best, >> Jörn >> >> >>> >>> Thanks for help with any of the above! >>> Best, >>> Tom >>> -- >>> Tom Marshall, MSc. >>> Neuronal Oscillations Group, Donders Centre for Cognitive Neuroimaging >>> tel: +31(0)243668487 >>> email: t.marshall at fcdonders.ru.nl >>> postal: PO Box 9101, 6500HB, Nijmegen, The Netherlands >>> visiting: Kapittelweg 29, 6525EN, Nijmegen, The Netherlands >>> >>> >>> _______________________________________________ >>> fieldtrip mailing list >>> fieldtrip at donders.ru.nl >>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip >> >> >> -- >> Jörn M. Horschig >> PhD Student >> Donders Institute for Brain, Cognition and Behaviour >> Centre for Cognitive Neuroimaging >> Radboud University Nijmegen >> Neuronal Oscillations Group >> FieldTrip Development Team >> >> P.O. Box 9101 >> NL-6500 HB Nijmegen >> The Netherlands >> >> Contact: >> E-Mail: jm.horschig at donders.ru.nl >> Tel: +31-(0)24-36-68493 >> Web: http://www.ru.nl/donders >> >> Visiting address: >> Trigon, room 2.30 >> Kapittelweg 29 >> NL-6525 EN Nijmegen >> The Netherlands >> _______________________________________________ >> fieldtrip mailing list >> fieldtrip at donders.ru.nl >> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > > -- > > Linsey Roijendijk MSc. > Donders Institute for Brain, Cognition and Behaviour > Centre for Neuroscience > dept. of Biophysics > Radboud University Nijmegen > Montessorilaan 3, Room B.02.34 > P.O.Box 9104 > 6500 HE Nijmegen > The Netherlands > Tel: +31(0)24 3612616 > L.Roijendijk at donders.ru.nl > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip -- Linsey Roijendijk MSc. Donders Institute for Brain, Cognition and Behaviour Centre for Neuroscience dept. of Biophysics Radboud University Nijmegen Montessorilaan 3, Room B.02.34 P.O.Box 9104 6500 HE Nijmegen The Netherlands Tel: +31(0)24 3612616 L.Roijendijk at donders.ru.nl -------------- next part -------------- An HTML attachment was scrubbed... URL: From caspervanheck at gmail.com Fri Dec 21 13:43:38 2012 From: caspervanheck at gmail.com (Casper van Heck) Date: Fri, 21 Dec 2012 13:43:38 +0100 Subject: [FieldTrip] Source localization Message-ID: Dear Fieldtrippers, I've been trying to do a source analysis on a dataset with two conditions, mostly following the tutorials and examples from the Fieldtrip site, and come upon a problem; it won't plot. While the source analysis is not an essential component of the analysis of this dataset (it's more along the lines of "let's see if I can do this"), and I don't expect any result other than something like "there might be a source in the left hemisphere", I'd still would like to found out why it's not working. Essentially, I create a standard headmodel cause I don't have individual MRI's from the "Subject1.mri"-example, using ft_volumesegment, ft_prepare_headmodel and ft_prepare_sourcemodel, and try to get Fieldtrip to find sources based on individual EEG-datasets using ft_freqanalysis (with cfg.method = 'mtmfft') and ft_sourceanalysis. After that, I use ft_sourcegrandaverage to produce, well, a grand average source, and feed this via ft_sourceinterpolate to ft_sourceplot. Now, ft_sourceplot has three plotting possibilities: slice, ortho, and surface. Initially I tried to get the ortho-option to work, but this gave the following error: *Attempted to access dim(3); index out of bounds because numel(dim)=2.* *Error in ==> cornerpoints at 11 * *Error in ==> ft_plot_slice at 157 * So, I tried the surf-option, which gave me this: *Undefined function or variable "val".* *Error in ==> ft_sourceplot at 1174 * And finally, I tried the slice-option, which gave one of two errors (which seems to be based on how many slices I requested): *Out of memory* Which is pretty damn impressive since I've got 8GB of RAM and a pagefile topping of at 40GB Or, the mildly disconcerting: *Matlab has encountered an internal error and has to close* Which seems to have been caused by a so-called *"Segmentation violation" * Can anyone offer any insight in these errors? I think I'm doing something wrong pretty early on, but have no idea what it is. Sincerely, Casper van Heck -------------- next part -------------- An HTML attachment was scrubbed... URL: From bibi.raquel at gmail.com Sat Dec 22 01:55:36 2012 From: bibi.raquel at gmail.com (Raquel Bibi) Date: Fri, 21 Dec 2012 19:55:36 -0500 Subject: [FieldTrip] processing ANT cnt files with FieldTrip In-Reply-To: <9DF8F864-24FD-4A96-AD24-32AB4D86211C@cerco.ups-tlse.fr> References: <50D0A0FD.9020805@davebritton.com> <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> <50D0C16D.5060105@davebritton.com> <9DF8F864-24FD-4A96-AD24-32AB4D86211C@cerco.ups-tlse.fr> Message-ID: Arno & Dave, In 2010, ANT made their importer code open source, it can be downloaded at http://libeep.sourceforge.net. I hope this helps! Best, On Tue, Dec 18, 2012 at 3:44 PM, Arnaud Delorme wrote: > This is because ANT did not compile their program for your platform. > There is nothing you can do about (except maybe contact ANT so they > compile it for your machine). > Import your data on a different machine. Then you can process it on your > machine. > Best, > > Arno > -- > Arnaud Delorme, PhD > Centre de Recherche Cerveau et Cognition - UMR 5549 > Pavillon Baudot, Hopital Purpan, BP 25202 > 31052 Toulouse Cedex, France > > On 18 Dec 2012, at 11:18, Dave Britton wrote: > > > Thanks, Arno, > > I've installed the ant-eep import files into the EEGLAB plugin > directory, but EEGLAB gives me the same error message as I got with > FieldTrip: > > > > EEGLAB error in function loadeep() at line 75 > > Invalid MEX-file > '/home/dave/eeglab11_0_4_3b/plugins/anteeimport1.08/read_eep_cnt.mexa64': > libeep.so.3:cannot open shared object file:No such file or directory > > > > The libeep.so.3 file is copied into the same directory as the > read_eep_cnt.mexa64 file. > > > > Do you have any suggestions? > > -Dave > > > > > > On 12/18/2012 12:38 PM, Arnaud Delorme wrote: > >> Dear Dave, > >> > >> last time I have checked FileIO (the Fieldtrip data importer) will not > read ANT trigger files by default (it should read the raw data though - > Robert Oostenveld one of the main developer of Fieldtrip was interning at > ANT about 10 years ago and wrote the Matlab import function for them). > >> The anteepimport1.09 is a plugin for the EEGLAB software. It can import > events (and then you can save the data in EEGLAB format and reload it in > Fieldtrip). You cannot recompile the ANT project (their source files are > proprietary and not included). > >> > >> Hope this helps. > >> Best wishes, > >> > >> Arno > >> > >> On 18 Dec 2012, at 08:59, Dave Britton wrote: > >> > >>> I am trying to process EEG data in .cnt files obtained from the ANT > system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have > installed the Linux 64 bit version of FieldTrip. > >>> > >>> What is required to successfully process ANT eeg .cnt and .trg files? > The libraries that came with the FieldTrip download did not work as it came > from the download, and my investigations led me to what should be the > latest version of the ANT import library. I have downloaded the latest ANT > eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make > install, successfully, but MATLAB cannot find the shared object file it > should generate, even when I link the .so.3 file into the same directory as > the MATLAB m-file and the MATLAB current directory. > >>> > >>> Initially the system was unable to read the cnt files, which is why I > attempted to reinstall the anteepimort library. Is there more to this > installation than the standard" configure, make, make install" process? Any > clues would be appreciated! > >>> > >>> Dave Britton > >>> City College of New York > >>> Psychology Dept. > >>> > >>> Here is the error message I get: > >>> > >>>>> cfg.headerfile = > '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; > >>>>> cfg.hdr = ft_read_header(cfg.headerfile); > >>> Invalid MEX-file > >>> > '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': > >>> libeep.so.3: cannot open shared object file: No such file or directory > >>> > >>> Error in ft_read_header (line 627) > >>> hdr = read_eep_cnt(filename, 1, 1); > >>> > >>> > > > > _______________________________________________ > > fieldtrip mailing list > > fieldtrip at donders.ru.nl > > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip > -------------- next part -------------- An HTML attachment was scrubbed... URL: From r.oostenveld at donders.ru.nl Sat Dec 22 08:22:19 2012 From: r.oostenveld at donders.ru.nl (Robert Oostenveld) Date: Sat, 22 Dec 2012 08:22:19 +0100 Subject: [FieldTrip] processing ANT cnt files with FieldTrip In-Reply-To: <50D0C16D.5060105@davebritton.com> References: <50D0A0FD.9020805@davebritton.com> <0A69D0D8-967C-4F42-A6BA-99576BA16288@cerco.ups-tlse.fr> <50D0C16D.5060105@davebritton.com> Message-ID: Hi Dave, Raquels reply pointed me to the (now fully open source) version of libeep. Previous versions of the libraries and mex files were hard to compile, but I have just given it another try and after a few attempts I ended up with make distclean ./configure --enable-matlab --enable-shared=no make which compiled the mex files in a format that did not require the shared libraries. This is the best way for having them, because then you don't have to bother with the shared library path. I guess that this would also work on your 64-bit linux computer. It would be nice to have the latest version of the mex files working out of the box again, so I will file this as bug (see http://bugzilla.fcdonders.nl/show_bug.cgi?id=1913). Getting it on all three common (win/linux/.osx) platforms and in 32 and 64 bit versions will take some time though. best regards, Robert On 18 Dec 2012, at 20:18, Dave Britton wrote: > Thanks, Arno, > I've installed the ant-eep import files into the EEGLAB plugin directory, but EEGLAB gives me the same error message as I got with FieldTrip: > > EEGLAB error in function loadeep() at line 75 > Invalid MEX-file '/home/dave/eeglab11_0_4_3b/plugins/anteeimport1.08/read_eep_cnt.mexa64': libeep.so.3:cannot open shared object file:No such file or directory > > The libeep.so.3 file is copied into the same directory as the read_eep_cnt.mexa64 file. > > Do you have any suggestions? > -Dave > > > On 12/18/2012 12:38 PM, Arnaud Delorme wrote: >> Dear Dave, >> >> last time I have checked FileIO (the Fieldtrip data importer) will not read ANT trigger files by default (it should read the raw data though - Robert Oostenveld one of the main developer of Fieldtrip was interning at ANT about 10 years ago and wrote the Matlab import function for them). >> The anteepimport1.09 is a plugin for the EEGLAB software. It can import events (and then you can save the data in EEGLAB format and reload it in Fieldtrip). You cannot recompile the ANT project (their source files are proprietary and not included). >> >> Hope this helps. >> Best wishes, >> >> Arno >> >> On 18 Dec 2012, at 08:59, Dave Britton wrote: >> >>> I am trying to process EEG data in .cnt files obtained from the ANT system. I am using a Linux Xubuntu 12.04 amd64 operating system. I have installed the Linux 64 bit version of FieldTrip. >>> >>> What is required to successfully process ANT eeg .cnt and .trg files? The libraries that came with the FieldTrip download did not work as it came from the download, and my investigations led me to what should be the latest version of the ANT import library. I have downloaded the latest ANT eep library, anteepimport1.09/libeep-3.3.167, and run config, make and make install, successfully, but MATLAB cannot find the shared object file it should generate, even when I link the .so.3 file into the same directory as the MATLAB m-file and the MATLAB current directory. >>> >>> Initially the system was unable to read the cnt files, which is why I attempted to reinstall the anteepimort library. Is there more to this installation than the standard" configure, make, make install" process? Any clues would be appreciated! >>> >>> Dave Britton >>> City College of New York >>> Psychology Dept. >>> >>> Here is the error message I get: >>> >>>>> cfg.headerfile = '/home/dave/semanticflanker/data/ANT/20121114_1013.cnt'; >>>>> cfg.hdr = ft_read_header(cfg.headerfile); >>> Invalid MEX-file >>> '/home/dave/fieldtrip-20121102/anteepimport1.09/libeep-3.3.167/mex/matlab/read_eep_cnt.mexa64': >>> libeep.so.3: cannot open shared object file: No such file or directory >>> >>> Error in ft_read_header (line 627) >>> hdr = read_eep_cnt(filename, 1, 1); >>> >>> > > _______________________________________________ > fieldtrip mailing list > fieldtrip at donders.ru.nl > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip From zrhzrh at mail.ustc.edu.cn Wed Dec 26 14:16:46 2012 From: zrhzrh at mail.ustc.edu.cn (Rui-huai Zhang) Date: Wed, 26 Dec 2012 21:16:46 +0800 (CST) Subject: [FieldTrip] How make trials after remove eye artifact using ICA Message-ID: <5733894.2956881356527806760.JavaMail.coremail@mailweb> Hello everyone! I got a continue EEG file (NeuroScan cnt format) that contain so many trials which tagged with triggers. As far as I know, the best way to remove eye artifact is to use ICA method to remove them in the whole file. So I had remove the eye artifact using ICA in the whole continue file, but failed to find the convenience way to cut them into trials... Now I got a trivial solution is to load the info of sample points from the original file, and then cut the trials manually. Anybody tells me the convenience way to make trials after removing the eye artifact using ICA in the whole continue file? Rui-Huai Zhang -------------- next part -------------- An HTML attachment was scrubbed... URL: From e.maris at psych.ru.nl Sat Dec 29 00:33:40 2012 From: e.maris at psych.ru.nl (Eric Maris) Date: Sat, 29 Dec 2012 00:33:40 +0100 (CET) Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) Message-ID: <010101cde553$c6d70900$54851b00$@maris@psych.ru.nl> Dear Artemy, Thank you greatly for the response! Keep in mind that in the "Bullmore-style" cluster-mass statistic \sum_i (t_i - c_i) where t_i > c_i the thresholds c_i are constants (i.e., independent of the data). I think I see what you mean. In my case, t_i are t-statistics, and I use a p < 0.05 cutoff on both tails for the t distribution. So yes, here all the c_i would be the same for all voxels and independent of the data. Importantly, these constants also enter in the permutation distribution that is used to evaluated the significance of the maximum cluster-mass statistic, to the effect that the Bullmore-style and the Fieldtrip-style permutation distributions are shifted versions of each other. As a result, the p-values that roll out of the two approaches are identical. If I understand correctly, having the same resulting p-values could only be if the two methods assign the same rank-ordering to a given a set of clusters. But I don't think that is the case. Let's imagine that the t-statistic cutoff 'c' is equal to 1, and the data contains two suprathreshold clusters (let's say this is a spatial test and the clusters are composed of electrodes): - The first cluster has 10 electrodes, each one with a t-statistic equal to 1.1 - The second cluster has 2 electrodes, both with a t-statistic equal to 3 As I understand, Bullmore's method would assign cluster 1 a mass of 10*(1.1-1) = 1 and cluster 2 a mass of 2*(3-1)=4 , while your method would assign cluster 1 a mass of 10*1.1 = 11 and cluster 2 a mass of 2*3 = 6. Hence, given a null distribution, it should be possible to choose a cluster-based threshold that indicates as significant only cluster 1 under Bullmore's method, and only cluster 2 under yours. I think your reasoning is correct: when the data contain more than one suprathreshold cluster, my argument does not apply anymore. Your example shows that the Bullmore- and Fieldtrip-style cluster statistics have different sensitivities. Thank you for pointing this out. For every test statistic, the decisions based on the permutation p-value controls the type-I error rate, but the type-II error rate (the complement of sensitivity) depends on the exact test statistic. Best, Eric Maris Thanks again, -Artemy -------------- next part -------------- An HTML attachment was scrubbed... URL: From akolchin at indiana.edu Mon Dec 31 05:03:34 2012 From: akolchin at indiana.edu (Artemy Kolchinsky) Date: Sun, 30 Dec 2012 23:03:34 -0500 Subject: [FieldTrip] Question about cluster-based statistical testing (sum of t-stats or suprathreshold t-stats?) In-Reply-To: <50de2ca3.e90b320a.2720.73b5SMTPIN_ADDED_BROKEN@mx.google.com> References: <50de2ca3.e90b320a.2720.73b5SMTPIN_ADDED_BROKEN@mx.google.com> Message-ID: > > > Importantly, these constants also enter in the permutation distribution > that is used to evaluated the significance of the maximum cluster-mass > statistic, to the effect that the Bullmore-style and the Fieldtrip-style > permutation distributions are shifted versions of each other. As a result, > the p-values that roll out of the two approaches are identical.**** > > ** ** > > If I understand correctly, having the same resulting p-values could only > be if the two methods assign the same rank-ordering to a given a set of > clusters. But I don't think that is the case. Let's imagine that the > t-statistic cutoff 'c' is equal to 1, and the data contains two > suprathreshold clusters (let's say this is a spatial test and the clusters > are composed of electrodes):**** > > ** ** > > - The first cluster has 10 electrodes, each one with a t-statistic equal > to 1.1**** > > - The second cluster has 2 electrodes, both with a t-statistic equal to 3* > *** > > ** ** > > As I understand, Bullmore's method would assign cluster 1 a mass of > 10*(1.1-1) = 1 and cluster 2 a mass of 2*(3-1)=4 , while your method would > assign cluster 1 a mass of 10*1.1 = 11 and cluster 2 a mass of 2*3 = 6. > Hence, given a null distribution, it should be possible to choose a > cluster-based threshold that indicates as significant only cluster 1 under > Bullmore's method, and only cluster 2 under yours.**** > > ** ** > > ** ** > > I think your reasoning is correct: when the data contain more than one > suprathreshold cluster, my argument does not apply anymore. Your example > shows that the Bullmore- and Fieldtrip-style cluster statistics have > different sensitivities. Thank you for pointing this out. For every test > statistic, the decisions based on the permutation p-value controls the > type-I error rate, but the type-II error rate (the complement of > sensitivity) depends on the exact test statistic. > Thanks for confirming that up. I should note, though, that this is an issue even in data without multiple suprathreshold clusters. The same logic as above -- which shows that the two measures gives different ranks to same set of clusters -- also applies to the distribution of clusters under the null hypothesis. Thus one can imagine a single cluster in the data that would be judged significant under Fieldtrip's method and not significant under Bullmore, or vice-versa. I believe that generally, in comparison to Bullmore's method, Fieldtrip's method would tend to favor judging-as-significant large clusters (with many electrodes). I personally think of the distinction not so much in terms of controlling sensitivity, but rather as concerning the definition of what counts as a cluster of interest. Though both methods look for spatiotemoprally contiguous regions of electrodes that exceed threshold, for Bullmore the cluster is the sum of suprathreshold statistic values , while for Fieldtrip it's the sum of the entire statistic values in the region. I'm quite interested in the question of which gives more justifiable/better results in real-world settings, though unfortunately I have not seen any work done on the matter. From what I have seen in my brief forays into the extensive analytic + numerical studies of cluster-based significance testing in the fMRI literature, in that field they always refer to Bullmore-style clusters. Thanks again & happy holidays... -Artemy -------------- next part -------------- An HTML attachment was scrubbed... URL: