Differences

This shows you the differences between two versions of the page.

--- workshop:meg-uk-2015:fieldtrip-connectivity-demo [2018/10/21 15:13]
42.49.180.224 [FieldTrip connectivity demo]
+++ workshop:meg-uk-2015:fieldtrip-connectivity-demo [2017/08/17 11:21] (current)
127.0.0.1 external edit
@@ Line 7: / Line 7: @@
 Please use the [[general|general MATLAB instructions]] to get started.
-====== FieldTrip connectivity demo ======
+===== Part 1 - virtual channel connectivity =====
+<code matlab>
+%% start with data that was preprocessed in FieldTrip
+subj = 15;
+prefix = sprintf('Sub%02d', subj);
+load([prefix '_raw']);  % this is called "data" rather than "raw"
+% load the results from beamformer_part1
+load vol
+load sens
+load mri_realigned
+%% deal with maxfilter
+% the data has been maxfiltered and subsequently contatenated
+% this results in an ill-conditioned estimate of covariance or CSD
+cfg             = [];
+cfg.method      = 'pca';
+cfg.updatesens  = 'no';
+cfg.channel     = 'MEGMAG';
+comp = ft_componentanalysis(cfg, data);
+cfg             = [];
+cfg.updatesens  = 'no';
+cfg.component   = comp.label(51:end);
+data_fix = ft_rejectcomponent(cfg, comp);
+%%
+pos1 = [21 -64 30];
+pos2 = [0 35 83];
+cfg = [];
+cfg.location = pos1;
+figure; ft_sourceplot(cfg, mri_realigned);
+cfg.location = pos2;
+figure; ft_sourceplot(cfg, mri_realigned);
+</code>
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.10.png?300|}}
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.26.png?300|}}
+<code matlab>
+%%
+% timelock2 was computed in http://fieldtrip.fcdonders.nl/workshop/meg-uk-2015/fieldtrip-beamformer-demo#part_3_-_reconstruct_single-trial_cortical_responses
+cfg             = [];
+cfg.vol         = vol;
+cfg.grad        = sens;
+cfg.senstype    = 'meg';
+cfg.method      = 'lcmv';
+cfg.lcmv.keepfilter = 'yes';
+cfg.lcmv.projectmom = 'yes';
+cfg.grid.unit   = 'mm';
+cfg.grid.pos    = pos1;
+source1 = ft_sourceanalysis(cfg, timelock2);
+cfg.grid.pos = pos2;
+source2 = ft_sourceanalysis(cfg, timelock2);
+%% construct single-trial virtual channel representation
+virtualchannel_raw = [];
+virtualchannel_raw.label = {'pos1'; 'pos2'};
+virtualchannel_raw.trialinfo = data_fix.trialinfo;
+for i=1:882
+  % note that this is the non-filtered raw data
+  virtualchannel_raw.time{i}       = data_fix.time{i};
+  virtualchannel_raw.trial{i}(1,:) = source1.avg.filter{1} * data_fix.trial{i}(:,:);
+  virtualchannel_raw.trial{i}(2,:) = source2.avg.filter{1} * data_fix.trial{i}(:,:);
+end
+%%
+cfg                 = [];
+cfg.keeptrials      = 'yes';
+cfg.preproc.demean  = 'yes';
+cfg.preproc.baselinewindow = [-inf 0];
+virtualchannel_avg = ft_timelockanalysis(cfg, virtualchannel_raw);
+figure
+plot(virtualchannel_avg.time, virtualchannel_avg.avg)
+legend(virtualchannel_avg.label);
+</code>
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.31.png? |}}
+<code matlab>
+%%
+cfg         = [];
+cfg.method  = 'wavelet';
+cfg.output  = 'powandcsd';
+cfg.foi     = 4:2:70;
+cfg.toi     = -0.200:0.020:1.000;
+virtualchannel_wavelet = ft_freqanalysis(cfg, virtualchannel_raw);
+cfg                 = [];
+cfg.baselinetype    = 'relative';
+cfg.baseline        = [-inf 0];
+cfg.channel         = {'pos1'};
+figure; ft_singleplotTFR(cfg, virtualchannel_wavelet);
+cfg.channel         = {'pos2'};
+figure; ft_singleplotTFR(cfg, virtualchannel_wavelet);
+</code>
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.36.png?300|}}
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.39.png?300|}}
+<code matlab>
+%%
+cfg = [];
+cfg.method = 'coh';
+coherence = ft_connectivityanalysis(cfg, virtualchannel_wavelet);
+figure
+imagesc(coherence.time, coherence.freq, squeeze(coherence.cohspctrm(1,:,:)));
+axis xy
+</code>
+{{:workshop:meg-uk-2015:screen_shot_2015-01-05_at_12.02.54.png? |}}
-In this demonstration we will use the [[dataset|face recognition dataset]].
-Please use the [[general|general MATLAB instructions]] to get started.
 ===== Part 2 - whole brain connectivity =====