===== Artifact Removal =====

A nice artifact detection and removal walkthrough can be found at [[http://fieldtrip.fcdonders.nl/tutorial/artifacts|FieldTrip Artifact Documentation]]

==== Visual Artifact Removal ====

This section addresses visual artifact removal using the FieldTrip toolbox [[http://fieldtrip.fcdonders.nl/tutorial/visual_artifact_rejection|FieldTrip Visual Artifact walkthrough]]

Load the data and define the trials as per [[analysis:amplipex|Loading Amplipex data into fieldtrip]].  Once this has been done we can begin by setting up our cfg (configuration) structure and using the ft_databrowser function


<code matlab>
cfg = data_trl.cfg;
cfg.channel     = 'all';
cfg.trials      = 'all';
cfg.latency     = 'maxperlength';
cfg.method      = 'Summary';
cfg.keepchannel = 'yes';
cfg.metric      = 'var';
cfg.alim        = [];
cfg.coninuous   = 'yes';
cfg.viewmode    = 'vertical'; % this shows all of the channels vertically
cfg = ft_databrowser(cfg,data_trl);
data_trl.cfg = cfg; % rewrite the cfg back into the data set
</code>

The browser will let you scroll through the trials and use the cursor to highlight a section of the data containing an artifact.
{{:analysis:amplipex:artifact_wiki.png?750|}}

The selected artifact regions are not altered within the data itself, instead the start and end of each artifact are saved as part of the cfg structure: data_trl.cfg.artfctdef.visual.artifact . So now can change these regions to any value of our choice.  


<code matlab>
%% Mannual Artifact removal
art_value = 0; %Value to replace the data within artifact regions
artifacts = data_trl.cfg.artfctdef.visual.artifact;
trials = data_trl.cfg.trl;
for iArt = 1:length(artifacts)
    for iTrial = 1:length(trials)
        if trials(iTrial,1)<= artifacts(iArt,1) && trials(iTrial,2)>= artifacts(iArt,2)
            if artifacts(iArt,1)-trials(iTrial,1) ==0
                data_trl.trial{1,iTrial}(1:16,(1:artifacts(iArt,2)-trials(iTrial,1)))=art_value;
            else
            data_trl.trial{1,iTrial}(1:16,(artifacts(iArt,1)-trials(iTrial,1):artifacts(iArt,2)-trials(iTrial,1)))=art_value;
            end
        end
    end
end
</code>


Now we can visualize the data regions that have been replaced to ensure they have been correctly converted. 

<code matlab>
cfg.viewmode    = 'vertical'; % this shows all of the channels vertically
ft_databrowser(cfg,data_trl);
</code>

From here we can run any further ft analysis.