<html><head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head><body style="color: rgb(0, 0, 0); background-color: rgb(255, 255,
255);" bgcolor="#FFFFFF" text="#000000">
Dear FieldTrip community,<br>
I highpass-filter and downsample my data in order to prepare them for
doing an ICA. I want to apply the resulting unmixing matrix and my list
of suspicious components to the original data (before downsampling and
filtering).<br>
<br>
In the ECG artifacts tutorial [1], this is done like this:<br>
<br>
<pre class="code" style="box-sizing: border-box; overflow: auto; font-family: Menlo, Monaco, Consolas, "Courier New", monospace; font-size: 13px; display: block; padding: 9.5px; margin: 0px 0px 10px; line-height: 1.42857; word-break: break-all; word-wrap: break-word; color: rgb(51, 51, 51); background-color: rgb(245, 245, 245); border: 1px solid rgb(204, 204, 204); border-radius: 4px; white-space: pre-wrap; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">% decompose the original data as it was prior to downsampling to 150Hz
cfg = [];
cfg.unmixing = comp.unmixing;
cfg.topolabel = comp.topolabel;
comp_orig = ft_componentanalysis(cfg, data_orig);
% the original data can now be reconstructed, excluding those components
cfg = [];
cfg.component = [4 17];
data_clean = ft_rejectcomponent(cfg, <span style="font-weight: bold;">comp_orig</span>, data_orig);
</pre>
<br>
However, based on some older posts on this mailinglist I tried to
directly provide the components calculated on prepared data like
this:<br>
<br>
<pre class="code" style="box-sizing: border-box; overflow: auto; font-family: Menlo, Monaco, Consolas, "Courier New", monospace; font-size: 13px; display: block; padding: 9.5px; margin: 0px 0px 10px; line-height: 1.42857; word-break: break-all; word-wrap: break-word; color: rgb(51, 51, 51); background-color: rgb(245, 245, 245); border: 1px solid rgb(204, 204, 204); border-radius: 4px; white-space: pre-wrap; font-style: normal; font-variant-ligatures: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; orphans: 2; text-align: start; text-indent: 0px; text-transform: none; widows: 2; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration-style: initial; text-decoration-color: initial;">data_clean2 = ft_rejectcomponent(cfg, <span style="font-weight: bold;">comp</span>, data_orig);</pre>
<br>
The output for both versions seems to be identical. This is maybe not
surprising considering that the only fields ft_rejectcomponents seems to
use from the second input comp are .label, .topolabel, .unmixing and
.topo. The first three are identical for comp and comp_orig. The values
in comp.topo are however slightly different (in the order of 10^-14,
looks like a precision issue?) compared to the values in comp_orig.topo.<br>
<br>
So my questions are:<br>
a) Are the two solutions equivalent or am I missing something?<br>
b) Where do the tiny differences in comp/comp_orig.topo come from?<br>
c) Should the tutorial be altered to the simpler version or are there
scenarios in which the two versions are not equivalent?<br>
<br>
Thanks,<br>
Jens<br>
<br>
<br>
[1]
<a class="moz-txt-link-freetext"
href="http://www.fieldtriptoolbox.org/example/use_independent_component_analysis_ica_to_remove_ecg_artifacts">http://www.fieldtriptoolbox.org/example/use_independent_component_analysis_ica_to_remove_ecg_artifacts</a><br>
<br>
</body></html>