We have released an update of the TU Artifact Corpus (TUAR) at this URL:
It is also available from the anonymous rsync server.
We have annotated all artifact events in 310 EEG files. This is a fairly
time-consuming process that is much more difficult than it might seem.
This data should allow you to develop and evaluate your artifact
This release is what we call the first-pass annotations. They have not
been reviewed by a team of our senior annotators and adjudicated. We are
in the process of doing that and will release an update soon that will
be more accurate - hopefully by the end of August.
However, we are releasing the data now so that we can get feedback on
the data. If you are interested in artifact removal/reduction
technology, please take a look at the data and let us know what you think.