Hi Lakshmi,<br><br>first a note on the current sources of documentation:<div><br>* The quickest way into the toolbox are the tutorial scripts in the userscripts folder. They illustrate how to set up and run (currently over 40) different computational approaches and exercise perhaps ca. half of the available components. These also give a good idea of how to use the scripting interfaces.</div>
<div><br><div>* There is no up-to-date GUI tutorial yet (although there are some walkthroughs for the previous versions on <a href="ftp://sccn.ucsd.edu/pub/bcilab/presentations">ftp://sccn.ucsd.edu/pub/bcilab/presentations</a>). I am planning to record a series of tutorial videos, but have so far not found time for this. Thus, for the time being, the best way to get started with the GUI is after one has inspected the scripts (because then one knows what would be the next step). There is a 1:1 mapping between features available in the scripts and in the GUI.</div>
<div><br></div><div>* Each component and function of the toolbox is documented individually, including the internal functions (this help is accessible through the Help menu), adding up to 100-200 pages of text. This covers nearly all of the functionality and inner workings of the toolbox in a "bottom-up" fashion. The key functions and components of BCILAB also each have a list of short scripting examples that may provide additional in-depth help (together >200). </div>
<div><br></div><div>* If all else fails, the code of each function is documented fairly extensively and to a high standard. Thus, if you run into errors, the first place to took besides the error message is the documentation around the place where it happened and the documentation of the responsible functions. To understand the inner workings of the toolbox one would probably need to read about 20% of the function-level documentation (and to fully understand all aspects of BCILAB there is currently no better way than glancing over some of the code-level documentation).</div>
<div><br></div><div>* If you implement your own plugins (e.g. filters, machine learning methods, BCI paradigms, device I/O), you basically have to look at how the other components of the respective type are written and take a look at the documentation for the toolbox features that they are using (e.g., the arg system or declare_properties). Most of them serve as working examples of how to build own components for the toolbox. </div>
<div><br></div><div>* There is a relatively up-to-date wiki for BCILAB, which tries to give a conceptual top-down overview of how some pieces of the toolbox work together and digs down into how to write certain types of plugins (<a href="http://sccn.ucsd.edu/wiki/BCILAB" target="_blank">sccn.ucsd.edu/wiki/BCILAB</a>). Although it is perhaps the most readable document, this wiki currently covers only ~20% of the toolbox functionality (not even talking about particular methods). Turning this into a complete documentation is currently not within our reach as it basically amounts to writing a full book, for which we don't have the manpower.</div>
<div><br></div><div>* There are slides of recent talks on the toolbox (some on on earlier versions) including some theory in <a href="ftp://sccn.ucsd.edu/pub/bcilab/presentations" target="_blank">ftp://sccn.ucsd.edu/pub/bcilab/presentations</a>. It is best starting with the most recent one as the slides have evolved somewhat over time.</div>
<div><br></div><div><br></div><div>Please let me know where your sticking points are with the toolbox so that I can improve the documentation in those areas (or fix bugs if any) -- at this stage of development we are very much dependent on feedback from the community about all these issues from understanding to bugs or missing features. </div>
<div><br></div><div>Thanks,</div><div>Christian</div><div><br></div><div>PS: If you are doing research in signal processing or machine learning and just want to leverage some of the facilities of BCILAB without having to learn the framework you might also consider using just the components that you need directly (e.g., cross-validation or some classifiers or filters) and ignore the layers on top.</div>
<div><br></div><div><br></div><div><br>Am Donnerstag, 12. Januar 2012 schrieb Lakshmi Nambiar <<a href="mailto:lakshmimavila@gmail.com" target="_blank">lakshmimavila@gmail.com</a>>:<br>
><br>> Hello<br>><br>> I am doing my PhD research in EEG signal processing for BCI. I am using the EEGlab for the signal processing. When I came to know about the BCI lab toolbox, I tried using it and I am stuck.<br>
><br>> Is there any existing tutorial available for BCI lab similar to the EEGlab? Could you please share it any documents that can help me familiarize with the tool. Can can you please send some data that I can use for processing.<br>
><br>> Looking forward for further discussions.<br>><br>> Regards<br>> Lakshmi<br>>
</div></div>