User Tutorial:Interpreting the Results

From BCI2000 Wiki
Jump to: navigation, search

In User Tutorial:Performing an Offline Analysis of EEG Data, we used BCI2000 Offline Analysis to generate a feature plot, topographies and a spectra plot for the eeg1 sample dataset comprising the data files eeg1_1.dat, eeg1_2.dat and eeg1_3.dat. In this continuation of the tutorial, we'll compare our results from User Tutorial:Performing an Offline Analysis of EEG Data to analyses of other data from the same subject.

Imagined vs. Real

BCI experiments will typically consist of sessions where the subject is asked to perform a movement followed by sessions where the user is asked to imagine making that movement. Comparing these data can help us to understand the relationship between doing and imagining. The images that follow show the results from the eeg1 data that were derived in User Tutorial:Performing an Offline Analysis of EEG Data followed by the results of an analysis performed on a different set of data recorded from the same subject. In the latter recordings, the subject was asked to imagine moving his feet or hands rather than actually performing these movements.

Results for actual movement:


Results for imagined movement:


Results for actual movement:


Results for imagined movement:


Results for actual movement:


Results for imagined movement:


As we can see, the images above are remarkably similar. This is often the case for data related in this manner.

Contralateral Response

As we've seen, it's important to evaluate the physiological plausibility of a potential feature during the selection process. For instance, we would certainly want to question a response to foot movement that is centered over the frontal lobe. Additionally, we should expect the response of the motor cortex to a given action to be contralateral to that action. The following topography reflects a subjects response to movement of the right hand. Notice that the response is focused primarily over the left hemisphere.


Next Step

While EEGs are commonly used in BCI, there has also been considerable success using MEG and ECoG as well. If you would like to learn how to use BCI2000 Offline Analysis to analyze these types of data files, please see User Tutorial:Performing an Offline Analysis of ECoG Data and/or User Tutorial:Performing an Offline Analysis of MEG Data.

See also

User Tutorial:Obtaining Mu Rhythm Parameters in an Initial Session