Towards Brain Computer Interfacing

 

By Klaus-Robert Müller (Germany)

Reviewed by Alexandra Branzan Albu (Canada)

Imagine raising your left hand. Did you know that a brain-computer interface can read your intention and use it for controlling the screen motion of a cursor? The Intelligent Data Analysis group at the TU Berlin and Fraunhofer Institute FIRST is currently working on non-invasive brain-computer interfacing via computational neuroscience techniques.

Research on non-invasive brain computer interfaces (BCI) aims at translating human intentions into a technical control signal without the use of any muscle activity or peripheral nerves. All mental activities, such as visual perception, audio perception, articulating words, and intention to move, are expressed via excitation and inhibition of distributed neural structures or networks. Adequate sensors are used to record changes in electrical potentials or magnetic fields; this represents input data for the BCI.

The Brain Computer Interface developed in Berlin uses non-invasive EEG techniques to record the electrical activity of the brain. Their Berlin BCI (BBCI) is harvesting advanced machine learning, signal processing and pattern recognition technology. Their paradigm uses healthy subjects untrained for BCI, and requires a training phase and an on-line feedback session. During the training phase, subjects are asked to imagine right/left hand movements, so that their electrical activation patterns can be learned by the computer.

One of the main challenges faced by the processing of EEG signals is due to variances: single trial versus averaging, session to session variability, and inter-subject differences. This is why a machine learning approach was used in the BBCI for EEG single-trial preprocessing. 

Non-invasive brain computer interfaces will be useful for a variety of applications. They are valuable communication tools for disabled subjects, as shown by Dr. Müller in the BCI-based spelling example [1]. Other applications include control of prosthetic robots and cognitive workload assessment for predicting drowsiness on the road.

[1] B. Blankertz, M. Krauledat, G. Dornhege, J. Williamson, R. Murray-Smith, and K.-R. Müller, “A note on brain actuated spelling with the Berlin Brain-Computer Interface,” in Universal Access in HCI, Part II, HCII 2007, ser. LNCS, C. Stephanidis, Ed., vol. 4555. Berlin Heidelberg: Springer, 2007, pp. 759–768.

More information:

 

 

 

 

Berlin Brain-Computer Interface

web site

 

www.bbci.de

Return to Home Page