Brain-Computer Interfaces

I am interested in machine learning, cognitive neuroscience, sensorimotor learning and brain-computer interfaces (among other things).

I'm currently working with Rajesh Rao on the topic of Brain-computer interfaces. The goal is to use the tiny electrical signals generated by the brain (measured via a variety of invasive and noninvasive methods) to directly communicate with the external world.

I investigate various aspects of the brain-computer interface scenario: (1) the interpretation problem, consisting of signal processing, feature selection, classification and tracking to extract information from the brain, (2) the interface problem, where we consider a variety of ways to translate this information into an effective user interface, and (3) the behavior modelling aspect, where we examine and exploit models describing the user's behavior.

Another stream of research that I am interested in is that of adaptivity. The Brain-computer interface is a "human-computer interface" but in a very immediate sense, directly utilizing the brain's plasticity to achieve communication. This scenario raises intriguing questions for the interface designer -- not just in terms of "quality", "usability" and other fuzzy terms but because the interface is a moving target, i.e., the communication protocol we use is (almost) arbitrary. Since the subject is extremely adaptable, and so (we hope) is the computer, we have to answer the question of which arbitrary protocol to use, whether the protocol should be adaptive, and finally what the dynamics of such adaptivity should be.

More detailed descriptions will come soon; in the meantime, please take a look at my publications.

Research History

Here are some of the research groups I have worked with:

• Neural Systems Laboratory at the University of Washington (current).
• VIBE Lab at Microsoft Research (current).
• Intelligent Data Analysis group at Fraunhofer FIRST, Berlin Germany.
• Database Group at the University of Washington.
• Information Sciences Dept. at Bell Laboratories, New Jersey.
• Database Group at the Indian Institute of Technology, Bombay.
• Database Lab at the Indian Institute of Science, Bangalore.
• Tata Institute of Fundamental Research, Bombay.

My previous research experience has been mainly in the area of databases. Why databases?... A cute answer from Berkeley :)

I've worked in the past on projects involving XML data manipulation -- in particular, indexing and relational exchange of XML data. This was work I did at Bell Laboratories. (Sadly, the group I worked with there no longer exists). I have also worked on problems in corpus-based schema integration, traditional query optimization, query result caching and association rule mining.