The father of the Mac Jef Raskin's belief that one day, computer users will have "direct mind input" to their machines, may be in the process of coming true.
In a giant theoretical leap over traditional input devices, brain machine interface (BMI) research is being conducted at Singapore's Institute for Infocomm Research (I2R) to fine-tune ways in which users can control computers by thought alone.
The NeuroInformatics Lab of I2R is currently demonstrating a real-time system – the NeuroComm platform – that it built from scratch to support BMI research into brain signal acquisition and analysis.
I2R integrates R&D in communications and information technology to develop holistic solutions across the ICT value chain. The Institute's research capabilities are in wireless and optical communications, and information technology and science.
"BMI provides a new communication method that does not depend on the brain's normal output pathways of peripheral nerves and muscles. It gives new hope to hundreds of thousands of people with severe disability or neurological disorders," said Lawrence Wong, executive director of I2R. "Much more than that, brain-machine-interface adds a new dimension to existing multi-modal human-computer interface, incorporating machines such as computers, robots, and other devices into 'neural space' as extensions of our muscles or senses."
BMI made the headlines recently when a medical device company Cyberkinetics announced plans to implant chips in the brains of paralyzed people to enable them to operate a computer by thought alone.
I2R, however, is pursuing a "non-invasive" approach to signal acquisition and analysis, using it in combination with electroencephalography (EEG) to improve ease of use and performance of future BMI systems.
I2R's NeuroInformatics program was launched in October last year to start up research in this area. Its NeuroComm platform supports data acquisition, signal processing, classification, feedback and graphic display, and is targeted at both research and an application development.
"BMI is a complex human-machine interactive system, so it is critical to have a real-time system in order to study various cognitive behaviors online," said NeuroInformatics Lab head Guan Cuntai.
Built from scratch on the Windows platform, the fundamental parts of the system integrated a wide range of expertise that I2R had acquired in the past, for example, in areas such as signal processing and pattern recognition.
The platform will enable I2R to perform new cognitive studies, verify new user paradigms, build demos and develop applications, said Guan.
I2R's research is focused on brain signal acquisition and analysis. "We will investigate novel approaches to signal acquisition and its combination with EEG to improve ease of use and performance of future BMI systems. We will investigate new approaches, for instance, source-model based approaches, with emphasis on facilitating human-computer adaptation and learning," said Guan.
With the NeuroComm platform I2R is hoping to address some of the challenges facing BMI research. For example, a subject's brain state during cognitive tasks keeps changing, and the brain signal varies from subject to subject even for the same mental task.
"In real-world environments, a brain signal contains a lot of noise caused by ambient interruptions and distractions," explained Guan. "Since BMI is a highly adaptive close-loop system, and the brain signal is constantly changing and complex, so one of the challenges here is to build a novel mathematical model for the brain neuro-system and for the brain signal itself."
The core part of the NeuroComm platform is built in standard C/C++, so it is easy to port it into other platforms, for instance, Linux, Unix and WinCE.
The platform will support the development of proof of concept for applications. Examples include applications that enable users to leverage on brain signals to select letters of the alphabet in order to spell words, or to control machines such as artificial limbs.
Guan admitted that BMI research is still at its infancy, but said through close collaboration with local and international organizations, I2R hopes to deliver significant research results, prototypes, and possible applications in two to three years' time.