The Age of Mind-Controlled Computers

A guest post from new blogger Simon Oxenham. Help me to welcome Simon onto SV!

For those not yet seduced by the relentless march of technology, the letters "BCI" may mean very little. For the technophiles out there, the letters represent a bubble of potentially earth shattering proportions just waiting to burst. For the families of individuals with severe paralysis desperate to hear their silent loved ones' wishes, these same three letters represent a ray of light on the horizon. Electroencephalography (EEG) based Brain Computer Interfacing (BCI) is an emerging technology that uses electrodes attached to an individual's scalp to map brain activity. This allows an individual to control a computer just by thought alone, for example by imagining that their leg is moving.

One of the most effective imagined mental activities that is used with BCI is motor activity, which is simply "imagined movement" in laymen's terms. We have known for a long time that separate body areas are controlled by separate parts of the brain, in a pattern known as the homunculus (latin for "little human"). Furthermore, we know that both the left and right brain hemispheres control opposing sides of the body, so if motor imagery can be generated for a localised area of the brain, then the computer can tell us which side of the body is being imagined. According to the sensory homunculus model, the hands have the largest relative brain area. Thus, imagined hand movement is typically the first mental activity used to control brain computer interfaces. This imagined hand movement is currently the focus of my own research in BCI.

Ever since BCI first began making waves in the biomedical field, a major problem for clinical application of BCI has been the time consuming and frustrating training that is needed to achieve accurate control of the computer through thought alone (Huang et al, 2011). to sidestep this problem, myself and my colleagues are attempting to pioneer a simple yet effective psychological approach that harnesses our current understanding of a psychological theory called mirror neuron theory. Mirror neuron theory (Ramachandran, 1996) is based on the concept that similar brain areas are activated when a physical movement is watched as when it is actually made by the individual. The theory has famously been used to relieve the pain caused by phantom limbs, which typically occurs when an individual loses an arm. Ramachandran (1996) found that by using a "mirror box" to give the illusion that a severed limb is present again, the pain associated with the phantom limb is relieved.

In our BCI study, participants and an experimenter must wear matching black gloves and sweaters for reasons that will soon become clear. In the first trial, the system is calibrated using the traditional BCI method, which involves the experimenter asking the participant to follow the computer's cue and imagine the required hand movement. The results in this first trial are typically poor, and we find very weak levels of synchronised brain activity in this condition. In the second trial, the system is calibrated by having the participant do precisely the same mental activity, but at the same time the experimenter places his/her arm under the participant's and in front of them, giving them the illusion that their (the participant's) arm is moving when in fact it is stationary down by their side. While the participant imagines their hand moving, the experimenter moves his/her hand in the precisely the way that the participant is imagining. While this is happening the participant imagines that the experimenter's hand is their own hand. The black garments and gloves help make the illusion feel as real as possible.

According to mirror neuron theory, realistic simulation of arm movement causes corresponding neurons in the motor area to fire (Ramachandran, 1996). Preliminary results appear to demonstrate greatly enhanced synchronised brain activity in the condition when the participant's arm is simulated using an experimenter's arm. The implication of this finding is that we are able to use this method to build a powerful tailored model of a specific brain activity pattern in only three minutes. It appears that this simulation-based model is both similar and more powerful than the model that is generated when an individual only imagines a specific type of brain activity. An analogy is that rather than drawing a map from scratch we are tracing the highways but manually defining the side streets, a much faster and easier process indeed. The beauty of this method is that once the simulation-based calibration model is created, the experimenter (or the experimenter's arm at least) is no longer needed and the individual is able to continue alone. I must note here that this is only a preliminary finding in a single pilot study and that these results should not be considered definitive until more research is conducted. Others may disagree, but after witnessing BCI in action, I am confident that the field of BCI promises to be a staggeringly fruitful area of research over coming years.

References:

Huang,D. Qian, K. Oxenham, S. Fei, S. Bai, O. (2011) Event-Related Desynchronization / Synchronization-Based Brain-Computer Interface towards Volitional Cursor Control in a 2D Center-Out Paradigm. Institute of Electronics and Electrical Engineers (IEEE) Symposium Series on Computational Intelligence. Cognitive Algorithms, Mind and Brain. Conference Proceedings. Pgs. 151-158 (PDF)

Ramachandran, V. (1996) Synaesthesia in Phantom Limbs Induced with Mirrors. Biological Sciences, Vol. 263, No. 1369, 377-386 (Jstor)