Oct 06 2011

Brain-Machine-Brain

It’s just a matter of time. The cyborg revolution is coming (although I won’t dare to make specific predictions about the timeline).

All the necessary basic principles have been demonstrated. We can train animals and people to operate either a robotic or virtual actuator with their thoughts alone. We can trick the brain into occupying a virtual body or “owning” an artificial limb. And now we can even provide specific sensory feedback directly to the brain – so called Brain-Machine-Brain-Interface (BMBI – I wonder if the researchers are calling it “Bambi”).

Just published in Nature Magazine is research involving rhesus monkeys that were taught to control a virtual arm with their thoughts alone. This much has been done before – various research teams are working on this technology, either involving implantable electrodes or surface electrodes. The new research, however, adds a new dimension – providing sensory feedback to the monkeys.

In this research implantable brain electrodes were used, and the monkeys were taught to control a virtual (rather than robotic) arm. These electrodes were placed in the primary motor cortex, and as with prior research the monkeys learned to control the virtual limb as if it were their own. In addition, the researchers provided sensory feedback through intracortical microstimulation (ICMS) to the primary somatosensory cortex. They had the monkey explore three different virtual objects that were visually identical but provided different sensory feedback. The virtual objects, in other words, had different texture. The monkeys were able to learn to distinguish the objects based solely on the sensory feedback.

The obvious applications are for exoskeletons (orthotics) and artificial limbs (prosthetics). It is obviously better to have an artificial limb that can feel as well as move. The sensory feedback is also critical to the process by which we feel as if we own and control a limb. If motor intention syncs with sensory feedback from vision and proprioception (feeling where a limb is in three-dimensional space), then this feedback circuit creates the sensation that we own and control the body part. Otherwise the body part feels as if it is “alien” or just an artificial part attached to our bodies.

Sensation is also critical for exploring and manipulating the world.

Obviously current technology is extremely crude, although rapidly approaching the point where it is functionally useful. There are several limiting factors. The first is that, for optimal control and “resolution” we need to implant wires into the brain. This is invasive and carries risks (although can be done effectively). External caps are safer and easier, but are not as precise.

Also, the whole thing needs to be powered, which means carrying around a heavy battery. This problem will likely be solved by devices that capture energy from biological processes (moving, breathing, etc) in order to power implantable devices. Improved battery technology wouldn’t hurt either.

And of course we also need to improve our knowledge of exactly where to place electrodes in the brain and what kind of sensory feedback to provide – in other words, how to communicate with the brain. We already have a great deal of knowledge about functional brain anatomy, but at the same time have a lot more to learn.

The plausibility of this technology is enhanced by the plasticity of the brain. It has a tremendous capacity to not only learn but to actually alter its wiring to adapt to new functions. There does not appear to be any absolute or even relative barriers to developing brain-machine interface technology. The proof of concepts have all already been done. Now it is just a matter of refinement.

16 responses so far