Brains, Programming Languages, Disconnected Enthusiasms

Tuesday, December 12, 2006

The State of Machine-Mediated Telepathy

What is Machine-Mediated Telepathy?
A term I've made up to describe the use of a two-way neural interface to project mental states between two people.

Has any progress been made towards this freakish machine?
Much of the Brain-Computer Interface research also contributes towards direct Brain-to-Brain communication. A fine grained bi-directional BCI might be all that's necessary. The Johns-Hopkins BCI group seem to already have Brain-to-Brain interaction on their radar. However, even mainstrean BCI research is still embryonic. Still, some of the foundations have already been laid down, making the quest for telepahy not entirely quixotic.

Is this even possible?
Transmission of mental states (thoughts, memories, emotions, etc...) between brains hinges on advances in three research areas:

  1. Brain "Reading": Neuroimaging and electrophysiology are both highly developed fields that offer researchers a dizzying variety of techniques for ascertaining brain state. Unfortunately, no single method has satisfactory temporal and spatial resolution. For a comparison of the resolution/invasiveness tradeoffs, see this chart. Not mentioned in the previous link are (1) electrode recording: highly invasive, spatially limited and temporally extremely precise (2) and EROS: non-invasive, sub-cm spatial resolution, ~100ms temporal resolution. MRS is the sole imaging technique which can discriminate neurotransmitter concentrations (I suspect this might be useful).
  2. Brain "Writing": Not writing, in the sense of permanently altering, but rather the induction of neural activity based on inputs. Aside from direct electrode stimulation, the only technology I'm aware of in this field is Transcranial Magentic Stimulation. TMS has horrible spatial specificity (could this be improved on?).
  3. Neural Correlates of Consciousness: Equally important as the mechanisms for brain access is the knowledge of what areas are relevant. I don't think it's necessary (nor possible) to fully map the interplay between brain and mind. Luckily, we could probably get by with a rudimentary knowledge. The activity of one brain will be interpreted by a second, so there's no need to make explicit what that activity means. Still, we *will* have to know which areas are relevant.
How long will it take?
Optimistically, decades. Of the three fields above, only "reading" (neuroimaging/electrophysiology) is advanced enough to be useful for a brain interface. Brain-effecting technologies will probably increase in the near future (as researchers delve deeper into bi-directional BCI), though their arrival is not certain. The correlation of conscious phenomena with neural activity has already begun in the Koch lab (as well as others). It's unclear how long it will take until such research crystallizes into a clear vision of how the brain implements conscious thoughts.

No comments: