Halfbakery: The Overclocked Brain

Alarmingly, it turns out that 50% of people have below- average intelligence. Even more alarmingly, those with above- average intelligence are not really **that** much smarter. Clearly, something needs to be done about smartness.

To the extent that smartness depends on the number of and interconnections betwixt brain cells, it's going to be tough to engineer it. We really don't know how the brain decides how many cells to have, or how to interconnect them. We have a few vague ideas about intelligence-related genes, but improving the brain's architecture is going to be decades down the line.

Howevertheless, if you can't redesign a computer, you can at least overclock it a bit. Most of the time spent in having a thought is actually all about impulses travelling along axons. This happens about 100m/s. So, if we guesstimate that the average transmission distance is 5cm (a lot less than the size of your brain, but then again neurons are somewhat clustered by function for this very reason), then the brain's clock speed is something like 2kHz. A thought that involves a mere 100 consecutive signals will take a whopping 50ms.

Now, nerve conduction is pretty well understood. The ion channels in one piece of the axon pop open, allowing a flow of ions which creates a voltage change. This change, in turn, triggers nearby ion channels to pop open, thereby propagating the wave. We know the relevant proteins down to atomic detail.

So, how do we speed this up? Well, one of the great things about cells is that they are pretty much transparent, which means we can use optical transmission to carry nerve impulses. First, we need to create a hybrid protein comprising the regular voltage-sensitive ion channel and luciferase. This will need a bit of tinkering, because we want the luciferase domain to be active only when the channel is open. Once we've done this, and bunged the new gene into someone, we'll have a brain that literally lights up while it's thinking*.

Next, we need to further engineer the protein so that it is light-sensitive. This shouldn't be so hard, since we can start with things like bacteriorhodopsin, and splice them into the protein. Again, it'll need a bit of tinkering, but what are protein engineers for?

Once we've done this, the whole process of transmitting nerve impulses will become a lot faster. As soon as an impulse starts, it will create a little burst of light. That light will tend to propagate down the axon, fibreoptically, at least for a few millimetres. And wherever the light goes, it will trigger the activation of more of the light-sensitive, light-emitting ion channel. The result will be that most of the transmission distance will be covered at 3x10^9 m/s instead of 100m/s. There will still be delays (assuming that the optical signal can't travel the whole axon in one go), but we might well expect the speed of thinking to increase by 10 or 100-fold.

[*This will, in itself, be a boon to psychiatrists, neurologists and brain surgeons - a gentle glow should be visible through the skull in a dark room, and with the skull opened you should see precisely which bits of grey goo are active at any one moment]