New? Uncheck. Useful? Uncheck. Accurate? Uncheck.
If you're going to spend money on a project, you normally want to be sure it's new, useful, and accurate.

Here's a project guaranteed to miss all of those goals:

Leading neuroscientists are joining forces to study the brain — in much the same way that physicists team up in mega-projects to hunt for new particles.

The International Brain Lab (IBL), launched on 19 September, combines 21 of the foremost neuroscience laboratories in the United States and Europe into a giant collaboration that will develop theories of how the brain works by focusing on a single behaviour shared by all animals: foraging.

So much for the ACCURATE. Foraging is NOT shared by all animals. It's certainly common among our familiar furry and feathered friends. Chordates, mollusks, insects. But it's NOT common in many phyla that live on the ocean floor, waving tentacles and filtering food. No intelligence needed for filter-feeding. Even among the smart set, some types have chosen filtering. Clams, barnacles, tunicates, sperm whales. Whales are plenty smart, but some of them don't bother to forage. Just cruise along with your mouth open and let the filter do the work.

The pilot effort is an attempt to shake up cellular neuroscience, conventionally done by individual labs studying the role of a limited number of brain circuits during simple behaviours. The ‘virtual’ IBL lab will instead ask how a mouse brain, in its entirety, generates complex behaviours in constantly changing environments that mirror natural conditions.

Is this new? Well, I guess you could say that simulating only a tiny fraction of what SOME animals do, and getting it incurably wrong, is new. Yes.

All of this has been done before, VASTLY better and VASTLY more precise and complete.

Why better? Because the earlier efforts were ANALOG.

The real system is ANALOG, with INFINITE layers of excitation, inhibition, feedback upon feedback upon feedback, delta detection, delta-squared detection, delta-cubed detection, smoothing, integration, memory, and above all PURPOSE.

LIFE IS PURPOSE.

You can do all of those things with a few components when the components are REAL AND ANALOG. Everything is simultaneous in an analog system.

You can't even properly simulate one feedback loop with software, because the software is always operating in STEPS OF TIME and STEPS OF VALUE. Nothing is simultaneous.

Here's a purely wonderful series of articles in British Practical Electronics mag, all by G.C. Brown, laying out a project that was actually built in a mechanical 'animal', and showing the reader how to build the same 'animal'. Typical of the period, it was a ground-dwelling wheeled critter, an ancestor of today's Roomba.

Just listing some of the topics to give a sense of the TOTAL coverage:

11/68 intro, conditioned reflex, cutting out distraction

12/68 Excitation/inhibition/feedback, memory, delta, touch sensor, light sensor, avoidance. (Also includes the Stock Game that I mentioned earlier, which is a separate example of an analog 'learning' ckt.)

2/69 SELF-RECOGNITION with a ckt to do it, mutual recognition, auditory perception

3/69 Building the animal, work and reward!!!



There's your PURPOSE.


Each of these articles includes a wonderfully clear and precise and ACCURATE description of how real animals perform the specified function; a block diagram of how to simulate it, and an ACTUAL BUILDABLE CIRCUIT that does the function. Each of the actual circuits has around 10 transistors and 20 R and C passive components. The final result is complex, but it's no worse than something like an oscilloscope.

These articles could be used to develop a lab course in REAL neurology and REAL social science. The ckts would need to be kit-ified** using breadboarding systems like Snap Circuits. Teamwork and discussion could lead to understanding both neurology and philosophy, far more accurately than academic software or academic "philosophy".


= = = = =

**Kits: Needless to say, Kit 10 would be a module that causes the machine to extend its front sensor repeatedly and beep randomly when a button on the dorsal side is pushed.

Labels: Asked and badly answered, Experiential education, Grand Blueprint

  ¶ 3:55 PM