Revolutions in Learning: A Better Reinforcement Schedule?

Researchers of learning have been looking into schedules for reinforcement and learning for quite some time. Now, U. Texas scientists led by John H. Byrne have returned to the Aplysia sea slug to write a new lesson in optimal reinforcement schedules of learning.

Byrne's team deployed a computer to model 10,000 permutations of intervals between pulses to try to coordinate activation of enzymes and to maximize their interaction. The optimal protocol, it turned out, was not the usual, even-spaced one, but an irregular series of two serotonin pulses emitted 10 minutes apart, then one five minutes later, with a final spritz 30 minutes afterward. With this regimen, interaction between the two enzymes rose by 50 percent—an indication that the learning process was operating more efficiently.

So should you be studying Riemann sums every other day for two weeks and then take a month off before going back to them? Too early to say. The timing protocol Byrne found may be the slugs' adaptation to lobster claws crunching their tails. Studying integral calculus might be a bit different. But the implication of Byrne's work is that the best way to learn may not occur in simple time chunks—and that leaves a meaty set of new research questions for neuroscientists to pursue. "The dream of cognitive neuroscience is going from molecules to behavior by way of the brain," says Gary Marcus, a psychologist at New York University, and author of Guitar Zero: The New Musician and the Science of Learning. "This is a terrific step in that direction."

For their part, Byrne and company will now use these same techniques to try to optimize other aspects of the memory formation process in sea slugs. If that proves successful, they may eventually move on to humans. Motor skills would probably be the first target—throwing a baseball, doing the high jump, or helping a stroke victim to walk again. Science homework will have to wait. Researchers know more about the brain circuits in the cerebellum, involved with movement, than in the hippocampus, a locus for initiating the type of factual memories needed for organic chemistry.

Better ways to learn based on brain science would have enormous ramifications for educational practices. "It's not going to be an easy direction to follow because it means a lot of painstaking and detailed work to understand the biochemistry of learning," Byrne says. "But I think what it demonstrates is that if you have that information you may be able to make some big advancements in improving learning abilities by being in sync with the underlying molecular dynamics. Rather than taking cognitive enhancement drugs, you could have better training procedures." _SciAm

A sea slug is not a human, and yet an incredible amount of information about neural processes in sea slugs has proved applicable to higher animals, including humans.

This is how some revolutions begin: Lowly and slowly, then picking up speed and rising in general awareness and scale of application. These developments could eventually overturn the way that sclerotic and ineffective educational institutions operate -- which might just save the world.

Electromagnetic Brain Stimulation More Popular

Neuroscientists at the University of New Mexico asked volunteers to play a video game called “DARWARS Ambush!”, developed to help train American military personnel. Half of the players received 2 milliamps of electricity to the scalp, using a device powered by a simple 9-volt battery, and they played twice as well as those receiving a much tinier jolt. The DARPA-funded study suggests direct current applied to the brain could improve learning.

This type of brain stimulation, called transcranial direct current stimulation (tDCS), is controversial but could show promise for treatment of various neurological disorders and cognitive impairments _PopSci

ImpactLab
The wide field of electromagnetic brain stimulation is likely to prove to be a fertile area of research. Because the brain itself runs on electrical currents -- with it corresponding magnetic fields -- anything that might influence or interfere with these electrical and magnetic fields are likely to influence brain activity. But many of these researchers are discovering ways to selectively augment or inhibit particular parts of the brain, reversibly. Being able to do that safely provides an incredibly powerful research tool.

The technique, which has roots in research done more than two centuries ago, is experiencing something of a revival. Clark and others see tDCS as a way to tease apart the mechanisms of learning and cognition. As the technique is refined, researchers could, with the flick of a switch, amplify or mute activity in many areas of the brain and watch what happens behaviourally. The field is "going to explode very soon and give us all sorts of new information and new questions", says Clark. And as with some other interventions for stimulating brain activity, such as high-powered magnets or surgically implanted electrodes, researchers are attempting to use tDCS to treat neurological conditions, including depression and stroke. But given the simplicity of building tDCS devices, one of the most important questions will be whether it is ethical to tinker with healthy minds — to improve learning and cognition, for example. The effects seen in experimental settings "are big enough that they would definitely have real-world consequences", says Martha Farah, a neuroethicist at the University of Pennsylvania in Philadelphia. _Nature

And certainly, the techniques will not be used only in research and therapeutic situations. They will also be used by students, bankers, lawyers, salesmen, recreational mind trippers, sex fiends, and a wide range of individuals wanting to make more or less of themselves, depending upon their particular inclinations and needs.

We live in a foolish and dysfunctional world. But there is no reason why parts of the world cannot wake up and discover how to make itself more rational, prosperous, and fulfilled.

Mind the Enchantment

The human mind is subject to various forms of enchantment. Not a magical enchantment, but more like a trance, sometimes pleasant, sometimes not.

Because our minds are "self organised", they are subject to falling into distinctly different states, at particular "bifurcations."An illustration of this phenomenan is the "bistability" of particular images. Following the series of images above, can you say exactly where the transition occurs? What if you saw only that one image?

But the deeper you dive into the mechanisms of consciousness, the larger the number of possible mind states, so that bistability becomes tristability and so on. Just the single topic of synaptic plasticity quickly acquires a complexity to confound most scientists.

Hypnosis takes advantage of the inherent ambiguity of consciousness, and "adjusts the weighting" of various competing states of mind. Since mind is inherently a self-organizing, ongoing trance-like process, it is often likened to "riding the wave," or staying on the "bucking bronco." From the moment of waking to the release of sleep, that "blinking cursor" of consciousness compels us to provide answers and solutions, even to unknown or nonexistent problems.

For anyone who is curious about some of the underlying neurophilosophy of consciousness, I suggests looking over this article by Edelman and Tononi--two prolific and respected students of consciousness. Or look over this overview of Models of Consciousness from Scholarpedia.

Understanding human consciousness is difficult enough. But a lot of people wish to create intelligence in machines. This dream goes back hundreds, if not thousands, of years. But since the computer age beginning in the 1940s, multiple generations of ingenious scientists of mind and computation have dashed their skulls against the wall of computational complexity (not to mention a lack of understanding of the complexity of human cognition or intentionality).

Each person experiences a consciousness of an enchanted mind. Not a mind of equations and computations. Rather a mind of metaphor and narrative. An entranced mind where real world expediencies intrude on waking dreams. Complex trances of strange attractors and slippery bistable conscious surfaces.

There would be no point in trying to emulate all of that in a machine. Not unless that is the only way we can find to create a conscious machine. Perhaps it is better to settle for machines that only seem conscious or intelligent, as viewed by a simple Turing test. After all, we are only looking for help in making better decisions and devising a better world for smarter, healthier, longer-lived people.

We may be entranced, but why burden our machines with all of that? It is our trance that we wish to enjoy far into the future, not the trance of a machine.

Originally posted at Al Fin