Monkey intentions and control of a robotic arm

In a brief departure from a current catch-up I'm doing on glia, I want to catch up some old posts, thoughts, themes and developments to do with the brain as movement simulator.

In reference to blogposts The devil is in the Details (Dec 18/07), More on Learning (Dec 21/07), and The User Illusion, Dec. 22/07:

Today Deric Bownds Mindblog post featured a short piece about a big big topic:
Science Hack - monkey brain moving robotic arm. Readers can follow the links and find the great little video - as usual, a food reward proved to be the most effective for getting a monkey to neuroplasticize its brain... this monkey can't move its own arm but it learned to move the robotic one to reach food and put it into its own mouth. Remarkable. It appears that with the right motivation a brain can learn whatever it needs to.

Additional Reading:
Differences Between Intention-Based and Stimulus-Based Actions (12 page pdf)
.

Nervous System Basics VIII: PLASTICITY

Angevine's 7th attribute is plasticity:

"Plasticity
Highly reliable in a healthy person, the human nervous system has inherent modifiability, though in adulthood this attribute cannot approach that in invertebrates (moths and snails) or certain other vertebrates (teleosts and amphibians). In mammalian development, neural plasticity is striking. In continues postnatally. Abnormal visual experience at certain sensitive periods profoundly affects ocular dominance and orientation columns in the visual cortex. If an eye is closed at birth, ocular dominance columns for the other eye enlarge at the expense of adjacent blind eye columns, with thalamic fibers arriving in the cortex expanding terminal fields into them. If, shortly after birth, visual stimuli are restricted for a few weeks or even days to stripes of one orientation, cortical cells develop a response preference to lines of that orientation.

In humans, PET imaging studies of cortical blood flow show that tasks requiring tactile discrimination activate visual cortex in people blind at birth or having lost sight in childhood. This suggests that cortical connections reorganize after blindness: that afferent fibers to nearby cortical areas serving polymodal sensory integration usurp the bereft visual cortex. Such plasticity may explain the well-known tactile acuity of the blind.

In later development, neural plasticity operates on many levels, as in fine-tuning circuits to changing body dimensions. Depth perception is recalibrated as the skull enlarges and interpupillary distance increases. Even in adulthood, plasticity persists. Vilayanur Ramachandran has shown that a stroke with a cottonswab on the cheek of a young man who had accidentally lost his left arm led him to feel touch on his missing left hand. Later, the whole hand could be mapped on his face. The findings suggest that the deprived somatosensory cortical region for the hand becomes innervated by fibers from the adjacent face areas and that secondary input to a cortical neuron's broad receptive field becomes functional when primary input is lost.

After injury to the CNS, intact neurons form new terminals, by axon sprouting, to replace those of other neurons lost to trauma and thus reoccupy vacated synapses. Such reactive synaptogenesis, the clinically proven effectiveness of long-range regrowth of PNS axons, and the evident potential for axon regeneration in the CNS (as in teleosts and amphibia) hold promise for circuit reestablishment. But in mammals, these factors are thwarted by myelin debris, glial scarring, usurpation of sprouts, unresponsive injured neurons, and complex central connections. Developmental neuroscience now focuses on the cerebral cortex. The human nervous system appears to learn very rapidly by using preconstructed circuits and by locking neurons into specific types and functions after cell origin."

About that last paragraph suggesting that deliberate neurogenesis is difficult in mammals, check out this new blogpost Growing new neurons by Kevin McHenry at painonline.com:

"Wernig et al in Proc Natl Acad Sci U S A May (2008) have achieved a real breakthrough. They have been able to convert fibroblasts to neurons. These converted cells form into neurons, glia, and even dopaminergic cells. There has always been concern that converted cells might form tumors, but these scientists painstakingly separated the cells turned into neurons from pluripotential cells with fluorescent stains."

Seems like ordinary cells can be turned into neurons if they can be recoded, using appropriate transcription factors, "Oct4, Sox2, Klf4, and c-Myc"

Also, work by Peter Eriksson and Fred Gage showed that neurogenesis is intrinsic to the human brain, even in elderly people on the brink of death (see this history module, The Growth of New Neurons in the Adult Human Brain).

Neuroplasticity has been a favorite topic on this blog. It's starting to dawn on a few of us PTs that this is what "improved outcomes", be they pain reduction or increased function, strength etc, have always been all about. Here are some old posts with extensive links:

1. Neuroplasticity Dec 11/07
2. Learning Dec 12/07
3. History of neuroplasticity Dec 12/07
4. About mirror therapy Dec 16/07
5. The devil is in the details Dec 18/07
6. A few types of learning Dec 18/07
7. Cart ruts: More about UN-doing something Dec 29/07
8. It's all about movement Dec 30/07
9. And it's about brain parts: like Hippocampus Dec 30/07
10. Function only Jan 15/08
11. Smart Prosthetics, smart nerves, smart brains Feb 10/08

And it's about brain parts: like Hippocampus

In reference to It's all about movement:

From our PT perspective it's might be about movement, Matthias, freedom from pain, but I would propose that from a functional perspective there are other considerations, such as having good working brain parts, understanding their contributions to movement, to pain processing, or purpose having us move from A to B in the first place.

The hippocampus has been a riveting study focus for me this fall and winter. Buzsáki has worked for decades to understand brain waves, and noticed that theta waves, his favorite, seem to come from there. My impression from reading his book is that theta waves are like a drum beat setting a rhythm for all the other kinds of waves.

If I may, further to the discussion about types of memory, Learning to Memory, I want to post a link to a fascinating lecture video, a 2005 talk by Sue Becker from McMaster called The Role of the Hippocampus in Memory, Contextual Gating, Stress and Depression. It touches on Hebbian learning, neurogenesis, topics we took a look at in History of Neuroplasticity. Becker is building on this to examine what the learning rules might be, and build models for them. She describes the hippocampus as a large convergence zone, where information from lots of other parts is "coded". Neurogenesis takes place here. A constant supply of new neurons seem to be necessary for coding memories over time, over temporal gaps. If neurogenesis is slowed by stress, new connections have trouble being made. (Sapolosky has mentioned this as well.) New neurons remain plastic (able to make new connections) for longer than old ones.

Single cell recordings of spatial coding cells in the hippocampus have been made with human subjects, as they move about in a virtual world (prior to this only rats had been examined). She touches on spatial hemi-neglect, neglect of the left half of a person's environment following a type of stroke; she is working on building a model that can account for both "egocentric" (sensory/self) and "allocentric" (other/outside) spatial coding.

Her third area of investigation is the role the hippocampus as a comparer and coder of personal behavior. This brings in the role of context. The hippocampus is crucial for determining context of a situation, and allow you to react appropriately in a stressful condition. Here we come back to producing movement again. If the hippocampus isn't working, how will you know what movement to choose? The hippocampus may exert a modulatory effect on other parts of the brain.

Also, here is a link to one of Ginger Campbell's brain science podcasts (Episode #3) which discusses Eric Kandel's book, In Search of Memory. In episode #12 she discussed another book, Memory: From Mind to Molecules, by Larry Squire and Eric Kandel.

The internet is absolutely full of great information.

Just do it

Reply to: A Few Types Of Learning

Diane,

the only additional form of learning I can think of is procedural learning.

It's basically learning by doing. I never learned to type - but after a few years writing emails and other stuff I have become fairly good at using the keyboard.

I think the same thing happens in sports - when you ask someone how he or she does a certain movement they are unable to tell you - they just do but can't explain why or how.

This type of learning requires a lot of practice over time - so it has limited use in treating chronic pain I think.

But: maybe this non-conscious type of learning is one of the causes of chronic pain?!

One of the most important things you have to do when in pain is to try to get rid of things that make the pain worse or cause it in the first place.

But when you react non-consciously towards certain stimuli that's a problem - because you don't become aware of this negative influence.

In this case you first have to develop the necessary mental skills - namely Metacognition - to be able to watch yourself - then change the behaviors that aren't helpful.

Just do it. ;-)

A Few Types of Learning

In reference to The Devil is in the Details:

Learning seems to be a recurring theme in this blog, so let's take it even further. Kandel says, "learning is in the synapses". How many kinds are there? Probably more than I have listed here, but this is a start:

1. Long Term Potentiation

Lately I've become caught up with Geörgy Buzsáki and his book Rhythms of the Brain. In chasing down things I've spotted in that book, namely the role of the hippocampus, the connection between theta rhythm and movement, orientation (space maps), memory and learning, it seems long term potentiation is A main if not THE main mechanism. "Learning" requires receptor breakdown and building, a process that is going on all the time anyway.. but when a signal comes in, repeatedly, the receptors will strengthen around the signal. (This is true for pain reduction, which is another sort of "learning", extinction learning, point 5.)

There is a very nice article from "Go Animal", a website linked in this blogpost by the same name. It's called "The Greatest Discovery You've Never Heard Of" by physical educator Frank Forencich.

Longterm potentiation (LTP) relates back to one of the other "tenets of neuroplasticity",

4. Initial changes are just temporary. While the brain can learn through impact (a powerful experience), usually it learns through lots of repetition.

In fact, we likely do NOT want to have to learn too many things through impact - such learning would tend to retain a heavy emotional overlay I should think, which could mess up integration, create large "hooks" for future learning of an inappropriate or counterproductive sort. Remember this other tenet:

5. Brain plasticity is a two-way street; it can change itself in positive or in negative directions. E.g., chronic pain, bad habits

... which leads us to another form of learning,


2. Sensitization

We know about this one from pain science, but its origins are with Kandel according to this Wiki page. This is described as short-term memory formation.

One of the most troubling (and confusing) aspects of pain states is mechanoreceptor sensitization, discussed in the Textbook of Pain, 5th ed.. This has led practitioners on merry goose chases after "the tissue at fault" (structure as opposed to function) for decades, centuries...

Sensitization can be "un"-learned, and quite rapidly, by something as easy as mirror therapy.

To overcome sensitization, and to utilize longterm potentiation, a "graded exposure" approach is usually recommended - this is a cognitive-behavioral approach which breaks down a learning or an un-learning movement objective into small, easily assimilative bits.

This ties in with

3. Habituation

This type of learning seems to be about just adapting to (learning to ignore) something in the environment or inside, consciously or unconsciously, be it initially positive or negative, although psychology definitions hold out for "decline in the response elicited by repeated stimulation, not due simply to adaption". My pick is "A decrease in the behavioural response to a repeated, benign stimulus." This would be best avoided in treatment of pain. I doubt we want our patients to "habituate" to either us OR our treatment. Which is why we should stay nimble and novel, in my opinion.

Habituation is classified as non-associative in psychology, which takes us to ..

4. Learning by Association, or Classical Conditioning

Here, the relationship between a stimulus and a response is strengthened, or (see above) "sensitized".

Here is one more:

5. Extinction Learning
We are now back with definitions that involve proteins at the neuron level of memory.

While this little list of definitions hops around among scholarly domains and perspectives, it is important to realize that, as per Kandel, nothing can happen with learning and memory regardless of domain, unless neurons and their receptors change. When dealing with patients whose overriding issue is persisting pain, the "locus of control" (another one of those catchy psych terms) must be with the patient at all times. They must be given ample opportunity to "learn" to downregulate their not-very-successful prior sensitization learning, given support, and sufficient time, especially if manual treatment (a form of exteroceptive neuromodulation) is involved.

Matthias, can you think of any other types of neural learning to add to the list?

History of Neuroplasticity

Matthias, I agree. And thanks for introducing Dean Kamen. Seems pretty clear his work designing arm prostheses that can "feel", will make the wider PT/OT rehab role lots easier. :)

You highlighted him as a good example of someone who is not afraid of "learning" and gave some of the important characteristics of that:

"creativity, an ability for lateral thinking, the ability to make connections between seemingly unrelated fields (of science) - and disrespect for authority."

I agree although I would translate "disrespect for authority" into "an ability to recognize and bypass conventional dogma or "groupthink" in favor of considered scientific reflection of all possibilities."

On the topic of neuroplasticity, Sharon Begley, in her book Train Your Mind, Change Your Brain, recounts how difficult it is for dogma to be overturned in favor of counter-information. She traces the history of the scientific study of neuroplasticity starting with William James, who said, in 1890: "organic matter, especially nervous tissue, seems endowed with a very extraordinary degree of plasticity", meaning, according to Begley, "a structure weak enough to yield to an influence". No one picked up on this at the time. Ramón y Cajal was in the forefront of neuroscience, and had dazzling, substantive, convincing and irrefutable research work to offer up. His opinion on fixedness, immutability, unchangingness in the nervous system was pretty much conflated with his actual research for the next century.

In the 1940's, Donald Hebb at McGill came up with

"When an axon of cell A is near enough to excite cell B and repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells such that A's efficiency, as one of the cells firing B, is increased"

..which became known as Hebbian learning, and gave rise to the quip, "Neurons that fire together wire together."

Along came researchers like Fred Gage (a shirttail relative, apparently, of the famous Phineas, according to Begley), who runs a lab at the Salk Institute in La Jolla, California.

Starting in the 1960's, other researchers, like Joseph Altman at MIT, Michael Kaplan at Boston U., and Fernando Nottebohm who studied birds that produce new songs continually throughout their lives, William Greenough, showed evidence that went against dogma, but weren't able to budge it. Elizabeth Gould showed in the 1990's that primates demonstrated neurogenesis. Slowly, the dogma finally started to dissolve.

Then researchers tried to figure out how to show neuroplasticity exists in the human brain - they wanted to know not just if the human brain could rewire itself, but also if it could grow new cells. It was hard to deal with all the ethical dilemmas involved, but they finally got permission to look at brains of a group of people who were terminally ill, at autopsy. Specifically, Peter Eriksson and Fred Gage worked together to determine that new neurons are born constantly, even in the brains of old and sick humans. You can read more about how the dogma was made to melt down in this history module.

Suffice it to say, human neuroplasticity is a fact of life, and cannot be pushed back into the shadows. From a physiotherapist point of view, this is good news! It shows that our work, i.e., helping people, helping patients in pain, bolstering them and supporting their efforts to improve, now has irrefutable science to back it. Now we can show people studies (should we have to) to prove they can recover from even the most debilitating injury, but even more so we can turn around and challenge various dogmas that still exist right in our own profession, dogmas about pain and how best to help relieve it, help move the profession from one that is still mesmerized by structure into one that smoothly understands and supports function.

Learning

Diane, I think that one of the most important statements from that program is this:

1. The focus, the inner attention, has to be on the process of learning the action, not the desired action itself.

That means that you have to learn how to learn.
If you develop good learning skills - you can apply those anywhere and everywhere - all your life.

If you look at a genius like Dean Kamen - a great inventor - you will see what I mean.
He is creative and simply doesn't give up.

All geniuses throughout history have shown some or all of the same skills:
creativity, an ability for lateral thinking, the ability to make connections between seemingly unrelated fields (of science) - and disrespect for authority.

Only by letting your mind wander to places that don't exist - by freeing yourself from the constraints society and/or peer pressure impose upon you can you really become creative and innovative.

In this case it's more important to un-learn constraining patterns of behavior than to learn new ones!

So where does pain fit in?

Learn to observe yourself - use Metacognition to identify situations, thoughts, habits, .... that aren't helpful. Start today.
Because if you learn to develop this crucial skill to it's full potential, you can use it - as stated above - anywhere, everywhere, anytime.

You have to realize that chronic pain is (mostly) a learned condition. The brain learned to be in pain all the time on it's own - what's called maladaptive plasticity.

But - by learning how to learn - and thus making use of "good" neuroplasticity - one can unlearn chronic pain.