A Paradigm Shift for the Motor Cortex?

By Neuroskeptic | December 11, 2015 5:56 am

Many people will be familiar with this rather strange image:


It’s a depiction of the motor homunculus, which is essentially a “map” of the body located in the brain. The image shows how different spots of the primary motor cortex control different parts of the body.

So, for instance, the spot I’ve highlighted in red corresponds to the muscles in the thumb. If you were to stimulate this spot, say using an electrode, it would cause the thumb to twitch. By stimulating different points and seeing what happened, the neurosurgeon Wilder Penfield first mapped the motor homunculus in 1937 and this account has become the orthodox view of how the motor cortex is organized.

But now, some neuroscientists argue that Penfield got it wrong. Michael S. A. Graziano of Princeton was the first to propose the new model, and he reviewed the evidence for it in a new paper in Trends in Cognitive Sciences. He’s not shy about explaining the importance of his theory: in the title of the paper he calls it “A Paradigm Shift for the Motor Cortex.”

According to the alternative view, the primary motor cortex doesn’t contain a neatly layed-out map of the body, and the different points don’t correspond to different muscles. Rather, the claim is that the cortex contains an ‘ethological action map’ – i.e. different areas encode different actions.

An action, in this sense, is a whole series of muscles acting in sequence, leading to a coordinated movement such as “raise hand to face”. Actions are far more complex than the single muscle contractions that the motor cortex encodes according to the orthodox view.

Here’s what an action map looks like, according to Graziano:


The colored spots on the brain are points on the motor cortex of a monkey. , Graziano says that stimulation of the different points evokes “complex actions that appear to come straight from the animal’s normal repertoire”, such as “reach to grasp”, “defense”, and “climbing/leaping”

On the other hand, Graziano adds that Penfield was partly right: there is a map of the body in the motor cortex, but it is distorted and fragmented because it has to share the space with the action map.

If that’s the case, how did no-one notice this before? Why was the action map only discovered in 2002 by Graziano’s team?

The trick, Graziano says, is that in order to evoke complex actions, you have to apply a prolonged electrical stimulation to the same spot – from 500 ms up to 1 second. Previous researchers, including Penfield, used very brief stimulation pulses, of perhaps 50 ms. Graziano says that these brief pulses only cause twitches (the beginnings of actions, perhaps?) and that the action map can only be uncovered with longer pulse trains.

Not everyone is convinced by the action map model however, and the long pulse trains are the main bone of contention. Skeptics, such as Paul D. Cheney, argue that prolonged stimulation causes activation to “leak” and spread over a wide area, triggering multiple muscles to move. Thus, the stimulation creates an artificial “action circuit”, rather than activating one that normally exists in the brain. Cheney calls this the “neural hijacking” hypothesis.

Graziano disagrees, saying that there’s a host of convergent evidence that confirms the theory, including data from experiments using “optogenetic stimulation, chemical stimulation, chemical inhibition, surgical lesions and cortical reorganization during recovery” as well as “the specific match between stimulation-evoked actions and neuronal response properties, the specific match between stimulation-evoked and natural behaviors”.

ResearchBlogging.orgGraziano MS (2015). Ethological Action Maps: A Paradigm Shift for the Motor Cortex. Trends in Cognitive Sciences PMID: 26628112

  • https://www.youtube.com/user/AmoralAtheistChannel/videos non_sig

    Hm, that is very interesting. I don’t don’t know what my opinion is, but what do they mean with “The greatest concern was that the stimulation might accidentally spread through connections within the motor cortex or among other parts of the motor system, thereby blurring and contaminating the investigation of the descending pathway” The activation does spread, doesn’t it? I thought that is why the eye blinks or the mouth moves or something like that, even if that area was not stimulated?

    And, I don’t know, but since the complex actions they describe are probably actions the monkeys do a lot and practice a lot, it maybe would make sense that these connections are stronger than others, and are more likely to be effected by an “accidental spread”?

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Regarding spread, I think the debate is over whether the stimulation causes “natural” spread along pathways that already exist, as opposed to “artificial” spread that is purely a product of the un-physiological nature of the prolonged electric stimulation.

      But I agree with you that the line between these two is a bit blurred because even artificial stimulation might preferentially travel along existing pathways (maybe along links formed through experience.)

    • feloniousgrammar

      My first thought was that the monkey’s thumb was simply the first thing that caught the researcher’s attention and that because the stimulation was so limited, that’s all they saw.

      It is fascinating. If there is a homunculus. There is no THE homunculus— a little mi in our heads. But the new paradigm makes sense to me, and makes more sense to me than the model based on isolated body parts twitching independently.

    • Anonymouse

      Even though he agrees that the stimulation is somewhat artificial (as is always the case in these kinds of experiments), Graziano actually devotes a fair chunk of his article to elaborate on why the “neural hijacking” hypothesis is very questionable, which include
      – replications of the finding with different methods of activation (chemical disinhibition and optogenetic)
      – consistent outcomes of experiments inhibiting or lesioning the respective areas
      – neuro-imaging suggesting that the activation spreads along existing neural pathways, rather than randomly
      – the particular behaviours that don’t appear to be random combinations of muscles that happened to be close in the motor cortex, especially since spatial deviation doesn’t result in partial loss of the movements (e.g. simply not turning the torso anymore for reaching)
      – a toy simulation of self-assembling maps that shows similar organizational patterns as the motor cortex (areas that are overlapping, rather than distinct with the occasional reversal and fractured area) when constrained not only somatopically, but also functionally and spatially (by the destination of the movement)

      – experience-dependent changes in the cortical organization when the animal is prevented from doing certain behaviours

      Overall I find this account very compelling and much more sensible than assuming basically nothing but an inborn anatomically driven organization that only allows for very a limited kind of plasticity (more or less area devoted to each body part). Presumably there are more constraints that underlie the organization, but this is a great start.

      Very glad you covered this, Neuroskeptic.

  • Audrey

    I have to agree that the original map by Wilder Penfield is
    too simple. It makes sense to me that there would be a motor map and action
    map, among other things. In Science Olympiad there is a category called
    Write-it. Do-it. One partner wrights all
    the instructions to a simple task such as making a peanut butter and jelly
    sandwich. Then a second partner makes the sandwich exactly according to the
    directions, and nothing else. This is used to demonstrate how precise one must
    be when writing their experiments down with the scientific method. However I
    think it can also translate to our brains. The complexity of our brains is hardly
    comprehended by them. The simple clarity of single paths with such detail as to
    make a sandwich would explain our movements. I am not saying there is a single
    path for every single movement we make, but that there could be trees of
    movement throughout our brain. As an example your right arm movement could have
    a tree, if you want to stimulate the right arm you must do so in the area
    mapped for the right arm, however as the complexity of the movement increases
    different paths are taken along the branches of the tree until the movement is

    To see if I can put it in the way of making the sandwich:
    The first partner has every branch of movement available to him to make a
    sandwich. The path he chooses determines the outcome. If he decides not to dip
    the knife in the peanut butter then there will be no peanut butter on the

    Meaning if (you) your brain does not decide to take the branch
    that crosses your right arm over the center of the body then it will be
    impossible for it to touch your left shoulder.

    I think while it is
    possible the prolonged exposure to electrical stimulation could be spreading to
    other reactions, it makes more sense that there are laid out paths for specific
    actions. And as non_sig suggested the pathways chosen by the electrical
    currents to produce a certain movement may be the ones more commonly used by
    the animals. However what that tells us is that these pathways do in fact
    exist. Commonly used or not, the path in your brain to move your right hand to
    your left shoulder is there. And there is a separate path to move your right
    hand to your right shoulder.

  • Pingback: Acute off-target effects of neural circuit manipulations()

  • Private_Eyescream

    Crude. Why not modulate the pulse signals with known and computer generated spike patterns in the same location (analog logic probe with prerecorded neural impulses) (digital logic probe with simulated neural impulses following a Grey Code binary pattern)? It’s like they haven’t even learned basic computer hardware hacking techniques. Obviously a stim to a neural location will enact a twitch or an action, but the more interesting hack is seeing if identical analog or digital signal patterns are bulk-matching to groups of humans per motor function usage.
    The second part is seeing if “error twitches” get identical brain correction signal spike patterns. Because the human brain begins to remap on multiple error codes in muscles, knowing what the patterns are would allow faster neural remapping in muscle therapy or after injury to speed healing and would be critical in basic cybernetic replacement limbs.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar