How Does the Brain Process Movement: Calculation or Intuition?

[KingNothing]

Here is the question before you today, ladies and gentlemen: Do our brains calculate, or do they measure with their own methods?

I posed myself this question when thinking about subconscious estimation...like when you catch a baseball, you don't know exactly how fast its going, but in under a second your brain calculates about how fast its going, so you can catch it.

Now, the question applied to this problem would be:
Does your brain actually use the laws of projectile motion, and a certain distance with a certain time, etc, to subconsciously calculate and tell you how and when to catch it?

--OR--

Are the laws of projectile motion and velocity, etc actually what your brain uses...or are they just ways that we have created to explain how things work..and maybe our brain uses some other method (that probably cannot be mathematically modeled) to figure things like this out?





My answer would be that it probably uses some other method, and my reasoning for this is how we learn things. If a quarterback throws to a receiver running the same route over and over, he will become more and more accurate. Assuming your brain's estimation of time and distance does not change, but rather just it's estimation (and your arm motion stays consistent in accuracy...any difference is related to brain estimation) of 'where to throw' changes, I do not know of a way this change could be accurately modeled...that is, if it actually used the laws of physics, it would come up with the exact same value every time...but hey, what do I know! THE question is, what do you think?

 

[expscv]

your brain is learned by experience, using experience to compare the situation . and give an estimated result.

more accurace answer usually required a maths&physics method...a pen and a paper

 

[hypnagogue]

expscv is essentially correct. The brain learns by trial and error, using experiences of motion to arrive at internal models of how motion works in the external world and how motion works with one's own body.

These internal models almost certainly do not involve anything like calculations of Newton's laws. The computational costs of having to do such calculations all the time would be prohibitive even with the brain's massive computing power. Another indication of this comes from AI, where even the simplest of motoric tasks (like picking up a cup) for a robot is inordinately difficult to implement using Newtonion laws as the basis.

 

[confutatis]

There are two simple, undisputable facts about this:

1 - anyone who can play tennis can also play baseball; this implies there are things in common between those two games
2 - being a good tennis player does not make one a good baseball player; this implies not everything is common between those two games

What do tennis and baseball have in common? The laws of physics. How do they differ? Basically on a very important point: the movements the player must make to catch the ball.

I think it's reasonable to say that we do calculate, in a strictly mathematical way, the trajectory of objects. I believe the fact that some autistic people can perform very complex calculations is evidence that we do have extremely powerful mathematical skills, even if we have trouble consciously accessing those skills.

The behaviour of our bodies, however, is a completely different story. We have no way to calculate how fast or how far our limbs will move, because we have no way to measure whatever it is that causes our limbs to move. We have the mathematical formula, but we don't know the values of some of the variables involved. And when that happens, there is only one alternative: we must use probabilities. We must guess what those variables could be, do the calculations, perform experiments, and verify how close we guessed. That process is known as "practice".

 

[hypnagogue]

What do tennis and baseball have in common? The laws of physics. How do they differ? Basically on a very important point: the movements the player must make to catch the ball.

I think it's reasonable to say that we do calculate, in a strictly mathematical way, the trajectory of objects. I believe the fact that some autistic people can perform very complex calculations is evidence that we do have extremely powerful mathematical skills, even if we have trouble consciously accessing those skills.

Clearly the activities of the brain can be cast in a purely computational way, and clearly when we react predictively to a body's motion we are in some sense calculating its trajectory. However, I took KingNothing's question to be inquiring a stronger claim: do our brains actually use Newtonian laws to accomplish this feat, or do they use some other computational means? I think the answer is clearly the latter. I highly doubt that somewhere in my brain a free body diagram is constructed to assess the forces acting on an object, and then its path is predicted based on application of F=ma and various velocity/position relations. The actual process is probably much more heuristic than that, and probably depends critically on continual, 'online' visual perception of the object.

The behaviour of our bodies, however, is a completely different story. We have no way to calculate how fast or how far our limbs will move, because we have no way to measure whatever it is that causes our limbs to move. We have the mathematical formula, but we don't know the values of some of the variables involved. And when that happens, there is only one alternative: we must use probabilities. We must guess what those variables could be, do the calculations, perform experiments, and verify how close we guessed. That process is known as "practice".

I imagine that such a process of 'practice' is also involved in creating our predictive models of the external world as well. An infant must learn how balls thrown into the air rise and fall just as much as it must learn how to walk.

 

[Njorl]

Most of what is involved in catching the ball, or hitting it in the case of tennis or baseball is pattern recognition. We don't calculate trajectories, we compare them to similar ones. Almost everything that is made easier by practice (neurologically speaking) is pattern recognition. Of course, practice also builds muscle and improves endurance, but many a great athlete can't hit a curve ball.

One thing neural networks do much, much better than serial calculators is pattern recognition.

Njorl

 

[confutatis]

Clearly the activities of the brain can be cast in a purely computational way, and clearly when we react predictively to a body's motion we are in some sense calculating its trajectory. However, I took KingNothing's question to be inquiring a stronger claim: do our brains actually use Newtonian laws to accomplish this feat, or do they use some other computational means?

I didn't get that from his question. If that is the case, then the answer is obviously no, as we have no intuitive concept of mass, force, acceleration, gravity. If we had, it wouldn't have taken thousands of years for Newton's laws to be invented.

I highly doubt that somewhere in my brain a free body diagram is constructed to assess the forces acting on an object, and then its path is predicted based on application of F=ma and various velocity/position relations.

Of course not, because even after four centuries of classical mechanics, we still think of the mechanical concept of force as being different from our sensations associated with muscle activity. All I said was that some mathematical model of the ball's behaviour probably exists in our minds, but I seriously doubt it resembles Newtonian mechanics.

The actual process is probably much more heuristic than that, and probably depends critically on continual, 'online' visual perception of the object.

The continual, 'online' visual perception is only required to correct errors in measurement. I often astonish myself, when driving, by thinking how precisely I can predict where an oncoming car will be at any point in time simply by taking a quick look at it. Yet, that astonishing skill is pretty much useless if the other car is accelerating, which explains a lot of stupid accidents out there.

I imagine that such a process of 'practice' is also involved in creating our predictive models of the external world as well. An infant must learn how balls thrown into the air rise and fall just as much as it must learn how to walk.

Surely. It's not only a matter of adjusting the variables, but also of postulating different models and assessing their relative merits. We are all scientists, and quite good at it.

 

[selfAdjoint]

But there are evidences of some forms of computation. For example recent experiments detected a short term memory limit of four or fewer "identified things". This is very much like a limit on register size in a CPU chip.

 

[hypnagogue]

I didn't get that from his question. If that is the case, then the answer is obviously no, as we have no intuitive concept of mass, force, acceleration, gravity. If we had, it wouldn't have taken thousands of years for Newton's laws to be invented.

We essentially agree, but this particular point doesn't follow. It could be possible in principle that our brains do have such internal models, but that these models are not privvy to conscious access. For instance, I have no intuitive concept of the neural algorithm that regulates my breathing, which presumably has some kind of computational model of the level of oxygen in my blood. Perhaps more relevant, I may have no conscious intuition of the subtler points of body language, even though people are regularly shown to pick up on such signals on an unconscious level. There is much that our brains do without our knowing it, and presumably this brain activity comprises meaningful (modeling) calculations.

 

[Imparcticle]

Here is the question before you today, ladies and gentlemen: Do our brains calculate, or do they measure with their own methods?

What is the difference [that you appear to see, and I don't] between "calculate" and "measure with their own methods"? Calculation is methodical itself, that was concieved by human minds. Therefore, it is the same thing, as far as I can see.

I posed myself this question when thinking about subconscious estimation...like when you catch a baseball, you don't know exactly how fast its going, but in under a second your brain calculates about how fast its going, so you can catch it.

Your brain doesn't calculate how fast something, a ball, is going so you can catch it (at least not neccesarily); rather, as others have said, this phenomenon is a product of previous experiences.

 

[expscv]

;) if we do have complex calulation in our internal system, do we still need a lot of pratice for a tennis game? we r human not machine, means that we are more felxiable

cause we learned from experience instead of apply mathmatical physics laws

we do things more beyond calulation and strict rules, it makes us improve

unlike calulation and laws they are constant they will not evolve


i think that our brain is like a data warehouse with a powerful! serach engine! and some process technic e.g "if then" "take average"

 

[selfAdjoint]

unlike calulation and laws they are constant they will not evolve

On the contrary, there is a whole technique of computer software development called genetic programming, that is based precisely on making computer calculations evolve.

 

[Jeebus]

… Perhaps more relevant, I may have no conscious intuition of the subtler points of body language, even though people are regularly shown to pick up on such signals on an unconscious level. There is much that our brains do without our knowing it, and presumably this brain activity comprises meaningful (modeling) calculations.

Really? I think the brain directly indicates what we do by our actions. If an action occurs, then the brain associates language by neural transmissions to 'mimic' our body language or correspond to get the 'point' across. The brains knows what we are doing, without 'us' as a holistic mechanism knowing what we are acting upon or thinking for. The brains knows, but its another system operating the "control panels" of our body language. It's like knowing what 1+1 is without really thinking about it, even though the brains knows, our external system calculates it like a second-nature operation.

 

[Imparcticle]

SelfAdjoint:
That is intriguing. Is that from this month's issue in SCIAM (I thought I saw something like that in there...). Anyway, how does it evolve?

 

[expscv]

unlike calulation and laws they are constant they will not evolve

On the contrary, there is a whole technique of computer software development called genetic programming, that is based precisely on making computer calculations evolve.

oh yeah, should mention that haha thx

how long do you think this neural system would be good enough. able to function like a human brain in a particluar area say music.