The Binding Problem

[gib65]

I've heard of the binding problem in the neuroscience: how seperate mental functions carried out by separate brain parts come together to form perceptions of something "whole". They say there is no part of the brain that all the electric signals being processed converge to - that is, no part that seems to integrate all this information so that we can have perceptions of whole things. I was wondering how far reaching the bind problem is considered to be. I mean, for example, I know that it means to say that there is no central part of the brain where ALL information converges to for integration, but what about less wholistic things? For example, when I look at an apple, I see that it is round and red. I also know that the roundness and redness of the apple are two distict pieces of information that are processed by different parts of the brain. But haven't there been parts of the brain found to store information about whole objects such as apples - like the memory of what an apple is, or the concept of an apple, or the perception of things shaped like apples (things like "geons"). I've read about a man who mistook his wife for a hat, which means there must be a part of the brain responsible for perceiving or conceiving of whole objects and recognizing what they are (as hats or apples). Have studies not shown that the parts of the brain responsible for perceiving shapes like the roundness of the apple and colors like the redness of the apple might somehow have neural connections to the parts of the brain responsible for perceiving whole objects?

[gib65]

So no one has anything to say about the binding problem? What if I asked a simpler question: What is the binding problem? I mean, what does the binding problem state exactly, just so I'm not misunderstanding it.

[gib65]

So no one has anything to say about the binding problem? What if I asked a simpler question: What is the binding problem? I mean, what does the binding problem state exactly, just so I'm not misunderstanding it.

[Bio-Hazard]

It reminds me of the computer file system that windows uses, I can't remember which one but I'm thinking of FAT32.

 

Simply put, the system will look towards a directory of little pieces of data as to where the next piece of data that will lead to the whole program lies.

 

Computer science and neuroscience in my opinion relate to a very large degree, or atleast in concept, but I'ven't kept up with computer science in a long time. I believe in my last CHECK IT OUT! thread I gave an article that said some stuff about MIT looking into the process of perception with encoding for the brain.

 

p.s. that might be an English word if tis'nt already. (i'ven't..)

 

 

I understand the binding problem to a miniscule degree that it relates to the inability for people to figure out where and how the brain stores tons of information, yet all of it scattered like marbles and there is a process the brain uses to collect all the marbles at a fast rate.

 

I think my computer science idea is going to be the closest thing to describe my idea of what happens. Just speculation and assumption from previous knowledge obtained how information electrical systems at work.

 

Wouldn't it be nice if we had access to an MIT Forum that has to do with neuroscience?

[gib65]

I think that makes sense as far as software goes. In software programs, the IP (Instruction Pointer) only reads one instruction at a time, and it increments itself linearly after every instruction is executed - so, it reads and executes instruction 1, then instruction 2, then 3, and so on. Once in a while, it encounters an instruction that tells it to jump to another instruction (like instruction #3682) at which point it continues linearly.

 

This sounds much like the brain in that it will process information locally and then "jump" to an totally different area.

 

But unlike computers, the brain has no "central processing unit". For every instruction the computer executes, it must take that instruction, send it through the CPU, and then return to the point it was before with the results. This results in a sort of bottlenecking, or a one-instruction-at-a-time rule. The bind problem states that there is no such bottlenecking in the brain because there has not been any discovery of a neurological CPU. All processing happens sporadically around the brain in parallel, as though each neural center has its own local CPU.

 

What I was wondering is how specific does the binding problem get. For example, I know there is the dorsal and ventral pathways that both start from the visual center in the occipital lobe, and process visual information for the location and movement through space of a perceived object (the dorsal pathway) and what the object is (the ventral pathway). Now, if you placed a carrot to the left of me and an apple to the right and asked me to point to the carrot, I would be able to do this with ease. This means I would be using location information in order to point (dorsal pathway) as well as recognizing which object is the carrot (ventral pathway). Therefore, the dorsal and ventral pathways must be communicating at some point. Signals from either one must be converging somewhere.

 

Does the binding problem say that this doesn't happen?

[ashennell]

binding problem -

 

http://www.dartmouth.edu/~adinar/publications/binding.pdf

 

I dont think that there is any useful link between turing machine like procesing and the binding problem - at least not any link that I know of.