How do we model time and space cognitively?

[BWE]

I need some sources from current neuroscience that describe how we model time and space cognitively to plan or predict future events.
...
I am trying to figure out a GIS application using math that is a bit over my head and I needed some help in figuring out what the model might be shooting for.
...
How do we run a mental simulation to decide if an action is likely to provide the desired results?
http://www.ncgia.ucsb.edu/giscc/units/u006/u006.html
http://www.eurekalert.org/pub_releases/2008-01/w-dbs012408.php
http://www.mindmodulations.com/mindmods/images/PDF/16.blanke%20o,%20arzy%20s.%20(2005)%20out-of-body%20experience,%20self,%20and%20the%20temporopa rietal%20junction.%20neuroscientist%2011%2016-24.pdf
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V91-4378TFF-W&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=20aa60b61c2dbf44fc19e0a03e6de2a6
http://www.journals.uchicago.edu/doi/abs/10.1086/522359

I've been through over a dozen, maybe a score of papers and it's too specific to chase down without knowing what words I should be looking for. These ones are interesting for anyone who wants to look.

[Don Alhambra]

My work focuses on predictive motor control, and there's been a hell of a lot of interest in recent years on how the brain uses internal forward models to predict the sensory consequences of actions. There's a good primer here:

http://learning.eng.cam.ac.uk/wolpert/publications/papers/WolFla01.pdf

and some reviews:

http://learning.eng.cam.ac.uk/zoubin/papers/natneuro.pdf
http://www.iop.org/EJ/article/1741-2552/2/3/S11/jne5_3_s11.pdf
http://www.springerlink.com/content/2253374773600417/fulltext.pdf

Don't know if you're looking specifically for motor control, but that's the best I can come up with. :)

[BWE]

My work focuses on predictive motor control, and there's been a hell of a lot of interest in recent years on how the brain uses internal forward models to predict the sensory consequences of actions. There's a good primer here:

http://learning.eng.cam.ac.uk/wolpert/publications/papers/WolFla01.pdf

and some reviews:

http://learning.eng.cam.ac.uk/zoubin/papers/natneuro.pdf
http://www.iop.org/EJ/article/1741-2552/2/3/S11/jne5_3_s11.pdf
http://www.springerlink.com/content/2253374773600417/fulltext.pdf

Don't know if you're looking specifically for motor control, but that's the best I can come up with. :)
Sonofabitch. Just sonofabitch. Thank you.

The biggest impact of science on the 21st century will be computer modeling.

Thanks again.

[Don Alhambra]

Any time. :)

[Barbarian]

http://www.iop.org/EJ/article/1741-2552/2/3/S11/jne5_3_s11.pdf
http://www.springerlink.com/content/2253374773600417/fulltext.pdfVerily, these are the paths leading into confusion, with no substance behind them whatsoever, meant to mislead the righteous.

IOW linky no worky. :(

[Febble]

I need some sources from current neuroscience that describe how we model time and space cognitively to plan or predict future events.
...
I am trying to figure out a GIS application using math that is a bit over my head and I needed some help in figuring out what the model might be shooting for.
...
How do we run a mental simulation to decide if an action is likely to provide the desired results?
http://www.ncgia.ucsb.edu/giscc/units/u006/u006.html
http://www.eurekalert.org/pub_releases/2008-01/w-dbs012408.php
http://www.mindmodulations.com/mindmods/images/PDF/16.blanke%20o,%20arzy%20s.%20(2005)%20out-of-body%20experience,%20self,%20and%20the%20temporopa rietal%20junction.%20neuroscientist%2011%2016-24.pdf
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V91-4378TFF-W&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=20aa60b61c2dbf44fc19e0a03e6de2a6
http://www.journals.uchicago.edu/doi/abs/10.1086/522359

I've been through over a dozen, maybe a score of papers and it's too specific to chase down without knowing what words I should be looking for. These ones are interesting for anyone who wants to look.

Can you be more specific about the problem?

[BWE]

I need some sources from current neuroscience that describe how we model time and space cognitively to plan or predict future events.
...
I am trying to figure out a GIS application using math that is a bit over my head and I needed some help in figuring out what the model might be shooting for.
...
How do we run a mental simulation to decide if an action is likely to provide the desired results?
http://www.ncgia.ucsb.edu/giscc/units/u006/u006.html
http://www.eurekalert.org/pub_releases/2008-01/w-dbs012408.php
http://www.mindmodulations.com/mindmods/images/PDF/16.blanke%20o,%20arzy%20s.%20(2005)%20out-of-body%20experience,%20self,%20and%20the%20temporopa rietal%20junction.%20neuroscientist%2011%2016-24.pdf
http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6V91-4378TFF-W&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=20aa60b61c2dbf44fc19e0a03e6de2a6
http://www.journals.uchicago.edu/doi/abs/10.1086/522359

I've been through over a dozen, maybe a score of papers and it's too specific to chase down without knowing what words I should be looking for. These ones are interesting for anyone who wants to look.

Can you be more specific about the problem?

Yes. In most complex GIS 3-d/4-d modeling, the more variables you can run, the more accurate the results. Think climate modeling or disney 3d animation with mondo ray tracing and physics. That gets pretty processor intensive pretty fast. But our brain approximates several features of observed reality and seems to know which features are good candidates for the lower processing areas and which features need to be thought out more carefully. I figured out a feedback loop that recognizes variables that aren't affecting the important isolated pattern very much and made a function that begins reducing the number of calculations done on the less important features immediately. I'm adding in another function to monitor unusual changes that the lower processing might accidentally cause.

I think I can model specific effects to populations, including human neighborhood demographics BTW, across very broad areas, possibly worldwide, brought about by very subtle environmental or economic factors (which I'm treating as environmental. I talked it over with a guy who deals with economics and he said it's good enough considering that isn't what it's designed for. So much human data is already available through arcsoft that it isn't too much of a stretch. Anyway, that's only an unintended bonus.).

My inspiration was the human brain and I wondered where it limits itself on purpose and what limits are imposed by its processing power. i.e. for larger changes it is the same level of processing over vast time spans as it is for very complicated changes over short time spans.

It's cooler than hell and seems to work pretty well with tweaks being relatively easy. It's also infernally complicated so I'd like to have a better idea of how people will naturally relate to it as it runs and I am modeling it after the way I imagined we do it so I also want to know how we do it in case an obvious improvement becomes apparent.

Look ma no punctuation



!

Dangit.

[Febble]

Well, the current models of human perception and action tend to be feed-forward models (Don Alhambra will happily provide references from the top of his head, I'd have to look them up) rather than feedback models. The idea is that at time we have a representation of what the world will look like on completion of the action, then after a small time increment, a current model generated from feedback is compared with the old forward model, and the forward model tweaked to take into account the new current model. And so on, iteratively.

It's getting late, so that's probably as clear as mud. Don's field is motor control so his stuff is probably more relevant, but I did some work on eye movements and perception, and developed a model again based on feed-forward, that accounts for the fact that when you move your eyes, the image moves across the retina, but the world appears to stay still (whereas, if you manually poke your eyeball, the world appears to move) - which is one strong piece of evidence for predictive processes (i.e. forward models) rather than sole feedback models.

Gotta get my broken foot elevated - I'll be back in the morning, probably!

[BWE]

Also, I have the equivalent of slow sensory input built in, that's what it uses to check the accuracy of its neglegted elements. That's why I was so jazzed about those other papers from Don. Essentially, I modified a standard arcview output through autocad so I can graphically represent the output as a movie if I want to. It's a kind of weird movie because the environment morphs a lot from lack of adat but I have that part figured out through simple parralel processing of the origninal map with the data superimposed from a separate processor.

[BWE]

Well, the current models of human perception and action tend to be feed-forward models (Don Alhambra will happily provide references from the top of his head, I'd have to look them up) rather than feedback models. The idea is that at time we have a representation of what the world will look like on completion of the action, then after a small time increment, a current model generated from feedback is compared with the old forward model, and the forward model tweaked to take into account the new current model. And so on, iteratively.

It's getting late, so that's probably as clear as mud. Don's field is motor control so his stuff is probably more relevant, but I did some work on eye movements and perception, and developed a model again based on feed-forward, that accounts for the fact that when you move your eyes, the image moves across the retina, but the world appears to stay still (whereas, if you manually poke your eyeball, the world appears to move) - which is one strong piece of evidence for predictive processes (i.e. forward models) rather than sole feedback models.

Gotta get my broken foot elevated - I'll be back in the morning, probably!
Oh ouch! It didn't have anything to do with kicking a live giant magnet wearing steel toed boots after it quit feeding you data in the middle of a test did it?

Anyway, this is quite explicitly a feed forward model. That's the only way to do it. Otherwise you have to process everything. I'm able to run a model that would have taken 5 or 6 hours on a newer dual core pc running xp pro in about 5 minutes with what appears to be identical critical accuracy.

[Don Alhambra]

http://www.iop.org/EJ/article/1741-2552/2/3/S11/jne5_3_s11.pdf
http://www.springerlink.com/content/2253374773600417/fulltext.pdfVerily, these are the paths leading into confusion, with no substance behind them whatsoever, meant to mislead the righteous.

IOW linky no worky. :(

Sorry dude - I know some of them are behind paywalls.

If anyone's interested, I absolutely cannot possibly break copyright and send you copies. Just PM me...

[Barbarian]

Sorry dude - I know some of them are behind paywalls.I swear they were plainly unavailable when I last tried. Now they seem to be up, one asking for money, the other asking for patience until site maintenance is completed. Internets haet me.

[RAFH]

My work focuses on predictive motor control, and there's been a hell of a lot of interest in recent years on how the brain uses internal forward models to predict the sensory consequences of actions. There's a good primer here:

http://learning.eng.cam.ac.uk/wolpert/publications/papers/WolFla01.pdf

and some reviews:

http://learning.eng.cam.ac.uk/zoubin/papers/natneuro.pdf
http://www.iop.org/EJ/article/1741-2552/2/3/S11/jne5_3_s11.pdf
http://www.springerlink.com/content/2253374773600417/fulltext.pdf

Don't know if you're looking specifically for motor control, but that's the best I can come up with. :)
Sonofabitch. Just sonofabitch. Thank you.

The biggest impact of science on the 21st century will be computer modeling.

Thanks again.

Perhaps the biggest impact on humanity ever will be computer modeling.

Ever read Alan Kaye's "Are we there yet?"

http://www.openp2p.com/pub/a/p2p/2003/04/03/alan_kay.html
http://craphound.com/kayetcon2003
http://www.dmoz.org/Computers/History/Pioneers/Kay,_Alan/