How are we aware of the location of touch on our bodies?

[Unity+]

I have been thinking of artificial intelligence and trying to find my own way around the research. I want to first get associated with feeling and such and see where that leads me.

 

So, it is obvious that we are capable of feeling touch. We are able to touch and the particular part of the body touched feels that touch. However, in many cases when a person loses an arm they can still feel itching in that particular area, which implies that the brain has much responsibility over feeling rather than the nerves itself, though the nerves themselves play a big role in it.

 

My question is how is the brain capable of making us aware of the location of the touch? I know it deals with nerves and the messages sent from the brain to that particular part of the body, but how can we spatially feel the location of the touch? Sorry if this seems confusing.

Edited June 9, 2014 by Unity+

[Roamer]

So, it is obvious that we are capable of feeling touch. We are able to touch and the particular part of the body touched feels that touch. However, in many cases when a person loses an arm they can still feel itching in that particular area, which implies that the brain has much responsibility over feeling rather than the nerves itself, though the nerves themselves play a big role in it.

 

 

? Are you confusing an itch with a tendency(brain function) to use the arm ?

 

 

My question is how is the brain capable of making us aware of the location of the touch? I know it deals with nerves and the messages sent from the brain to that particular part of the body, but how can we spatially feel the location of the touch? Sorry if this seems confusing.

 

 

I'm not familiar with the, umm, genetic basis for this, but since we learn from the day we're born, such basis could be small to non-existant.

 

Suppose a nerve of a cut-off arm still sends signals(an itch), then doctors attach a small device to it to stop the itch, but it has to be activated, and they put an activation-sensor in the belly, so you just need to rub your belly to get rid of the itch.

Very soon the subject would start "imagining" the itch comes from his belly and he'll scratch his belly to stop it.

[Unity+]

 

Very soon the subject would start "imagining" the itch comes from his belly and he'll scratch his belly to stop it.

I am referring to the subject being able to "imagine" a symptom and feel pain or such. How can the mind do this?

[Roamer]

I once had the opposite happening to me;

i spilled (close to) boiling water on my belly, but, it first had to move through a few levels of clothing,

by the time it reached my skin i was already dealing with taking those clothes off and getting some cold water,

hence i did not feel any pain(i suppose i didnt need the warning) even though a chunk of my skin was burned.

 

I have no doubt there are people on this planet who, in same situation, would start imagining pain before the boiled water reached the skin,

which might actually make sure they scream/groin/autsj and thus get immediate help from other people present(i was alone)

[CPG]

That's a really complex topic and an interesting question. The conceptual part that seems mot difficult to me is the brain "mapping" the entire body and the space around it. If you first look at how such body mapping is thought to occur, localization intuitively seems like a straightforward process. There is a book by Sandra Blakeslee and Matthew Blakeslee that deals with this topic really well. Even shorter version than the book, Ginger Campbell's "Brain Science Podcast" has an episode or two about that book and interviewing Sandra Blakeslee. Those podcasts tend to be an hour long or so. Depends how deeply you want to get into the topic, but those are my recommendations

Also, check out the wiki article on "referred pain". Similar to the phantom limb, I think examples of this process gone wrong that are informative about how it works.

 

http://en.wikipedia.org/wiki/Referred_pain

[metacogitans]

 

My question is how is the brain capable of making us aware of the location of the touch? I know it deals with nerves and the messages sent from the brain to that particular part of the body, but how can we spatially feel the location of the touch? Sorry if this seems confusing.

The brains mechanism for mapping nerves responsible for the sensation of touch throughout the body is, like many other mechanisms and processes in the brain, almost completely subconscious. You'll never have to think about it. Although the shape of your body may change as you gain or lose muscle mass, body fat, or as bone growth makes you taller, nerve mapping and the sensation of touch is one of the oldest and most instinctively rooted jobs the brain performs. The constant input from those nerves to the CNS allows for it to be intuitive and subconscious, and re-mapping of never location is done constantly, efficiently, and instantaneously.

Edited June 29, 2014 by metacogitans

[iNow]

The simple answer is signal interpretation. The impulse travels through very specific channels and nerve bundles and pathways arrives to the brain with a specific "signature." Those pathways differ slightly based on where in the body the stimulus occurred. It is through time that we learn to differentiate those signals and become able to associate the signal with a specific stimulus and specific location.

[metacogitans]

 

 

The simple answer is signal interpretation. The impulse travels through very specific channels and nerve bundles and pathways arrives to the brain with a specific "signature." Those pathways differ slightly based on where in the body the stimulus occurred. It is through time that we learn to differentiate those signals and become able to associate the signal with a specific stimulus and specific location.

Where did you get the idea of a "signature"? I'm pretty sure that the "all or nothing" nature of action potentials implies that either a signal is sent or it isn't - it's either firing along the axon or the stimulus wasn't strong enough to trigger an action potential. The signal carries no other information or 'signature'. Now, there are other things to consider, like how frequently repeated signals are sent, levels of neurotransmitters already present in the synapse, etc. You also have to consider every other neuron connected to the process.

Also, when you say 'through time' it is actually almost instantaneous, and when you say "we learn", the stimulus is actually processed subconsciously; it is not as if we need to consciously guess, test, and revise to try and figure out which sensations correspond to which locations on our body; it's all done without even thinking about it. Think of If you felt a bug crawling on your shoulder - without even looking you'll likely swing your hand from your other arm over to brush it off with a high level of precision and no visual feedback required. For different types of stimulus (feeling heat versus feeling something pushing up against you, for example), these have different nerves with dedicated roles for detecting particular types of stimuli. A neuron which detects heat would not also detect a sharp object poking the skin.

 

 

Edited June 29, 2014 by metacogitans

[iNow]

Where did you get the idea of a "signature"?

I simply used that word as a rhetorical shorthand to more simply convey the essence of the idea to someone who is intelligent, but quite clearly still new to the underlying mechanisms of the human nervous system.

 

Simplified, part of interpreting the signal is done by incorporating information about which pathways and nerve bundles through which that signal has travelled to arrive at its destination. My point was that that extra information is critical to accurately understanding both the source and nature of the stimulus. As long as we're on the subject, it should also be noted that these pathways are directly relevant to which parts of the brain ultimately receive that stimulus and become responsible for its interpretation (see also: homonculus, in context of body mapping and motor cortex representation).

 

I'm pretty sure that the "all or nothing" nature of action potentials implies that either a signal is sent or it isn't - it's either firing along the axon or the stimulus wasn't strong enough to trigger an action potential.

Sure, I don't disagree, but that changes nothing whatsoever about the actual point I was making.

 

The signal carries no other information or 'signature'.

Perhaps not the signal itself, but the path traversed by the signal, as well as its intensity and any parallel or neighboring signals very much do provide additional information that is later used to help interpret and understand the origin of the sensation. This was my core point that I think you may have originally missed.

 

Also, when you say 'through time' it is actually almost instantaneous

I'm sorry, but you're confused. I was referring to the concept of learning (again, as a rhetorical shorthand to describe the process of associating neural signals to specific stimuli and localizing where on the body that stimuli may have initiated the neural cascade). It is through time that we learn and become increasingly conscious of those signals, where they originated, and what generated them.

 

Also, while neural signals propagate rather quickly, the action potential should never be described as "almost instantaneous." It doesn't even happen as rapidly as electricity flowing through a copper wire, which (as fast as it is relative to our average everyday human velocities) is also not "almost instantaneous." Relatively speaking, the cascade of an action potential is a pretty slow process that is dependent on a number of other important variables (sodium/potassium concentration, density of dendritic connections, axon length, health and thickness of the myelin sheathing, distance between source and brain, etc.).

 

Think of If you felt a bug crawling on your shoulder - without even looking you'll likely swing your hand from your other arm over to brush it off with a high level of precision and no visual feedback required.

Yes, and despite what you may think you had to learn to do this. You were not born with this skill, but it was instead acquired... It was learned.

 

From the moment your nerves began sending impulses in the womb, through your infancy and adolescence, all the way through your current life today you have been learning and further refining your interpretation of the myriad neural signals cascading through your body in each passing moment, plastically pruning and reinforcing various neural connections.

 

On another note, I encourage you to reconsider your continued usage of the terms "conscious" and "unconscious" as they are largely a meaningless carryover from the time of Freud and they fail to adequately capture the nuance of modern neurobiology or offer even a minimally relevant explanatory framework for these issues. They tend to create more confusion than they ameliorate.

 

For different types of stimulus (feeling heat versus feeling something pushing up against you, for example), these have different nerves with dedicated roles for detecting particular types of stimuli. A neuron which detects heat would not also detect a sharp object poking the skin.

It's not quite that simple and we don't have different nerve receptors for every single different type of stimulus (even though I stipulate that we do have some specialized receptors for things like temperature).

 

You have to remember that the same nerve receptors responding to a sharp knife also respond to a dull one, and yet we can differentiate between the two. We can tell the difference between a spoon touching us and a fork or between a wet towel and a dry one or between sandpaper and cashmere or velvet and we can do this because of the nature of the signal being sent to and received by the brain... The signal has a certain "signature" when reaching the brain since different networks were activated to bring it there and we have learned through time what the different signals and different patterns of network activation tend to mean. For comparison, it's almost as if we perform a type of Fourier Analysis on the many incoming waves of information.

 

Now look... I appreciate that you are trying to help ensure accurate understanding of these complex issues and that your intentions are good, but I ask that you please validate your own comprehension of what others are posting prior to jumping in and responding so critically and aggressively yourself. I always welcome clarification and correction when appropriate, but ask that the feedback at least be based on a clear understanding of my actual message and not some misreading of my words. Cheers.

Edited June 29, 2014 by iNow

[CPG]

Action potentials can carry far more information than the binary yes/no implied above. Even if we are only considering all-or-none events at a single synapse, you'd also always be dealing with the precise timing and rate at which they occur. Firing at different frequencies (2 Hz, 50 Hz, etc) can have entirely different effects on the post-synaptic neuron, and can make the difference in how the signal is processed (ie filtered out or amplified). The synaptic as well as intrinsic properties of the neurons which are sending or receiving a signal can display complex temporal dynamics that are fundamental to neuronal signal processing. The paper linked below makes for a nice theoretical introduction with its graphical depictions which in my opinion really go a long way towards assisting an understanding of how neurons can have such "frequency preference" characteristics.

 

 

Bursts as a unit of neural information: selective communication via resonance (Izhikevich et al. 2003)

http://www.ncbi.nlm.nih.gov/pubmed/12591219

[Genecks]

God did it....

 

No, but really, I don't know. Fourier transformations?

Edited July 15, 2014 by Genecks

[Endy0816]

Generally hear the process described as mapping signals onto a mental simulacrum.

 

Hence you can lose a limb but still feel an "itch". Thankfully there are fairly unsophisticated ways to trick it.

 

 

and in only slightly more professional settings these methods are used to help with phantom limb pain.

 

 

The Pinocchio illusion is similar. Essentially your brain concludes that your nose is extending. Stupidly presumptive system. Poor input validation.

Edited July 15, 2014 by Endy0816

[CPG]

God did it....

 

No, but really, I don't know. Fourier transformations?

 

 

Fourier transformations are strange to look at at first, I'll admit. If you want to calculate the frequency of an event occuring, the straightforward measurement would be to count how many times the event occurs and divide it by the total amount of time, right? However, in this situation, what they want to do is look at different frequency inputs a neuron might receive (X axis), and how effective that frequency is at changing the membrane potential (Y - axis). The goal to describe the neuron's response to different frequencies of input. With frequency as our independent variable and impedance as the dependent variable you can now see that time is removed from consideration - and yet we are still dealing with frequency... what the heck?

 

The Fourier transform is a way to take any signal (let's say it's a recording of current injected into a cell) and break it down by how much of that signal is made up of a given frequency. Let's say hypothetically you had a perfect sine wave with a nice steady 1Hz frequency. This means it goes through one full cycle - peak to peak - exactly once per second. The most abundant frequency present in a signal will turn out to be the peak on a Fourier transform so your ideal sine wave would show a peak at 1Hz and nowhere else. In the process of doing that, time actually gets dropped out so it's a mathematical way to look at frequency without consideration of the time domain. But where does it come from? There are many different ways to actually do the math but I just want to stimulate some intuition for it here.

 

If you want to take a signal and calculate frequency, sticking with our sine wave, we can take the 2nd derivative at any given point. The rate of change of the slope gives you an instantaneous frequency. Do this for the whole signal (ie the whole current injection) and tally up how much you get for each frequency. That is the "power" for each frequency. If you are injecting current you will get a voltage response from the cell and do the same transformation. This can give you a profile of frequency-specific impedance. Here's how and why that matters:

 

 

Tying it together with my previous post:

a) Ohm's law tells us that Voltage/Current = Resistance

b) Take the Fourier Transform of Voltage, divide that by the Fourier Transform of the current the cell receives, and that gives you a curve that describes the frequency-specific resistance (Impedance) of the neuronal membrane. If there is a peak in impedance at a specific frequency, biological input at that frequency would be more effective at changing membrane potential than at other frequencies and thus more effectively drive that cell.

 

Tying it together with the rest of the thread:

a) I originally posted that link because it seemed to someone that neuronal output can't carry a "signature" or the richness of information that they do.

b) Neurons are capable of firing at different frequencies, so it's not unfair to call a particular output frequency a "signature" that helps describe a specific stimulus. I have heard pain researchers say that other types of sensory fibers (ie carrying information from touch or temperature or something else) can also be interpreted as a painful stimulus by the brain if it comes at a particular frequency - even though it's coming from nerves that usually provides a very different sort of feedback.

 

I would have to do some searching to find citations in the literature documenting this exact phenomenon with respect to pain, but I have been told that this occurs. True, that way of answering leaves some evidence to be desired, but having said all this I'm curious and intend to look around the literature when I get the chance and see what I can find out. I don't want to forget about posting, so if anyone is curious or skeptical, let me know and I'll be happy to reply after I read a few papers. Just remind me so I don't forget!

[barfbag]

It occurred to me that research regarding strokes might be of benefit to this topic as I believe many stroke victims lose understanding of their sensations and must re-learn things like where to swat if you feel a bee.

 

It is only a possible area to search this idea further, but I have no interest in researching this and honestly do not quite understand what stroke victims do or do not feel. It is just an avenue not mentioned here yet, and it seems logical. Methods used on stroke victims could be analyzed for use in AI.

 

Just a thought,

 

Cheers

[CharonY]

The main reasons for stroke is that they create lesions in the brain. If these are the areas that receive afferent information from a given part of your body, you will lose the ability to sense it. Re-learning is typically done by having other brain areas take over that function.

Obviously for sensing the actual pathway (i.e. fibers leading from transducing neurons to specific brain areas) is used in conjunction with the the integrated signal (as typically heavy processing is done before signals reach the brain).

[barfbag]

@ Charony,

 

I believe the OP is wanting to teach (in theory) a robot or AI to learn spatial touch as we would. I would think we are still centuries away from stuff like that, but who knows.

 

So it is the recovery therapies that might be of more interest than the causes or effects of the stroke if you know any.

 

I'll take 1 minute and google "Spatial touch stroke therapy" and see if I get any hits.

http://stroke.org.au/pdf/Cognitiveproblemsfollowingstroke.pdf

 

Any unique therapy that can help an adult re learn might be of interest to OP. I've read of some in past but cannot find them in short notice.

[iNow]

Try using the phrase Somatosensory Perception.

 

http://nba.uth.tmc.edu/neuroscience/m/s2/chapter02.html