Transcript for:
Non-Associative Learning and Perceptual Learning

so now that we've learned about different types of non-associative learning habituation sensitization and perceptual learning priming is an example of perceptual learning those discrimination tasks are examples of perceptual learning let's zoom in a little bit and see how what we know anyway about how this type of learning is supported in the brain we're in the nervous system it happens how some early efforts to figure this out were produced mixed results Pavlov being the monster that he was cut out a dog's cortex it observed that habituation ceased what part of the cortex we don't know things weren't that precise that those days Katz show habituation however even below a spinal cord cut meaning if you disconnect the spinal cord from the brain then below that they still see evidence of a situation suggesting it doesn't require or involve the brain at all in some cases roaches show a magician and they don't really even have any cortex to speak of so not entirely clear from these rather crude ways of probing that question and part of the reason it's difficult is that there are so many neurons and they interact in such complex ways it's difficult to answer questions about how neurons do X or Y so one common approach is to study very relatively simple nervous systems such as in please iya which is a si hair basically a sea snail kind of slug kind of thing which is oft studied because it only has only about 20,000 neurons in its central nervous system which is a lot but a lot less than the one 100 billion that we have and also the neurons are very large so they're relatively easy to study so it's didn't studied a lot particularly in the area of things like habituation and sensitization and stuff like that because we can actually watch what's happening in the neural network and figure out how habituation happens at a neural level so essentially what is happening in a habitual study using Aplysia here is we might touch the siphon which is I don't know some to be part of a snail with a brush or to stimulate it somehow and it basically with Drac with drive withdraws or sucks in its Gill and this is I don't know some defensive response so that when a prayers around it doesn't have its tender bits out or something like that anyway we can actually map out what's going on here when you touch the siphon it stimulates the sensory neurons priceless which release glutamate on to motor neurons glutamate is the primary neuron neurotransmitter that is released on the muscles to cause them to flex and that causes the motor neuron to activate the gill muscle and tada the gill muscle is sucked in so like I said the nervous system of these animals is relatively simple so we can actually kind of map out what's going on there's a beautiful animation if you can see that so that's a story touch activate a sensory neuron it releases glutamate onto motor neurons which cause the contraction of gill muscles and know what you do is you measure how long the snail waits so to speak before it relaxes its Gill again and here's what happens so again just to be clear on the process you touch the siphon it withdraws or retracts its Gill and then if you do that every minute for 10 to 15 minutes and each time you measure how long it takes for it to release or extrude its Gill again the amount of time it waits gets shorter and shorter so basically the first time you touch you that sucks that thing up and hold it for a while it eventually releases it then the second time you touch it it sucks it up and holds it but for not as long and so on and so forth and as you can see here on the bottom axis you've got trials 1 through 12 and how long it waits before it releases its Gill back out at first it waits a long time two and a half I guess seconds yes but then after you've done it every minute for nine minutes in a row for example it only holds it in for half a second so one fifth that amount of time and this is an example of habituation basically right the response to a stimulus is weakening over repeated administrations of that stimulus typically this habituation recovers very quickly so it is weighted and didn't touch it for 20 minutes or a day and then touched it again it would go right back to the long Gill withdrawal response much like it exhibited before you did any habituation but if you do this every day for a month for 10 minutes or something like that then you get or lasting habituation so it can last longer if you do a bunch of training spread out over time but if you only do it once it recovers very quickly how so we can look at how this happens zoom in a little bit look at how this happens in the neural network using this rough schematic of the nervous system of a sea snail or slug or whatever so you touch this iPhone activates the sensory neuron releases glutamate onto the motor neuron you get the picture we went over that so what is happening in habituation well one simple mechanism by which habituation is occurring in this neural network is that over repeated activations of that sensory neuron it just sort of runs out of glutamate it has less glutamate storage saved in its little vesicle water balloons there and so each subsequent stimulus results in the release of less glutamate tada so you have a less strong motor neuron response and the gill is released more quickly if you keep doing this we can actually see a mechanism of long-term habituation because there's less glutamate being released generally the nervous system has a kind of use-it-or-lose-it philosophy you might say and since there's less glutamate in the synapse their nervous system will actually prune away glutamate receptors so it would reduce the number of available receptors resulting in long-term habituation so even when the glutamate stores have had a chance to be replenished if you touch the siphon you're still gonna get a weaker response that is a situation because we've actually have these long-term structural changes where you've got glutamate receptors that aren't there anymore so you can also study sensitization in a please yeah again gentle touch the siphon produces a Gil withdrawal now we administer a small mild aversive shock to the tail and the next time we touch the siphon the Gil withdrawal is longer than it was at baseline so of course that makes sense as you remember habituation and sensitization kind of are two sides of the same coin causing opposite effects again sensitization recovers quickly but can become long lasting with multiple sessions and if we look at the mechanism for this at the neural level using this little they're all structural schematic here it turns out that there are more neurons than were John before there are inter neurons so when you touch the siphon you get the response ba-ba-ba-ba-ba but then if you shock the tail it actually activates inter neurons indicated by those blue lines there which release serotonin onto the sensory neurons which modulate their activity and causing them basically to release more acetylcholine on the next activation so it has to be an aversive strong dish stimulus in order to activate these internal inter neurons otherwise they're quiet so only when you get a strong a versiv stimulus you get these inter neurons inter neurons kick in and actually have the opposite effect that habituation has they cause a stronger response to a siphon touch on the next trial and similar to the neural mechanisms by which habituation work how does this work long-term well you get actually the addition of acetylcholine receptors like I said the same by the same token if there are receptors that aren't being used that much they'll tend to get pruned out if there are receptors that are being used a bunch there will be even more made so if you make more receptors that's gonna generally make the synapse more sensitive and you'll generally see a stronger response that is long term sensitization this slide just sort of sums up what we learn looking at it please iya they're habituation and sensitization caused opposing changes in synapse function one causes depression and one causes facilitation or activate I'm sorry reduces activation or increases it respectively causing weakening versus strengthening of reflex circuits and you got a very cursory glimpse learning can involve direct changes in sense function but also modulation the involvement of modulatory inter neurons that are only active in some cases and serve to increase or decrease neural activity at certain synapses and of course with repeated training long or lasting learning occurs typically by more structural kind of changes like the adding and subtracting those synapses or even I'm sorry of receptors or synapses and this is essentially the mechanisms of long-term memory so perceptual learning is a little different than situation and sensitization just to refresh your memory perceptual learning repeated exposures to a stimulus lead to improved ability to recognize process that stimulus how does the brain get better at processing an input well let's look at how the brain processes input a little bit to get a feel for this so here is the primary somatosensory cortex or s1 sometimes it's called and this is in the gyrus just posterior right behind the central sulcus that crack which separates the frontal from the parietal lobe we've been over this there there's like I said there's a sensory cortex for every sense touch taste hearing blah blah blah blah this is the one for touch and each neuron in here has a distinct receptive field there's that eye that concept I introduced before again which is essentially for a given neuron in a sensory cortex its receptive field is the part of the world it responds to so in this case for example you'll see their face drawn on the cortex neurons near in that area receive touch information from lips from the lips basically so if I stuck an electrode in there and delivered some a little zap of electricity you would feel as if there was some sensation coming from your lips you feel I don't know what tingling or something like that even though nothing is really touching your lips I'm just going in there and kind of sending a signal directly into the hardware that receives the signals from the lips so that the I would point out again that when we heard that word receptive field before we were talking about the visual sense and there are neurons back there in the visual cortex right feature detectors that are called that respond to a bar of light at a certain point in space moving a certain way that's the feature that neuron likes in other words that point in space that feature out there in that place is that neurons receptive field in the case of the touch cortex since that neuron they're in that lip part of the cortex the lip is that neurons receptive field okay and when the receptive fields for adjacent neurons represent places out there in the world that are next to each other we say that the cortex preserves a topographic map all right the organization in the cortex reflects the organization out there in the real world this is a little bit of review just to orient us when we go in here to look at how perceptual learning occurs so how do we find and define receptive fields well for the visual stimulus it's some sense it's pretty easy monitor bunch of neurons and then show the retina a bunch of things in different parts of the retina stuff and you wait to see if you can find a neuron that likes a certain point in space this example is looking at the auditory cortex which is in the top of the temporal lobe primary auditory cortex and this is where the auditory sense goes after from the ears to the thalamus of course but then to this primary auditory cortex and what you find is if you monitor using single unit recording the activity of individual neurons there you can find ones who respond to a certain pitch as determined by the frequency of the sound so what's shown here is the normal response and spikes per second you can see that this neuron really doesn't fire much at all in response to 0.1 Hertz that's the cycles per second it's just a frequency that determines the pitch of a sound and as you increase the frequency of that pitch in other words it goes up the neuron doesn't really care much until you get close to around 1 Hertz or point 9 hurt kilohertz I'm sorry these are kilohertz not Hertz and then it starts firing like crazy and then once you get up and keep moving the pitch up going higher and higher it reduces responding so from looking at this graph you can pretty much tell okay we found a neuron whose receptive field is about nine kilohertz in the auditory cortex now how my experience increase this organisms ability to detect here tell the difference between different pitches it's actually really interesting if you train animals to on some pitch distinguish the pitches tasks like you know when it's one point one kilohertz you get a treat for pressing the button if it's 0.8 kilohertz or lower or you get a shock for pressing the button something like that where it matters and you can see behaviorally whether they can tell the difference so if you try to do a bunch of training like that and then you go back and look at these graphs you actually see that the humps get skinnier they get tuned so the neuron this receptive field has gotten smaller it's a sharper basically there you've tuned up this sensory cortex so that it is better able to make fine distinctions between in this case different pitches different stimuli so there is a cool example that we can see in real time how perceptual learning can manifest in sensory cortex by sharpening or tuning the receptive fields we'll look at another example of perceptual learning in the brain in a sec here this slides basically just the visual to go along with my description of topographic maps when adjacent neurons and sensory cortex have receptive fields that are adjacent out in the world the cortical arrangement is called topographic that's a much cleaner way of saying it then I did a couple slides ago for example cell with the receptive field for the thumb in sensory cortex that is somatosensory cortex body says touch is right next to that with the receptive field for the finger the auditory cortex which is shown in that picture on the bottom there that little pink spot on the top of the temporal lobe is also topographically organize we'd call it a tonotopic method because it's organized by pitch or tone frequency of sound and for example it goes from bass ear sounds to more treble e sounds going from anterior to posterior front to back as you can see in that zoomed in picture there so you know cells with receptive fields for a pitch of 500 Hertz are next to those for a thousand which are next to those four two thousand four thousand etc so now we know our visual cortex the touch primary sense or somatosensory cortex and the auditory cortex and the motor cortex actually are all topographically organized so it might seem like we're talking about some very specific thing here you know perceptual learning but the truth is this whole concept of how experience sharpens and shapes cortical networks and for example narrows and sharpens the receptive fields for individual neurons and the auditory cortex is really a good way of understanding a lot of how what happens in the brain during development it's this type of pruning narrowing of receptive fields sharpening of function elimination of some synapses and neurons so like I don't know if you've heard this like at 1 year or something like that you've got the most neurons and synapses that you'll ever have development that happens after that and of course I would argue it's probably most of the development certainly from a cognitive point of view happens by virtue of tuning pruning away there's more pruning than making those synapses let me put it that way but certain synapses are strengthened but for the most part it happens by pairing out the extra stuff it's kind of like starting with a big block of cheese and then life whittles away parts of it to reveal a beautiful cheese sculpture of a cow or something like that and to make that point these are these are just examples of that point so if you deprive an animal of perceptual input at birth say vision and then restore it later ok you can open your eyes they will never develop normal vision I mentioned this when we were within the context of cats and we were talking about though feature detectors and perception there but like I said you've got them neural machinery there that is able to learn things but it has to interact with the world be as experience prunes shapes sharpens feature detectors and narrows receptive fields and a bunch of other stuff like that and without that experience the neural network doesn't tune properly so the ability to see is not just you know light going into the eyes obviously there's an incredible amount of machinery back there that has to be shaped by evolution and experience and also interestingly you know to illustrate that it is a plastic a fluid system if you blind an opossum at birth what you'll find in the adult is that the visual cortex responds to tactile and auditory pottery auditory input and sensory and auditory cortices are enlarged so they've got all this neural hardware that the animal is born with to help it to see right but if no input is put into that part of the brain it doesn't just die or sit there idling it is shrinks probably but also parts of it are co-opted by other senses so you actually get visual cortex responding to sense from the auditory or tactile modality meaning touch and hearing so in other words the visual cortex didn't have anything to do because no visual input was coming in so it said well I'll help out with hearing it's not didn't obviously say that but the nature of how experience shapes cortical Maps and what does what resulted in what would normally be used for seeing being used for touching and of course also the sensory and auditory cortices were enlarged relative to normal because the animal had to the importance of those senses was increased because they didn't have any visual sense and so they became basically more expert in hearing and touching just the way a blind human might be have a increased sense of touch it would have finer receptive fields in the area of touch we will look at an example of that in a second alright so the last bullet there I've made this point hundred times we're born with the perceptual hardware but experience is required to tune those systems organize topographic maps etc or they don't work okay so zooming back out a little bit to our larger thread here how does perceptual learning happen in the brain via cortical plasticity of course and refinement in the receptive fields of neurons on the sensory cortex do to develop enable and experience primarily and even modest practice can at least temporarily rewire these maps and that's how you can see evidence of perceptual learning in a single experiment but of course repeating practice can make this permanent right if you're a professional musician your auditory cortex is without a doubt more finely tuned shall we say then your average Joe Blow because of a lifetime of perceptual learning has resulted in long-lasting synaptic changes and that result in finer receptive fields in the auditory cortex so here's an example of perceptual learning and cortical plasticity from a study and the task here the senses is touch and the task is where you basically take two toothpicks or pins and you touch them to someone's finger or you can do it anywhere in the body and basically you start with them really really close and people aren't looking right their eyes are closed or if they're looking away and it feels like one pin because they are right on top of each other and then you start separating them a tiny bit at a time and keep pricking pricking pricking until someone the participant can say okay wait I feel two pricks now not one and that is that distance right it's much smaller for your finger than your back right because you have much better touch in your finger then in your back because you have much more sensory cortex devoted to processing touch information from your head and your back and as a result your ability to sense pinpricks is much finer on your hands so the distance I have to pull those apart before you can tell there's two things there it's much smaller for your finger than your back so that's called sensitivity right the the closer together they can be where you can still tell there's two of them then the more sensitive you are anyway what they had here was they took a bunch of people they measured their baseline sensitivity which again is a distance the smaller it is the finer your ability to discriminate touch stimuli and they scanned people doing this task and looked at where activity was in the brain where do you think it would be that's right this primary somatosensory your body sense cortex that is associated with the finger so they did two hours of training on this after they took those initial scans and got the initial sensitivity they did two hours of training with the right hand only all right and so they touch and if people got it wrong they give them feedback or even if they got it right they give them feedback basically they practiced a bunch and then they found their smallest distance they could tell there were two pinpricks basically re-measured the sensitivity and of course gave them another fMRI scan and found that they got much better meaning they could sense when there was two pricks at a much closer distance after training than before and also only in the hand that they trained on right because the right hand is this information from the right hand goes to the left sensory cortex so the learning occurred primarily only in that one half of the brain so it only applied to that one hand and also I guess it was going to happen they showed greater area of activation in the part of the sensory cortex that correspond it to the right finger so basically they saw even with just two hours of training sensory learning right people got perceptual learning rather people got had more finely tuned receptive fields for pinpricks on the finger and they saw in the brain a corresponding increase in activity in that part of the brain these pictures it's not really clear if you don't know what you're looking for it's not like one's all dark and then ones all lit up but trust me that part of the sensory cortex devoted to the right finger has more activation in it in the after picture than in the before picture and this is not as true with the left finger so just two hours of practice can rewire literally the cortex resulting in perceptual learning however these changes in this case were temporary they brought them back the next day and found that the improvements that were made on the their-their threshold went back to normal so they didn't mean then last the night basically and so some as we've already pointed out some of these mechanisms happen very quickly and don't last as long some last longer and take longer and apparently the ones that were in play in this finger-prick experiment we're not permanent didn't last very long and that graph there left i F stands for Lex index finger and right is right index finger and the point here is the right index fender where they practiced the threshold distance went down after all the practice that demonstrates perceptual learning but when they came back the next day was right back up where it was before they practiced however of course if they were practiced every day for a month then different longer-lasting neural mechanisms would come into play and the cortical plasticity would be longer-lasting obviously in order to their old plasticity and this is looking at this in the context of perceptual learning but really these mechanisms underlie all learning and memory and ways that experience is recorded saved leaves an impression a lasting useful or not useful impression in your brain it's all through these types of neural plasticity mechanisms long term potentiation long term depression which we talked about heavy and learning which is why neurons that fire together wire together remember that ed was a pioneer in the field of how neural networks do what they do and yeah came up with that catchy phrase neurons that fire together wire together we already looked at an example of this long-term potentiation and the flip side of that coin of course is that cells out of synch lose their link another catchy phrase as in long term depression that is when you get that you intentionally cause cells to fire out of sync the connection between them weakens and hem suggested that this kind of experience based natural selection for the most useful synaptic connections was a primary mechanism driving learning and memory so it's basically like I said before if synapses neurons synapses are used a lot they strengthen if they're not they weaken and you might consider how this explains priming right the form of perceptual learning we talked about before where at the beginning of the lecture some of you may have consciously or unconsciously seeing the word moth on the bottom of the slide I admit anything and then when I showed you the word completion task mo blank and said pick a word that completes this put in some letters here it could make a word priming would have been demonstrated by you being more likely produced the word moth than someone who hadn't seen moth recently who might produce a motel which is a more common mo word apparently and if you think about this in terms of heavy learning this is just an analogy it's not like there's literally a neuron for him and a neuron for oh but if you think of each of those nodes right those little circles in the spider webby things at the top of this as representing a letter in this case when you see the word moth the M and the O and the T and the H are all active at the same time of course the concept moth and all sorts of other things about it are activated too this is a oversimplified example but they're active at the same time okay so what happens when neurons are active at the same time the connections between them are strengthened right so then later when over on the right here when you just see the M and the O indicated by those two bottom purple nodes what happens well because the connections between those and the T and the H have been strengthened that activity spreads to the T and AIDS which become activated and the end result in your little hallucination you call reality is that you go oh moth it pops into your head why well it's been primed how does that affect what pops into your head via long-term potentiation because neurons that fire together wire together all right another type of learning that is covered in this chapter it's not sensitization or situation or really perceptual learning but it would fall into the same categories of non-associative learning that is it's not about learning that two things go together it certainly wouldn't fall under the category of classical or operant conditioning so it's in this chapter is the formation of mental maps and this type of learning is also incidental there's no animals do it without being rewarded it just happens and we actually can see how it happens in neural networks pretty clearly because there are these really cool neurons that have been discovered in the hippocampus called place cells which are neurons that if you use single unit recording and get an electrode in there to monitor their activity they respond to or their receptive field you might say is a certain point in space the organism being in in place and so in the example here if you just look at the leftmost figures you've got a three arm maze that a rat can explore your monitoring neurons in the hippocampus in this case they found a play cell that likes a certain place if you look at the bottom row the bottom-left figure you'll see on the bottom left leg there are is a little dark patch basically what they're doing here is letting the animal run free and explore the space while we monitor this neurons activity every time the animal goes into that little spot that's dark it late in leg two there this neuron fires like crazy and it turns out there are other neurons that represent other points in space and some that actually have a grid like organization the point is these place cells seem to as soon as an organism is in an environment start to get tuned up that is create a neural substrate for a mental map of space and what is demonstrated here is a study that tried to figure out okay well what cues or information is the animal's brain here using to tune this map up that is what exactly is the bat neuron responding to because one interesting point is it's not visual input and on a simple level because if the animal is facing one direction they might have a completely different view of the room but then if they're facing in the opposite direction but this neuron fires equally in either case so it's not responding to a simple visual input however it has to be responding to something has to be telling the animal where it is it could be you know birds have some like kind of ability to sense magnetic fields and navigate over great distances maybe on the basis of the Earth's magnetic field so that sounds crazy but it's not as crazy as you might think so is that what's going on with the rat here or what and so what they did is they found they had a that card is basically a cue it's just a something in the room that helps distinguish one light from another and they learn no that's between legs two and three not in that sense but you get the point so what they did then was they were Swan during you know okay so is that neuron tuning in response to that local cue or larger room cues or even larger like I don't know the cardinal directions like which ones north and south and so what they did is they rotated the arms of the maize either with or without the card the cue that is right so they found that this place cell liked that lower leg on the far left there in leg - then what they did is they rotated the whole maze all right and they were like well if it's responding to some cardinal direction like the earth or the larger room then it should this place cell should now be firing when it's in leg one because that's where leg two was they took the rat out before they did the rotation of course and they cleaned all the smells out and all that stuff but they found that actually the place cells receptive field moved with the cue so it appears that what's actually happening is that this is being controlled this place L is controlled by that local cue that's really all this shows and then on the last column on the right this just shows that if they move the the maze and the card two different vases in a different orientation within the room and the car is in a different place relative to the maze the tuning still follows the relative location of the card it's always basically to the left of the card is the place that in Iran that reasons the larger point here though is just that look place cells aren't those cool and this is an example of experience causing a change in neural activity and in this case the changes result in a mapping out of space and this is an example of incidental learning because like I said it just happens automatically you put a rat in an environment it will just start you know looking for food sniffing running around the walls but its brain will automatically start creating incidentally right without being rewarded or punished or anything like that will immediately start forming a mental map of the space that it's in and this map is controlled shaped at least in this example by local visual cues alright here's just a summary slide again I will let you read this on your own since it is stuff I have already said you can pause it and go through it so I'll finish up with a couple clinical perspectives here that is sort of looking at a few clinical situations where we might bring what we've learned in this chapter to bear when thinking about how to treat these conditions so stroke for example is when you have disruption of blood flow to the brain which often results in cell death and trouble you know there's a stroke in a certain place you might have paralysis on the certainly half of the body because the contralateral motor cortex or has been damaged or something like that and as if I'm para fine you know I have paralysis in my left arm it might be because there was a stroke in my right motor cortex that controls my left arm and stroke is a relatively common unfortunately the leading cause of brain damage in the u.s. apparently and one of the pernicious things that happens when you have a stroke for example if my left arm were paralyzed due to a stroke what am I going to do in life I'm gonna use my right hand more because it works better right my left one is maybe it's not completely paralyzed but it's not as efficient it's clumsy whatever and what happens based on what we've learned about how neural networks work how plasticity works how they're shaped well neurons that fire together wire together right survive on the fittest natural selection selects active synapses over non active synapses not being used receptors are pruned away etc so if you I'm not using my left hand because that part of the cortex has been damaged it just gets worse and worse and worse it's a vicious cycle really because if you don't use those synapses they continue to degenerate and get weaker and so the condition gets worse and worse so one very clever can anyone think of a simple clever way to treat this condition for example if I had a stroke to my right cortex and my left off arm have reduced function in that arm well if anyone said strap your right arm to your body so you can't use it and force the person to use their left arm that's exactly right it's called constraint induced movement therapy or C IMT and it is very effective here's some data I don't know if these are even but it's probably real data but it's pretty crude here but you can basically see that we've got two groups one who used constraint induced movement therapy so let's say we've got two groups of stroke patients they all have reduced the ability to use their left arm we strap half of their right arms to their body and force them use their left arms and then measure their ability to use their damaged arm their left arm right after treating a month later a year later and as you can see not only do they have greatly improved outcomes over the group that didn't use constraint induced movement therapy but these benefits last up to a year later so by forcing you to use even though it's a struggle it's not as efficient at first what happens let Burt the natural neural plasticity that brains have the ability to do this incredible ability to heal and recover work and in order for them to do that you have to force them to you have to sort of force the brain to try to do what it doesn't want to in order to get it to get better at it this is true in general if you hate doing math you tend to avoid it right but really you should be leaning into that to force your brain to practice that and rewire so that you minimize that weakness so that it gets better at doing that oh and as an aside here just at the end so what is actually happening up there well if a neuron the dead it's dead it's not gonna come back to life okay but neurons nearby that may have been doing something slightly different if you keep trying to use those now dead neurons the brain has a way of co-opting nearby neurons to take on that Duty where synapse may have been only mildly related to using your left arm before because other neurons were doing the bulk of the work when those other neurons died well that partial functionality that that synapse may have had towards using left arm now is going to explode increased and that synapse will become much more strongly associated with controlling left arm and because it's there it's alive it can another example of the incredible plasticity of our neural networks is demonstrated by man-machine interfaces I already showed you that one word you can read the population vector from a bunch of neurons in the motor cortex to get someone to literally control a machine with their mind so to speak another example here is cochlear implants a much common much more common example of man-machine interfaces and these are in fairly regular use I think and what happens here and this is obviously to treat deafness is that say your eardrum is broken or the little bones in your ear are atrophied or whatever normally you know sound waves hit the eardrum that your drum vibrates it Wiggles those little bones you can see in the picture on the right to stir up that blob a hammer they involve whatever those things are called they jiggle a little window on that snail full of jelly and the vibrations cause a little tone that is in that snail to jiggle and the jelly and little hairs touching that tongue move and those hairs turn that movement into action potentials which depending on where in the jelly it's moving the thinner part responds more to high frequencies the fatter part low frequencies etc determines what pitch ends up popping into your little hallucination you call reality but if some of those bones are damaged that jiggling can't get in there so one solution that they've come up with is you put a little microphone on the outside of a person and run a wire from it into that snail full of jelly alright now remember in a normal functioning person there's a membrane in there like a tongue that kind of jiggles when the vibrations come in and hairs touch this and they move at certain frequencies in at different places that and turned that movement into action potentials it's kind of like the photoreceptor cells in the eye turn photons into action potentials these turn that jiggling into action potentials so what we're doing now is microphone and turning it into an electrical signal and just sending that electricity into the snail that is nothing like little vibrations being read by hairs it's just electricity in there that has a correspondence to the sounds in the outside world so the point here I'm trying to make is that avenge this does not it's not like you run this into someone's ear and they're like hey I can hear just like I could before of course they can't they're not getting vibrations like they did before they're just getting electrical signals so what does it sound like I've heard it's quite jarring but it's definitely a mess what has to happen is they have to interact with the world get a bunch of this input to go into those sensory neurons and eventually the neural network due to neural plasticity learns to interpret the data that it's getting and eventually the sounds that come from the outside world pop into your little hallucination you call reality the same way they used to because the neural network has taken the data even though it's in a very different form and learned how to process that in such a way that results in intelligible hearing a couple of notes about cochlear implants here I already said it takes obviously training too and for the brain to interpret these that's the neural plasticity and this learning happens quicker towards the beginning of rewiring and then the gains generally slow which is kind of true of everything the better you get at playing the violin the slower the strides you make are when you first start you're making these giant leaps because you suck and recency of hearing loss is a big factor an implant success probably because if you this machinery hasn't been interpreting information from the outside world for 30 years well what happens we know that neurons synapses that aren't use tend to get pruned away so a lot of the hardware you'll need isn't going to be as well connected anymore