Eyes on Reading: Dr. Stanislas Dehaene with Emily Hanford

good morning everyone I'm Ann fredman I'm the founder and CEO of Planet word and uh like many of you here I would guess a former or current reading teacher so I'm especially pleased that you came out early on this Saturday and especially if it's your usual day off to attend a lecture on the reading brain but also because I am a former reading teacher I'm not at all surprised to see any of you here because you're probably as eager as I was to always learn everything possible about how kids learn to read and to improve

your practice to benefit all your students we're especially grateful to Emily Hanford Planet word's very first journalist in residence for inviting Dr danan to speak to us today it's the first of many Explorations that we'll have in our continuing eyes on reading series as the producer and host of the award-winning podcast solda story Emily has helped to shift the discussion and practice about how we teach reading in America through her research and Reporting Emily has made us all aware of the importance of evidence-based instruction and has introduced the phrase the science of reading to A

Wider General audience one filled with parents seeking to understand why their children aren't learning to read proficiently and also reaching teachers who are worried about why their instructional methods aren't always effective so when Emily approached us about becoming a journalist in Residence at Planet word we jumped at the opportunity Planet word was created as a museum about words and language with a mission of finding ways to engage in Inspire visitors to love words language and reading and with a vision that the foundation of a strong democracy depended upon a literate populace but we knew that

even though we weren't about to teach reading ourselves we needed to be advocates for best practices in Reading instruction and become a hub for literacy how could we achieve our mission unless children were being taught to read successfully in our schools so joining up with Emily who would use her skills as a reporter and her curiosity would help us achieve our goals with her connections through the literacy space Emily could help attract leading researchers and thinkers to Planet word stage that was an opportunity too good for us to pass up and a top EMILY's List

of people to invite to speak at Planet word with Dr sisl daan we are thrilled that she is that he is able to join us on one of his rare visits to the United States and I want to also thank my colleagues Caitlyn Martin and uh Brit oats for helping set this up along with Emily and now here is Emily to tell you more about our special guest [Applause] good morning thank you all so much for coming uh I know we have people from all over the country I think we have Wisconsin Arkansas Pennsylvania Tennessee

anywhere else and all over Washington the Washington DC area so thank you so much for coming and I'm very grateful to have this opportunity to be uh journalists and residents at Planet word thank you very much for that nice introduction an I think this is such a a cool Museum first time I came here with my mother and my son a few years ago I thought I need to find out more about this place and I just think it's a wonderful we we're just hearing a lot of noise outside because I think there's a school

group here today so language and children uh and reading all my favorite things so let me introduce our speaker today uh Stannis danan has no idea how much he has taught me uh when I reached out to him about coming here to Planet word I sent him a photo of my copy of his book reading in the brain which I'm sure many of you have read and is this is a lot of my books look like this but I think this is probably um the ultimate example of the most sticky noted book on my shelf

um I first learned about Dr danan and his work back in 2018 I was just beginning to learn about this big body of research that people are referring to these days in the United States as the science of reading uh Dr danan comes to us from France he is a psychologist and cognitive neuroscientist who directs the neuros spin Center which is France's Advanced Brain Imaging Research Center his research investigates the neural bases of human cognitive functions such as reading and also calculation and language and he has a particular interest in the impact of education on

the brain there's a quote at the beginning of his 2020 book how we learn which I highly recommend if you haven't read that one yet this quote is from the former president of MIT and he said if we don't know how we learn how on Earth do we know how to teach that's what Dr danan is here to talk with us about today what we all need to know about learning about the reading brain and what we can do what we need to know about this to do a better job teaching all of our kids

to learn how to read so please welcome Dr stanos Duan thank you thank you so much for this wonderful hospitality and welcome um I am extremely happy to be here and to tell you a little bit about my work I'm I'm a brain scientist um but more and more I am interested in the impact of Education as you mentioned and also the impact of our science on uh education whether we can help uh to correct some of the difficulties that the education system has so um this is new spin the place that I direct uh

in the south of Paris about 15 km from Paris each of these arches that you see here um has a magnet an fmri system or another technique to image the brain we have all of these techniques the entire building is dedicated to trying to understand the brain um and in particular um we have currently the world's largest fmri magnet it's quite amazing it's 11.7 Tesla magnet that you see here this giant machine here which is just just starting to work so we're trying to put all of these techniques to the service of understanding the human

brain in particular because we think the human brain is very special it's the only one on the planet which has language which is able to do science mathematics and of course reading and to this aim uh my wife and I this is my wife Jen we're trying to put all of these machines to the service of understanding the child's brain in particular the baby's brain she's really working with very young babies and um we I'm working on adults we're joining when it comes to school children uh sort of um so you can see that basically

it's becoming feasible to deploy all of these techniques whether this is MRI or EG or just Behavior or you can see Meg here magnetoencephalography this is electroencephalography we can deploy all of these techniques safely uh and uh with the collaboration of these children who are like little cosmon notes in the magnet we can uh we can really try to explore how the brain works so that's what I'm going to tell you about um first discovery 20 something years ago uh my wife and I were the first to put uh infants few months old infants in

uh the magnet and uh to scan their brains and to see how the brains work when they listen to language and we knew that they have a lot of language abilities already at that age but it was very striking to see this gorgeous image showing that when they listen to spoke language they have a network of areas it's not the whole brain lighting up as you can see but they have a network of areas that is already wired to channel and to process spoken language and uh it's actually very similar to the adult language Network

which you can see on the left um essentially you can see the temporal lobe Brokers area in the inferior frontal region here lighting up of course I'm not claiming that the child is already understanding the word or the syntax of the language but the signal from speech is already channeled to these areas that are going to acquire uh spoken language and this is my first message when it comes to spoken language we have a clear Network which already there very early on in fact in the first year of Life way before the child starts to

speak this network is already learning enormously about the structure of spoken language and we know now that this spoken language is completely Universal um you you can see here the left hemisphere of uh I think something like 45 different uh people belonging to different language groups this is the recent work from fedorenko and MIT you can see that regardless of the language that you speak this network is going to light up when you process uh spoken language um it's also Universal in the sense that if you are bilingual you'll get very very similar if not

exactly identical activation to each of your two languages it might be a little bit different if you learn one of these languages later then you will have recruitment of additional areas but if you learn the languages early enough there will be a lot of overlap if not perfect overlap between two languages so we have this um we we think it's an evolved uh property of the human brain to have this spoken language Network and um what happens when we learn to read well we're going to reuse the very same network in order to understand written

language so uh written and spoken language converge to very very similar brain areas here you see an image from the work of Jack Gallant at Berkeley um it's a beautiful image because it's a single person now it's not a group just one person and the colors indicate different domains of meaning of the words the person is listening to so uh the colors indicate for instance in in pink and in red the social um related words these areas here are related to mathematics numbers and space but you can see the point of this slide is that

you see very very similar if not identical activations to spoken and written language this is the miracle of uh reading acquisition when we learn to read we essentially learn to access this existing spoken language system through a new modality Vision which was not anticipated by Evolution this is an invention now so we have this evolved spoken language system but we learn to access it with a new invention uh the invention of the alphabet and of writing systems so um just a very simple slide here of course I will try to simplify the findings um as

much as I can to to convey them to you but here is your left hemisphere before you learn to read you have this spoken language Network which has knowledge of the phology the Lexicon the syntax the meaning of spoken words and sentences and when you learn to you're going to access it through the occipital pole which has Vision which where all of the visual inputs enter into the brain you're going to develop a specific brain region that serves as the interface that recognizes the letters and sends the message about letters to the language areas and

in particular to a region that we call the planum temporal where um we have a representation of the fores of speech and we believe that this connection between letters and sounds is the key to unlocking reading after that you will be able to go faster and maybe more directly to the Lexicon but initially this is the key link here so this area which is the interface I've called the visual word form area was my colleague luron Coen but you might call it the brain's letter box because this is the one place in your brain that

has enormous knowledge of the statistics of letters and this is an area that has developed because you've learned to read okay so we see here the the logic of what I'm going to tell you learning to read is not a complete change of brain networks it's a recycling that's the term I've been using it's a reuse of existing areas turning them reorienting them into a new direction we're going to reuse the language system for sure but we're going to reuse it through vision and therefore we're going to have to recycle an area that already exist

in the child's brain but that's going to be reoriented to the recognition of written letters written words okay so first I would like to show you some experiments about what changes in the brain when we learn to read you're not able to learn to read this Rosetta Stone maybe there is an egyptologist in the room that can read it but okay so what what happens when you learn to read we did I think what was a key experiment difficult one we wanted to check sort of map of all of the differences between a literate and

an illiterate person and we didn't want illiterate people just because they dropped out of school we wanted people who just didn't have the chance to go to school at all so we collaborated with colleagues in Portugal and in Brazil it was a very International experiment we found people that uh lived in villages or poor families that simply didn't have the chance to be sent to school at all and didn't learn to read and we scanned them um and correlated their brain activity with their reading score so what you see here is brain activity versus how

many words they could read per minute so the illiterates are all the way on the bottom left um and we'll see increasing activation in an entire circuit in people who have better and better reading abilities of course there were some people who hadn't been to school we call them ex literates they were able to learn to read from their Community or by going to school later on as young adults so in this way we could correlate reading ability with changes in the brain and it's just one slide for a huge amount of work there are

many papers on this work but um simply speaking first of all you see here the left hemisphere language areas of the temporal and inferior frontal region Brokers area lighting up when we show a little sentence visually and of course the better a reader you are the more you can access the spoken language system so the more activity you have there that's a first finding accessing the language system that's what you can do when you're literate second you see here the bottom of the brains like we took the brain and looked at the bottom and you

see these visual areas that are hugely red because they have more and more activation the better a reader you are and this is the visual word for area this is the brain's letter box and Associated areas and third we made a finding that these more dorsal regions of the temporal lobe not exactly shown here they were also changing even for spoken language this is bit less intuitive because you've learned to read you are processing spoken language better you are hearing the phones better and you are able to do phological awareness task you're able to move

phones around in your mind you're able to play with phones and we found that there is almost a doubling of the brain activity to spoken language as a function of how good a reader you are as a function of reading score so that's a huge transformation we don't fully know you know at the cellular level what it means but there is a huge enhancement of responses in this area we also found the connection between these systems if we look just at the anatomy of the connections of the brain with diffusion Imaging was uh reinforced so

you can see the network I was showing you on the previous slide here visual areas connecting to phological areas and allowing access to the whole language system I want to focus a little bit on this area on the bottom here which I call the brains letter box the visual word form area because this is really the key the entry of vision into language and we know a lot about this region from many many studies so just to summarize first of all it's always at the same location all of you in this room regardless of what

you read you're going to have this left hemisphere activation plus or minus 5 to 10 millimeters it's always the same location um in all cultures it will activate more to written words than to anything else so you can see it here for instance if you have an electrode in the right location you see this activation to letter strings and pseudo words any WR material above anything else like seeing objects scenes faces and so on same here in fmri little patches of Cortex activate very selectively to words among other areas that respond to objects faces houses

and so on and uh it will activate only to the scripts that you know so if you have not learned Chinese it will not activate to Chinese but if you have learned Chinese it will okay it will not activate to Hebrew if you have not learned Hebrew and so on so forth and if you have a lesion in area this was the case in this patient here we have several such patients you will totally lose the ability to read so there's a causal relation and uh when I say totally these people describe waking up in

the morning a small infa for instance to this area and they say well I look at the newspaper and I don't understand why is it printed in Hebrew it looks like Hebrew to me I I cannot I can see this people are not blind but they cannot recognize any of the letters that's what this area is doing for you okay and it's very hard to relearn as an adult so it's really based on child plasticity that this area is able to uh be tuned to letters it works completely unconsciously when you are an adult uh

We've shown using subliminal materials that will activate whenever your eyes land on the word even if you're not aware of it um here is a very recent image from our seven Tesla magnet now we can do single subject images uh so this is one particular person who happens to be bilingual in English and French and we created this whole hierarchy of stemi they're all six letters you can see some of them are French words some of them are English words and then we have approximations to English or French and more and more distant statistics all

the way down to things that could be a word I suppose in Polish perhaps but not in French or in English right and you can see um very well this reading system here in the back of the brain uh visual areas was respond equally well to all of these because they have you know visual stimulation bars maybe even shapes of letters As you move to the front you get more and more selective to the stimuli that you've learned these brain areas have learned the statistics of French and English they've done the shapes of letters and

they they've learned how the letters go together and they do exactly the same job whether you're reading in English or in French if you're English French bilingual maybe some of you are bilingual also in Spanish for instance as long as the alphabet is common it looks like we using exactly the same patches of Cortex you can see how precise it is we used to call it the visual word form area but it's actually not one area it's a constellation of little patches of Cortex because you've learned to read you've specialized some of these patches of

Cortex um so yeah they provide a very fast encoding of letters and the relative locations of letters that's important as well you know that words can be anag s of each other I suppose in a museum you maybe you play with anagrams well your brain can distinguish anagrams because this area is caring about not just the existence of letters but the relative locations of these letters it tells you that this letter comes first this one comes last in the word that you're currently seeing here is the brain of another person who happens to be English

Chinese bilingual and here we made a little bit of a discovery uh reading Chinese is almost the same as reading in English and we have a lot of things that are in common so we created a similar hierarchy of stemi that are impossible or possible in English and impossible or possible in Chinese if you read Chinese you'll know that this is a word I think maybe this is a word but this one is not and so on so forth and you can see the hierarchy here so most of the patches of Cortex are common even

when you're reading Chinese and English but a few are specific to Chinese so here you see this very strong response to Chinese but not to English which is the lighter green bar and strong response to Chinese but not to English here we're still trying to understand the system there are these patches of Cortex that specialize and because the shapes of Chinese characters are a bit special we think that the cortex has to dedicate some special resources to that one thing we don't understand is that there is this proximity to faces and reading is a little

bit like face recognition once you're an expert it's very parallel processing and it turns out that the patch that respond to Chinese also respond to faces so sometimes we see competition sometimes we see overlap we don't fully understand it but it's very very similar let's say so these are all images from adults but we wanted to get images from children as well and so we started to do these Imaging experiments in children it's not so easy to scan children you have to give them stimula in a task we ended up choosing this nice task which

I highly recommend you just flash pictures some of them are pictures of words or faces or houses or the checkers and all they have to do is watch these and click when they see a star it's a bit like looking for Walo we actually use Walo in one of one of our experiments so they like it you know they they are clicking for Walo but we are interested in these categories that they are not clicking to and we see activations to them and if you scan a nine-year-old this is a group of nine-year-olds here you

can clearly see the reading system is already in place you have this huge activation to words more than to these other categories the other categories are are in other colors you can see for instance the face system there's a skin little activation to faces in the left hemisphere right next to the visual World form area there's a huge activation to faces in the right hemisphere you see this Topography of areas you all have them in this room by this whole systematic organization of objects faces places and words if we scan a six-year-old who is not

a reader that's what what we see much less organized system no visual word form area activation it's not yet responsive to letters at least not selectively it's not just a matter of age because if we scan kids of about the same age we see this very strong activation to written words and you see a sort of intermediate pattern here it's not fully organized not as strong and the face system is already not yet fully organized but it's coming there as soon as you learn to read you see this activation to written letters coming up um

it's really not a matter of age so in this more recent study uh We've scan more children we've added more children and we've especially been looking for children who learn to read Maybe by themselves uh but from preschool so in this way we could have very early beginning readers here that you can see already have this visual word form activation selective to words even though they're quite young and in this way we could decorrelate age versus experience in reading and what we find is that age does not really matter uh what matters is the experience

of reading for the development of this area so if we correlate brain activity with both age and how well you how good a reader you are it's really how good a reader you are which decides the amount of activation here it's completely different for faces for faces it's a very slow development does not depend on a critical input of of course you're meeting faces since you are born and very slowly this system specializes for recognizing faces um I think the analogy is interesting here with our schooling system we're taking this brain which has the ability

to learn things like faces or objects and we just at a very precise moment in first grade we decide that now you're going to get an enormous input of letters and we feed it with letters and letters and letters and we see that suddenly the system will develop patches of Cortex that specialize for letters uh we have a triggering let's say of this system by the the school input what I was showing you until now is group studies but we are also able to follow single children as they learn to read um and these are

called longitudinal studies as opposed to the cross-sectional studies that I was showing you before um in particular a few years ago we got the permission to Scan 10 children up to six or seven times each which is a bit unusual it's completely safe by the way we can talk about the safety of MRI well you know three Tesla MRI is a high magnetic field you could have questions about that so maybe you we can answer your questions later on but uh we got permission easily actually and we had no problem whatsoever in fact the kids

loved to come back to the lab we made a whole point of you know explaining to them the brain and so on we actually printed in 3D little pieces of their brains so little little brains like that that they could take home it was their actual brain their individual brain that we could print in 3D so they they loved it anyways we got good data so we were able our goal was to have two scans before school at the end of preschool four scans all along the school year and for eight kids we had one

more scan a year later so after two years of experience of reading this is quite unique experiment I don't think there's a similar experiment in the literature even now so you can see there were 10 kids here we selected 10 kids that uh started without knowing how to read so in preschool they had zero this is the school entry here they are they had essentially zero knowledge of graim Phim relationships which is perhaps a bit unusual most kids begin to have some knowledge of graim Phim relationships early on but we selected them for that so

that they very quickly the upon School Entry School in France is doing a relatively good job not perfect but you can see all of these kids shooting up learning fony relationships and a bit more slowly becoming fluent this is the number of words they could read per minute which is a bit slower okay so they were learning so what was changing in their brain here is the brain of one of these children and you can see the successive scan so preschool very little activation to the written word compared to other categories like I showed you

before suddenly this is school entry here and in September we have already this reading brain lighting up there are many areas but I'm going to zoom in on this ventral temporal lobe and occipital temporal region here you can see already the emergence of the visual wordform area and also this temporal Superior temporal ccus activation here they're already in place and they're just going to grow up a little bit more but they're already in place right from the start this is one of the scientific questions we wanted to ask you know do you have to search

for this circuit is there a massive activation at the beginning following by now going down or something like no there is a there is a circuit which is probably wired already maybe from birth which happens to connect vision and language and we recycle it in order to acquire the shapes of letters at the same time we see a very slow change in the face system so again there's this nice contrast faces you're seeing from the start and you we can see the face system even in preschool it's if you look very closely you see that

it's very very slightly changed is barely visible here but if we do the statistics we'll see that it's growing in the right Hemisphere and that's something that we've seen that there is a very fast change in the left hemisphere for learning to read and that leads to a little bit of a shifting of the faces towards the right hemisphere because of competition between these two types of learning so this is a vision it's a model but this is a model of how the visual cortex may change uh in these children before schooling there's already an

organized system and this system comprises patches of Cortex that care about faces patches that care about houses and places patches that care about tools and objects and some of them are still labile and we think that if you learn to read you're going to shift many of these patches towards recognizing written words specializing for the alphabet but if you don't learn to read you're going to also use these patches of Cortex and maybe you're going to use them for objects or for faces so if we scan a literate person we see this patches respond to

words maybe a little bit to objects uh if we scan the very same patches in an illiterate person we see that they respond to faces and two objects they're not just sitting there uh doing nothing they are part of an architecture which is involved in visual recognition we slightly shift it by giving them this very particular stimulus the written word right uh in fact we are beginning to understand the system better and better by simulating how it's learning this is very very recent work that we're still doing uh we can take artificial neuronal networks I'm

sure you know that there are now wonderful uh systems that are inspired by the brain a little bit but they are of course part of AI and they take an image as input and they can give you an output which is the category of the image maybe it's a bowl or it's a dog or it's a table something like that and we we take exactly the same network and we recycle it to learn words as well so we add 1,000 output which are 1,000 words that the system has to recognize regardless of font of size

of place on the screen and when we do that we find that just like the brain these artificial networks will develop a subset of units bit like the visual word form area specialize for Words tune to words more than to anything else like faces and bodies and so on and now we can analyze them this is something we cannot do yet with our fmri we can look at single units we cannot look at single neurons in the human brain so easily if we look at single units we begin to understand how it works there are

units in the system that will tune to specific letters as well as to their positions so here for instance this is a unit that likes the letter a and it likes the letter A on the left side of the word at the beginning first or second letter we find another unit that likes the letter R towards the end of the word we think that there are hundreds or thousands of such neurons that are tuned jointly to the letter and to the position with some fuzziness so they start on the left and on the right and

in the middle of the words it's a bit more fuzzy okay and that's the neural code that we think is being built up by learning to read which allows you then to take a string of letters and immediately to know well the first letter is an a the last one has got to be an r in between there are some other letters and I know which word it is is it's like each word has a little barcode in the brain and that's the result of learning to read each word has a different set of neurons

that are being activated and that allows you to recognize which word it is right I'm sorry this is a little bit abstract I'll go back to concrete stuff but just to give you get a very concrete picture of what must be happening in your brain right now you're fixating on a given word well think that there's an assembly of neurons that is being activated and the nature of which neur are being activated it's like a little map that or barcode that tells you which word it is right invariantly for the size the font whether it's

uppercase or lower case whether it's handwritten or not these neurons have this invariance property okay part two what happens when children struggle to learn to read uh we can also make images of children who are struggling and that's the first thing we did actually we did exactly the same experiment as before flashing words faces houses checks same exact task what's beautiful about the star detection task is that even children who cannot read can do it right illiterates can do it so it's just perfectly fine task and when we do that you can see the disorganization

of the reading system in 9-year-old dyslexics compared to 9-year-old readers here notice the absence of the visual word form area of course anatomically it's there the neurons are there but it's not specialized so it's not responding when we present words more than when we are showing other types of pictures and um you might notice that the face system is also a bit disorganized here and that's still something that we're showing to understand but it's very significant there's more activation in normal readers to words in the left hemisphere and there's more activation to faces in the

right hemisphere in normal readers so it points to some disorganization of the visual system in dyslexia um of course you have to be very careful it's not clear what is the cause and what is the consequence here of this Lexia and I really want to stress this point right we are we are showing images it's a bit better than a lot of the literature because a lot of the literature is in adults and this is not in adults this is in nine-year-olds when they are struggling with learning to read but still the absence of the

usual World form area we think is more the consequence of not learning to read they are not specializing this area the cause could be elsewhere I'll come back to that but the cause could be in phology for instance and still you would not be developing this region properly but this is still a matter of research so uh we are doing much more work in this area this is a recent study where uh we scan both Chinese and French uh struggling readers 10 year olds this was a collaboration of shiaa fang and uh we had the

very similar Paradigm just the words were in Chinese for the Chinese children and we found that this reduced visual word form activation is a universal marker of Dyslexia even in Chinese so you see here the French subjects and lack of activation of this visual word form activ area lack of selective activation and the same for the Chinese control subjects versus Chinese dyslexia here missing activation in fact you can see the activation here dropping dramatically in uh this brain region in children were struggling to read so at the very least it's a biomarker we don't know

whether it's a cause or a consequence but it's a very clear marker of how well you can read in fact it's quite amazing just by scanning you can estimate about 50% of the variance in reading scores the fluency scores just by looking at the activation of this area that's that's how good we are in brain Imaging only half right so behavior is still the main way to monitor the development of reading and behavior is cheap in one minute of reading you can get you know an estimate of fluency and uh that's already better than brain

Imaging I want to stress that because some people say oh my kid is dyslexic I need a brain scan well no this is this is research this is not very useful I'm afraid uh but it's useful it's useful to understand how the system works but it's not yet useful for individual children but we see this rather remarkable fact which is that even in Chinese even in a very different writing systems the same circuit is still being used in fact There is almost no interaction and no detectable difference between reading in Chinese or in French and

the same areas are being uh under activated um I don't want to overstress the role of the visual word form which is here because there are also these other areas that are being under activated so it's a circuit and all of these regions correlate with reading scores across dyslexics and normal readers so they show a sort of broad correlation right this is the only place in the brain where we see a main effect of language is more activation in Chinese in these parietal regions but there is no interaction whatsoever which suggest that a very similar

um uh lack of activation or deactivation reduced activation is present in uh struggling readers throughout the world but this is a circuit right it's not just one area I've insisted on this area but the connection to these other language system is equally important and we're still trying to sort out in research what is the cause and what is the consequence of the struggling reading so let's talk a little bit about dyslexias I would really like to insist on the S there is not one dyslexia I think there are lots of ways in which the system

can go wrong and can give you uh deficit in learning to read so this is my very schematic view I I I had a more refined schema I decided not to show it here I'm really simplifying okay because the the cognitive science of reading is much more refined than that there's beautiful model by Max cart already for several years decades and we are using more refined models to try to understand but let me just give you a sort of summary in one slide that if you think of this system connecting Vision to language any of

these uh spots could be the result of an impairment could be the cause of an impairment right so certainly impaired phology is one of the Prime suspects for many children it's not hearing well enough the sounds of language and Mak being able to categorize them properly separate them from each other is certainly one of the most important precursors of reading and could be the cause of some impairments connection between letters and sounds you have to develop this connection if the connectivity is not there you're going to struggle to learn to read so impaired graphine Phim

conversion itself could be a cause of Dyslexia but we are also finding I mean in this lecture I insist a lot on our findings from the visual system we know now that the visual code for letters can be one of the causes of some dyslexias and even there there are several subtypes and we're collaborating with Professor n fredman in Tel Aviv who's been putting together diagnos tests of how to distinguish this we are now finding them in French she's finding them in all sorts of languages in Turkish in English in Hebrew um what are these

subtypes well there are two very important ones to understand one is letter position dyslexia these kids know very well their phes they know how letter sounds what their problem is to figure out the positions of letters so you give them form and they will say from okay I'm sure you've seen this in your classroom the trouble is in the middle of the word typically the first letter and the last letter are correct but in the middle of the word it's all fuzzy we all suffer a little bit from that if I show you the word

bagged you might think it's badge right uh I'm sorry I should have an illustration of that we all suffer from that but these kids seem to suffer more from that okay for them finding the positions of the middle letters is difficult they only make the sort of Errors so the interesting Diagnostics tools is that you can separate these types of errors and show their phology is okay their attention is okay and so on so forth they only make these sort of Errors right what's the right solution I just showed it to you is to enhance

the differences between the positions of letters so Nar found that one way to help them is simply to ask them to point with their finger from the left to the right of the word and make sure that they scan the word and that helps them enormously because it helps them separate the positions of the letters you can also space the letters and this has been Pro proven to be helpful for many children just put a little bit of a space maybe half a space between the letters so that they are more separated from each other

but these children will be helped but not all dxic children will help just these that have this uh letter position dyslexia I was telling you that there is another type which we call attentional dyslexia it's perhaps not the best name what do they do they have trouble separating the current word from the other words when you read you have this typically sentence right or text and you have to focus on one word at a time moving your eyes and when you focus on one word don't get distracted by the others well that's their problem um

when you show them form and pale they might say F and pole okay and that's the diagnosis test you give them words like this that can create errors like this and suddenly they make a lot of them notice is a very particular form of error the position of the letters is respected so very different from the previous one they will take the second letter from the other word and bring it into the second position of the current word or vice versa so they don't have a problem with letter position they have a problem with focusing

on the current word and how can you help them well Nama tried a lot of things but the very simple one is you cut a piece of cardboard right it's very cheap uh cut a piece of cardboard and and show them how they can slide it so that they only see one word at a time and suddenly they're good readers which shows that their phology is fine graphin phon conversion is fine they're just being distracted by the other letters right so uh I think we are getting much better in uh describing different subtypes of Dyslexia

and this is ongoing work there is no General agreement I I would think that um maybe G who is here might agree with me Professor Eden that phological um source is perhaps to seal the dominant view of Dyslexia but I have no doubt now and I've changed my mind a little bit based on the data I have no doubt that there's these other subtypes that are very different and we need to have careful tools to distinguish them and to rehabilitate them what dyslexia is not ah maybe many of you think that this is dyslexia a

child is writing with mirror letters this is the classical Orton view 1930s it's very old almost a century old now you have kids I I love this this one here this is an Italian kid who writing Theodor thio Ben Theodor loves you right it's it's going like this left and right right right to left Etc it's called bedon in in technical terms maybe it's described in this Museum it's how ancient Greek was written it's like a like you would plow a field that's what bedon means you would you would just switch directions at every end

of the field right well this child uh has no problem whatsoever he going to be a fine reader it's not dyslexia it's just a natural property of the system to have this mirror inversion so um very simple test was done by Cornell 1985 it's a bit of an anecdotal paper but I like it because it's so clear you uh ask a child to write their names next to a DOT so lissy and Magie wrote their names next to the dot but here there's a problem the dot is on the right of the page right so

what do they do well Maggie she these are two sisters meie is one year older she squeezes all of the letters together cleverly to put her name here lissy is one year younger no problem whatsoever I'll right from right right to left right it's very simple when when you do this test Cornell reported that 90% of children will ride from right to left at a critical age you can see it here around the age of five or six so it's not dyslexia it's a normal feature of the system I just wanted to point this out

it's actually a good uh indication of this recycling idea that I was opposing to you the neurons in this area have this property that they recognize mirror images left right inversions it's very important you have to recognize that this is the same person right and yet it's a completely different image you may not notice but on your retina it's a completely different image it's a problem for the brain the problem is solved by the very same neurons that we're going to recycle in order to learn to read so initially children see this as identical you

don't for you it's very different words because you've learned to read so when we learn to read we have to unlearn this mirror invariance and that creates difficulties for all children all children have difficulties with B's and D's and p's and q's but it's only if it stays for a very long time maybe after three years that maybe it's a sign of Dyslexia it still is not the cause it's more of the consequence of not learning to read properly right you have to unlearn the system so let's really put this aside for I would say

99.99% of children this is not the cause of Dyslexia there are some very rare cases where this seems to be a particular problem in their brains but it's very very rare we have not been able to find many cases where this is really a causal factor in dyslexia you mention very kindly my book reading in the brain I insist perhaps a little bit too much on these mirror errors in this book and now I would I I think I should do a second addition to revise it a little bit because we know now this is

not the Central problem or it's it's a problem for all kids you have to attend to it but it's a problem for all kids one way to attend to it is to make them right because when you write the gesture is is very different for writing a p and a q or B and a d so they begin to understand that it's really the 2D shape that counts and not the 3D object that would be inverted uh in three dimensions okay I will finish with a few Reflections about how we can use this science for

for the teaching of reading and I I'm sure you understood that I'm a strong advocate of what I think every scientist in the field uh believes that uh explicit phonic is the right way because this really corresponds to the how this circuit works you have to detect the letters connect them to the to the Sounds in order to begin to acquire reading um so the first thing I want you to remember is that there are completely different timelines for spoken language versus reading acquisition I told you that spoken language starts uh on day one it

actually starts in the womb a little bit with Pro um so it really starts extremely early and it's a very slow process you are learning about 10 words per day sometimes 20 more you can um but that still takes years and years and years to achieve knowledge of a language or two languages or three languages and the early years are already trial way before preschool actually we know now there is sort of critical period for the first language if you're not exposed to a first language you are in trouble uh and that needs to be

something like first year beginning of second year but if you don't have a language before that you are in real trouble you have trouble all your life basically um I say this because sometimes you have kids who are deaf right and they can acquire sign language that's fine sign language is a perfect language and then they become completely fluent and and normal or they can be um equipped with Cleo implant and then they can learn spoken language but if they have neither of these Solutions then they're in trouble and we see that they may have

long lifelong deficits it's really important so very slow growth of the spoken language system based on a specialized Network what do we do with reading completely different we insert around first grade this massive input of um decoding basically uh phonic and we it can go quite fast typically the first half of first grade you begin to decode words at the end of first grade you should be fluent in France we have these guidelines of 50 words per minute uh which every child should attain and of course it's not finished yet you're going to learn many

many more words and sentence structures through reading reading now becomes another way to learn language to learn vocabulary even though it's the same areas that are doing the job for spoken and written language they are going to acquire words sometimes from vision sometimes from the spoken modality and fuse it together right so it's very important that we provide this decoding input it's like a silver bullet that you put uh to uh suddenly enable this new input to language um I've tried to put together in several books this book is not translated in English by the

way I should do it but we're working on it uh it's a very small book just for teachers to have the sort of key ideas that you should have uh when you're trying to learn to read so we're trying to develop this route from Vision to phology and later to automatize it so it becomes faster and faster and you can go straighter to the Lexicon so here are some of the key ideas explicit teaching of graim to Phim conversion what is called phonics a systematic progression we want to use the statistics of language to facilitate

the acquisition we start with the most frequent letters the most frequent and most the most regular graphine to Phim conversions that makes it easier for children we need a very careful guidance of attention it's not trivial for children that they have to go from left to right it's not written anywhere in the stimulus and uh if you think of it it's one of the most uh clever inventions I was going to say cleverest but I I think it is not exist uh one of the most clever inventions the time of the spoken word word is

turned into the space of the written word time to space right it doesn't get more abstract than this that's why it's such an amazing invention it was not invented so many times right it's a very very clever invention the alphabet but it's not trivial for children at all so they have to be taught you start from the left you move to the right each of these letters or groups of letters going to turn into a sound and you're going to be able to hear it okay um we need Active Learning and that involves writing as

well as reading there's a lot of evidence that learning to write helps learning to read uh and the two go together and I'm surprised that I have to say this but I had to say to French teachers you don't write words that you cannot read right uh it's very simple but I I was surprised that there are preschools where they write entire sentences and the kid cannot read them that's very silly I think the two go together you teach the very same letters in writing and in in Reading in a systematic order but don't anticipate

one over the other one more recommendation you need to automatize it's not sufficient to recognize letters you need to automatize you need to practice and make it more parallel I'll return to that one word about decoding and comprehension because this is a constant misunderstanding teachers oppose decoding and comprehension if I'm doing one I'm not doing the other that's not true and that's not how cognitive scientists think about it uh the correct formula well first of all understanding what you read implies both right you need to decode the characters and you need to have the vocabulary

you need to understand the target language so think of the Rosetta Stone you need to know the hieroglyphs but even if you knew the hieroglyph you need to know the target aisan language if you don't you're stuck right you need both so the correct equation is this and that's a very old word from G and and T tner um written language comprehension is the product of decoding skills and spoken language comprehension so if one of them is zero this is finished written language comprehension is zero right you need to have strong decoding skills and you

need to have spoken language comprehension they go together and the evidence is very clear if you enhance decoding you enhance written language comprehension so the two go together but you need to do both and I told you this one is slow spoken language comprehension is very slow so many teachers say well I I T decoding and still the kids are not good readers well sure they you know if they are from if French is their second language or in your case English is their second language they don't have a lot of vocabulary at home you

know that there is this 30 million vocabulary Gap in preschool that some poor families don't have enough vocabulary the kids are simply not getting enough spoken input well that takes a long time to compensate this process is very slow this one can be fast so at least give them this right give them this six months 12 months Silver Bullet of decoding and then you will give them the ability to learn vocabulary through reading so vocabulary will develop if they read more and more right and you can help them enormously even if in their families um

the spoken language is not very present you can give them written vocabulary and it will come to land on the same brain areas so this will be your contribution ution to their language cont comprehension right this is very important to understand stop opposing comprehension and decoding they go together there are just two dimensions then I want to speak a little bit about the importance of consolidation because I spoke about the Early times of reading you're learning graph phon but every learning this is uh what I discuss also in my other book how we learn every

learning starts with explicit conscious effortful processing and as teacher you know that the kid is struggling right with the string of letters but any learning needs to be automatized we need to transfer the knowledge to non-conscious circuits and you need to free the resources of the prefrontal cortex which is here one third of the brain struggling to uh apply the rules of reading so in the case of reading is very clear if you T first year readers the number of letters is a strong determinant of reading speed so you see here the time takes to

read a single word when it's two letters three letters four five letters and first year readers are struggling they have this linear increase in uh reading times as a function of number of letters by the second year it's almost gone and by the third year it's gone and for all of you in this room you can read two letter words just as well as nine letter words but you're almost flat right in this curve it's quite amazing it means what does it mean it does it mean that we are whole language readers no I hope

you understood no we are still we are still processing the letters your brain is still processing every one of these letters just the brain is processing them all in parallel right instead of one by one it's able to take them all at once but it's still processing the letters and still converting them into Psalms and still at the same time finding this BARC code which allows you to recognize the whole word right so um automatization is absolutely essential not just in reading by the way but in all sorts of domains if you're still struggling if

all of your resources consciously are at the level of trying to decipher the word how are you going to do mathematics how are you going to focus on the message right right now you're reading this completely effortlessly and that's because you've automatized reading and you can focus on my message but if you were still struggling could you let's see if you could can you read this does it make sense I translated it into phonic English I hope all right well first of all do you have the feeling of relief When You Reach This last two

okay okay so now you see the difference between automatized and non- automatized reading the some of your kids in your classroom are like this right they still struggling um now one question could you understand the sentence it's of course it's po it's it's not so easy but it's it's it's difficult to understand because you are focusing your resources on decoding and so you cannot focus your resources on the comprehension of the syntax is a little bit complex and so on right so this is an example of what struggling kids are going through I want to

go back to my uh experiment where I was following kids so at the end of preschool you remember not much activation for reading in first grade there was this explosion of activity and what I didn't tell you is by the end of second grade most of the activity is gone except for this extremely efficient posterior Visual and phological circuit so all of these additional activation is the struggling it's the attention it's the working memory that you are you are have to put together in order to decipher uh decode words so of course we want to

go beyond that stage in this sense decoding is not sufficient you need automatize decoding but it's an indispensable step we cannot go with we cannot do without we have to go through it there's no way to go straight from this all the way to this right okay I will finish because I think my time is up just to tell you that uh I have uh become much more interested in developing applications of this for education in France um the minister of Education six years ago called me to create this scientific Council for education and we

put together a group of scientists we are uh 26 at the moment uh advising on all matters of education and in particular on reading and there are several things that uh I think uh we can do to help one thing that we did is we U of course put together resources and one of them is an evaluation system at at this moment every first grader and second grader in France is getting a battery of test bit like in the UK it's called the phonics check it's very interesting and so they are we are checking not

just phonics but knowledge of letters fluency and so on so forth so every teacher gets a picture of their class this goes back to the teachers very important every teachers gets a picture of you know my 20 kids or 15 kids this one this one this one what are their needs we can see of course National statistics like this one here we can see for instance that in sixth grade uh there are many kids that are not meeting the standards the setting of national standards itself is of course very important we're setting the same exact

goals in terms of reading speed fluency year after year for every kid regardless of their socioeconomic status and origin so that the teachers are encouraged to try to reach this limit now what do you do to reach this uh you know ideal level which at the end of first grade is 50 words per minute here is a sample of the reading manuals just a sample there's more than 35 reading manuals just to learn to read French in first grade it's terrible I think um sometimes the same publisher has a whole language method and a phonic

method just to make money uh they really don't care um and uh so what should we do we have statistics here there's a very interesting study that was done by Jerome doo in sociology just asking 10,000 teachers what are you using which reading method are you using and what is the type of method and we found out that most French sister still use there is no longer a whole language that that is gone but there is a mixed or balanced reading methods with a little bit there's a little bit of everything and most teachers select

these Balan reading methods and they still believe that root memorization of site words is has to be done basically there's no way to avoid it it's something that you need in order to read sentences and so on so forth so um the percentage of teachers who use S words most of them introduce between around 20 or 30 site words early on in uh their um reading and very small minority down here use these manuals you don't know them but these are purely phonics manual that have no site words no sentences to guess and so on

so forth who is right well we can use these National statistics to correlate what the teachers do which manuals they use and the success of the kids and the results are quite clear the synthetics phonics without site words get the best results so once again this is not a new message we always have the same message um you don't need the site words you can you can learn site words when they are introduced as graphemes and as fores right but you don't need to wrot memorize words it's the wrong way teach is the wrong uh

system basically so we found out that you could have better results with popular schools low socioeconomic status schools with a syllabic or phonics method uh com and worse results in favored schools just because they were using the wrong manual they were using uh this balanced uh reading methods the evidence is the same in every country it's been already mentioned 20 years ago more than 20 years ago in this country uh I think you know we have to just accept the evidence as it is but this is a bit more subtle evidence now we don't even

need site words at all um so we're putting out documents for teachers trainers authors and Publishers with manual guidelines about which manuals work better also what every first grade teacher should know we I don't think we need to know so much about the brain but we need to know just what works and it helps a little bit to visualize how it changes the brain right um we are now in the process of labeling these manuals so in the future it will not be possible for Publishers to just put out anything out there because the manuals

will be labeled and our scientific council is uh contributing to the definition of the criteria by which a manual should be labeled as uh compatible with the science of reading so uh these are things that we could still see uh we we published a note of alert by our scientific Council because just a year ago a year and a half ago you could still see in first grade in the first week of first grade please read this text kids all right okay yeah and uh so don't do this we could still see this which is

the most amazing I've ever seen can you read this the letters have been turned into triangles and and circles and squares so that this would favor guessing my children but nobody can read this not even an adult in this room right nobody this is just ridiculous but this was found in first week of our first grade for some for some children in France so it makes me very angry that we have this very busy inventions right we have other things of course we have all word guessing um we also have I don't know if you

have this in the US we have the phonics phonetic alphabet professional researchers use the phonetic alphabet to label all of the phones in all of the world's languages but even professionals like G and I we don't know this alphabet very well it's very hard so why would you do this to kids you know they already have one alphabet to learn why would they have to learn another especially when this one this is U okay okay how could you this is just this is beyond me I don't know why you would invent these things okay so

just one last thing but I think I will stop sometimes you know researchers are are out of touch with the classroom this is not the case in my lab and I'm very happy about that because we have Cassandra Po wkins and also collaboration with other colleagues we are actually developing software and methods and we are testing them and and and doing randomized control trials as much as we can so we've put together this So-Cal kalulu software method there's software and there's a method as well and games to put together in one piece in one manual

and one soft one piece of software all of the principles we think are useful for children so systematic pedagogical progression and we have nice AI tools now to take any language and automatically identify what are the most frequent letters what are the most frequent letter to sound correspondences so that automatically we can design an order for any language about what to teach we divide the lessons into steps there's this very nice tracing of the letters a bit like Montes send letters that you all know about which really helps children to understand the shapes that they

have to recognize we the software includes automatic evaluation of how automatized is reading this is what I was showing you earlier the effect of the number of letters on reading so the software can can do that and of course it's being tested and we have uh positive effect on children but is we used it more as a tool to understand which principles should be used that we at what time and so on in teaching so my conclusions first of all the brain is a highly structured organ there is considerable plasticity within an existing architecture which

is that which is the universal architecture of the human brain whether you're French or Chinese we all have the same brain and the same learning algorithms so for me it's a superc computer honestly I I consider the baby is the best superc computer on Earth it's completely unequal uh there's no similar machine even with today's AI is learning as fast and certainly not with the same amount of power the brain consumes 20 watts all right so currently uh Google is spending 20 megawatts easily to learn a single language model to teach a single language model

um so I think we have to respect this system feeded with the best possible input well-designed curriculum with explicit teaching progressively um you know feeding it with the inputs that it needs uh I think furthermore and more broadly education must be based on scientific principles there's a lot we have to learn together it's not finished of course it's not that Neuroscience has all the answers but we have to work together to figure out the brain's learning algorithms to experiment in classrooms to test what pedagogies work best and to accept the results because these results on

reading what I told you today the science of the brain is advancing but the science of what works in the classroom is at least 30 years old and more right so just accept the results phonic works best the whole language method is just inacceptable in today's world it has been rejected so many times and uh with that I will just conclude and thank you very much thank you thank you thank you so much I think we're going to move right over here to sit down and um thank you that was fabulous and fascinating um I'm

going to ask a few questions and then we're going to take questions from the audience so think about the questions that you have I'm sure you have many um and we'll get to those thank you very much oh thanks I'm sorry I went a little bit over time no it's okay I we all could have listened longer are we on are we good okay great um well um okay I want to ask with one of one of my start with one of my favorite questions as a journalist I know that every career has a story

so I'm curious why are you a neuroscientist why are you so interested in this oh that's that's a nice question I don't know I I was uhoh didn't go very far I was no I was always fascinating by the brain I was very good in math I could have been a mathematician I suppose I was training in math and so on but I was more intrigued by the errors that we were making in mathematics uh you know it's very frustrating when you realize that you got a bad grade because you made a stupid error somewhere

and so I I I I was always very interested in the in the sort of meta level of asking myself why am I making this error or why is this kid I was teaching a little bit with because I was good in math so I was teaching young adolescence and why don't they understand I understand what what's wrong you know so I developed this interest in the brain also as a young mathematician I was fascinated with AI already at the time so that was like 40 years ago I guess I started to try to program

computers to do intelligent things I realized how hard it was and I realized that there was a whole discipline of cognitive science and then later Neuroscience that could help to get ansers fascinating yeah I mean I think that's one of the hardest things I think there are many teachers in the room uh I love that your curiosity was spurred by teaching math and why don't they understand it I think that's one of the hardest things is once you become an expert at something to then understand what it's like to be a novice and I think

a lot of this reading stuff is is mixed up in that problem how difficult is absolutely true in fact the the whole language idea came from psychology it came from the very curves I showed you there this flat curve where there's no length effect right we recognize short words just as well as long words that gave the impression that this is how you read you just flash a word and you immediately recognize it but there was no awareness that in order to get there you have to get through these intermediate stages otherwise you never get

there at all so I I think it's what you are saying is very deep there is a constant illusion that because I know it's easy it shouldn't be difficult we could go straight there why you spend so much time losing time decoding right where we could go straight to comprehension well it doesn't work this way yeah I love the image of what it's like in first grade like there there's nothing nothing there and then all that effort and then it gets organized I'm currently tutoring some kids and they're in the middle of that there's all

that effort going on and it's so it's like nails on a chalkboard sometimes to listen to them like sounding out those words but you can see in those images like you've got to go through that to get to get from point A to point C you've got to go through that point B yeah and still we can help right we can facilitate this process it's not true that you just have to struggle for a very long time we can make this period much shorter uh by proper teaching methods and by also rehearsal and so on

so there's a lot more about the science of learning which goes beyond reading of course but which is General um in in my book on how we learn I insist a lot on the role of daily practice and sleep um these are very essential ingredients uh we the brain consolidates the learning during sleep so as a parent you should take advantage of every day and every night and even every siest you say a nap every nap right is a way to consolidate The Learning in this way you make it much easier I'm a big believer

in the 20-minute afternoon nap so things are getting Consolidated right there so you talked I appreciate what you talked about in terms of what you're doing in France to try to connect this and all around the world to the to the world of education and there's certainly a lot of discussion that's why I think we have so many people here today about the so-called science of reading in American schools and a lot of people are concerned about how uh their kids are being taught to read so do you have you had a similar sort of

debate I think part of the problem here is there's been a tremendous amount of resistance there's been a tremendous amount of dismissing of some of the basic scientific findings is it similar in France and why yes uh there there is quite a lot of resistance um it depends very much on who you meet right there are many teachers when I give a talk like this they are enthusiastic they think they they they come to me and they say why wasn't I told this you know in teacher training school so there is that but there's also

a strong resistance very often is from the trainers not from the teachers but from the trainers of teachers who have been teaching for 20 years in a certain way and suddenly they have to change they have to accept that maybe you know it was not perfect what they were saying and that's very difficult um and then there is purely ideological reaction uh what has what has Neuroscience Got to Do With It uh these people want to reduce our kids to Little Machines they're going to put electrodes in schools you have all sort of fantasms that

are extraordinary I think and you know we're not thinking of scanning every child but um and I I have children and Grand children like everybody else and I care very much about how quickly they learn and whether we can help them and so on so this is I think this is perhaps the hardest is all of these fantasms that we have to dissipate which are very uh you know part of people's beliefs that we have to change yeah like it's exactly the same in France as it is here yeah we come together when we show

results I think so when you show in a classroom like with this calulu method we go to classrooms we convince teachers to try it and when they see that um every child is starting to read by January uh and at the end of the year there's maybe one child who is below 50 words per minute then then they sort of realize you know and we have many testimonies from teachers like that that say oh I I thought it was completely normal for five kids out of 20 not to read at the end of first grade

and now we realize no it's not it's not right at all it it shouldn't be this way it doesn't have to be this way so so that's that's where we all come together I think with the sort of conrete evidence from the teachers themselves because they are doing the teaching right I may have given the impression is that the the manual is doing the teaching that's not true there is this the teacher is doing the teaching the manual is only some support right for helping them to derive a proper progression a structural progression so maybe

that's something that you have also here in the US one source of opposition is a very deep belief in the teachers that they have to construct a curriculum if I if if I have not constructed my own curriculum day after day I'm not a good teacher and and that's strange thing you know there for me a good teacher is one that has made the kid progress the best you know to the to the most of their abilities regardless of where the curriculum comes from so we have we have these teachers in France who actually complain

to have a lot of work because every week they have to create the materials for the next week well they don't have to they they in fact it's very hard to design a good curriculum it takes years of effort to put it together in the proper order to try it feedback so they don't have to do that but that's one of the strong beliefs uh that they have that is also propagated by teacher training schools that you're not a good teacher if you don't design your own curriculum that's that's another big belief that has to

be changed which is quite hard I think I've got one more question and then we're going to go to the audience so I guess get your questions ready um so yeah I think what you described about what was going on in France is very similar to what's going on in the United States and you I think there was a little bit of an audible reaction in the room when you showed the testing that all kids in France are doing and the results that teachers are getting and you told me in the Green Room they're getting

them like in real time like teachers are getting this information and one or two weeks after one or two weeks after they get it so it's like progress monitoring for the nation so I'm curious since you've now working with uh Educators and policy makers it sounds like to try to bring some of this into education if you have any thoughts or advice for us in the United States in terms of policy there's a lot of policymaking going on right now we have an even bigger country and we sort of divide everything we don't do anything

as an as an entire nation really any thoughts well I don't know the local situation well enough I I I've been reading that Mississippi for instance been doing incredibly well lately I I hope it's true but what I was reading was really uh amazingly uh positive I'm also reading that New York is uh changing dramatically and uh in this case with changing teacher training uh which is you know uh one thing is that there is big money involved as well and so you have to change that right that's that's very very difficult there are people

who resist just because of the money involved uh so I I I think that these what I see in Mississippi and in New York what I read from the New York Times uh seems to be going in the right direction why can't you do that together for all states I don't know yeah good question okay do we have some questions from the audience I think there's a microphone um there's a woman going around Caitlyn's going around with the microphone see someone right there hi thank you so much fascinating and what a privilege to hear you

um I'm a dyslexia therapist and over the years I've had families come to me with children with a diagnosis of Dyslexia who have just spent years and thousands of dollars on vision therapy and say oh it's working but we just want to see if something else can help too and I said well let's try multisensory structured language syllable based lay and see how it goes and it always in fact it has worked every time in helping these children to read I always bite my tongue now when people ask me about vision therapy and I wondered

if you you do if you understand that's what we call it here um what do you call vision therapy exactly um maybe in one I'm not an expert but what I hear from parents is um a few times a week where you where the child goes and they follow um to strength tracking to strengthen their eye muscles or something to sort of uh as intervention for dyslexia ouch yeah I I'm sorry yeah the more I I I try to enter a little bit the field of Dyslexia the more I am surprised by the diversity of

just crazy it's just just no seriously so well okay so in France we have a company now that sells glasses for dyslexics which are blinking very fast and that's the cure for dyslexia right they think it's in the retina I mean we have we have every evidence in the world that it's not in the retina it's not in the muscles it's in the brain in different brain areas there's probably different subtypes but you know it's really a learning issue I I will not say that there's nothing in the eyes in the eye movements this is

actually a research question for me I uh for facility in all of the experien that we have done here we we avoid eye movements because eyes are creating artifacts in our scanners and so on so that's a little bit of a critique there is an eye movement component to reading which I've not described here today you have to learn to move your eyes very quickly to the next word and then to the next word and of course if you're not very precised with that you're not going to be a very fluent reader at least in

the late stages but at least initially the single word is really the good Paradigm here and um this has nothing to do with eye muscles or whatever right so no I I I I would advocate for being reasonable there is a diversity of dyslexias but they don't involve the muscles they don't involve the retina in they require focused cognitive training um and indeed as you say graim Phim practice is some of the most efficient with some of the additional tools that I described focusing on the letter position and focusing on the kids are being re

interference from the peripheral words that's that seems to be key another question uh I don't know there we go okay uh on your research thank you very much for this by the way you're showed as they are learning to read the facial recognition also increases does that mean they can find more details and faces they can tell them apart and if there is a way to increase facial recognition that also helps their reading is there do you do any from from the reverse like helping with facial recognition helps them with processing of language or uh

that's a great uh question but we don't know so much it's a bit of a mystery the relation between reading and faces it was a completely happen stance Discovery we but what we found was more competition these areas that specialize for letters cannot specialize for faces at the same time so there's a little bit of cortical competition um we so there is a tension between competition for the surface of the cortex but also enhancements that generalize across uh different categories and we found a little bit of both so I'm sorry I cannot give you a

very very precise answer here for instance being a reader helps you to become a very good uh in Vision precise Vision vision of very small letters it's not by chance that we start with large letters and we go to smaller and smaller ones I think because we're progressively training I think but that's a hypothesis more and more early areas of the cortex to discriminate these very fine shapes and that has some generalization not just to faces but to any shape actually so we have enhancement in early visual cortex for any shapes yeah um still I

think so I think what you are there is one bad idea in what you are saying is if we train faces would we enhance reading no I don't think so you have to train reading um there is also a very frequent mistake that we make by we hope for far transfer far generalization um we are actually doing an experiment in another domain right now in the lab we have kids that have music training and the hope is that this will generalize to math for instance well at the moment we're not finding that uh and in

the literature there's very very little evidence for that because far transfer is unlikely if you are not teaching pre precisely the content that you need to teach so you need to teach mathematics if you want the kids to know their mathematics you need to teach reading if you want them to be good readers that's that's the just a very simple rule yeah another question oh here uh can we get a mic up here oh sorry all right we'll go here and then we'll go back there hello thank you very much for today um I wanted

to ask a question I'm going just like therapist as well and work with students with dyslexia I've recently seen students who read actually very fluently but are having great difficulty with spelling um I'm assuming that their finding skills aren't as aren't as good as what might be and I try to work with that but do you have any suggestions have you looked at that idea whether spelling is really the issue and there's a whole category of uh deficits called dis graphia dis orog graphia uh which are different they're really different and um yeah in French

it's particularly a problem because French is relatively regular from the spelling to sound Direction but in the opposite direction from the sound to the spelling is very difficult there are many many irregularities um so you need additional training in that specifically I think but the the areas at least I don't know very well I don't think there is as much research at all on on graphia graphics writing as opposed to reading it's much more difficult also to study in the scanner so it's not very well studied thank you on the back hi hello I'd like

to ask um what changes in Reading acquisition occur uh maybe in children who are blind if there is less activation in the visual word form area and more in the sensory uh motor areas of the parietal lobe to accommodate for Braille yeah that's a wonderful question it turns out that when you read in Braille you still activate visual word for area and this is really formidable finding and S also a challenge right so somehow the information about Braille is getting to that area probably in a top- down manner because now this area is not being

inated by visual input but still it plays an important role yeah we don't fully know whether it plays a role in the recognition itself or maybe you're all in the higher level functions of language the the study of the blind is really super interesting for us we know that the the blind especially early blind are going to use the Royal cortex for something else for completely different functions so it's another evidence of recycling um and different bits of the occipital cortex will care about language and other bits will care about mathematics there's a lot of

mathematics activation in the occipital lobe in the in the blind um I speak a little bit about that in the in my book how we learn learned but we're still doing experiments I think it's unknown what do they do with these brain areas and how is the information getting there if it's not getting from the retina this is not really well known yeah I think we might have time I um have failed to wear a watch how much time do we have one minute oh great let's do one more question does someone have the microphone

in their hand right now or um I'm G okay we got a question on the back so I'm hoping that I can articulate this question because it refers to some of the the visual images that you were giving earlier but the visual word form portion of the brain if the person doesn't learn to read it typically that area is recycled for tools or for um uh other forms houses and faces but it in the dyslexic brain if if that area is not being used for reading and other pathways are found do we see that that

they have abilities that are particular to dyslexia because they're using that portion of their brain for for houses and faces and and words and then that they're using other portions of their brains for decoding and reading are we seeing those kind of differentiations that kind of characterize dyslexia so not specifically in this area uh and in general in the literature there is a small literature on whether the dyslexic children may have advantages um but uh I don't think that this is a very stable literature and I don't think that there is any very convincing demonstration

um one suggestion has been this mirroring uh but maybe it has nothing to do with dyslexia just if you're not unlearning mirror image recognition maybe you are able to see some mirror images a little bit better and we had some evidence for that in our alliterate subjects they they were suffering a little bit less in a task where you have to recognize mirror images and there's there is one paper that suggest this could be useful for scientists to recognize graphs this very very speculative uh and I think very very unstable scientific literature the the bulk

of the finding is that no this children are struggling and it doesn't look like they have very strong advantages maybe they have the advantage that if they surmount the difficulty they are showing character they are showing effort they're understanding the value of effort and maybe this has positive consequences later on in their lives Winston Churchill is a famous dyslexic who got the Nobel Prize right and he's an amazing figure but I don't think we should think of dyslexics as especially gifted in something else or in it's just kids that require a little bit of help

in a specific domain thank you well thank you all for coming and join me and thanking Dr danan coming from France thank you so much [Applause]