Neurociencia cognitiva demuestra que mayor parte de nuestra cognición es inconsciente. Transcript: Speaker 1 I think this is a key finding from a lot of experiments. In fact, most of the computation in the brain is unconscious, whether it is face recognition, word recognition, understanding the meaning of a sentence. This is something was and result is conscious, but the process itself is not. And by making experiments using this subliminal stem, and I were able to show how far non-conscious stimulus can go in the brain without being detected by the subject. So we could flash pictures. In my case, it was mostly pictures of words, written words, and show that it could go all the way, for instance, to the amygdala, where the meaning is accessed, and subjects have a notion Of whether a word is emotional or not. For instance, if you flash the word rape, they will strongly activate the amygdala, much more than for a neutral word. And they will do so even when the word is totally impossible to detect. So it’s just one example. We have many, many others showing that consciousness is already the tip of the iceberg, so near small part of our mental life. (Time 0:06:51)
2min amígdala cognición evidencia experimento inconsciente
2min evidencia experimento cognición
El aprendizaje se define como la construcción de modelos mentales. Transcript: Speaker 1 Learning in my opinion should be defined as the construction of a mental model. So it’s not the trivial thing to think that in our brain we host an enormous variety of models of the external world. We can close our eyes, for instance, and think about the spelling of words or think about the visit of our apartment and how things are laid over there in our apartment. We have in our brain internal models that we have captured, so experienced, and they are so rich that they allow us to generate almost an entire virtual reality. This is what happens in our dreams. We are immobile, there is no input, and yet the brain generates an entire reality. So this is what learning is about. It’s about adjusting the parameters of an internal model so it fits better with some aspect of the external world. That’s of course a very broad definition and we have to think that there are probably thousands of mental models, including some that are extremely abstract, like in mathematics. We are learning about equations in our they are organized, but also some very concrete ones. Like I learn the parameters of my body and how much force and aim I must put in order to reach a particular object, for instance. (Time 0:10:23)
2min aprendizaje definición modelos representación
2min aprendizaje definición modelos representación
Hasta los animales más simples modelan internamente su entorno. Transcript: Speaker 1 I think even in the case of habituation, I mean it’s very simple, but there is a variable inside the animal that corresponds to the amount of concentration of substance. And so if the substance has been high concentration for a long time, and there’s an internal variable that has habituated inside the animal, then that reflects the presence of the substance In the external world. So I think even in this case, even though it’s a minimal model with very few parameters, that somehow the animal is carrying with him and it can be even a very simple animal like a nematode Warm, the animal is carrying some knowledge of the external world. (Time 0:12:06)
2min animales modelos representación
Homo docens The unique niche of humans is the ability to teach and learn, making them a species that teaches itself. This distinguishes humans from other animals, whose niches are based on ecological factors. Transcript: Speaker 1 I claim that we should be called homodocans, the species that teaches and that learns, the teaches itself because I think this is the definition of our niche, other animals have a particular Ecological niche, but our niche is the ability to learn. (Time 0:13:14)
Homo docens
aprendizaje enseñanza homo_sapiens 2min
El aprendizaje humano se construye sobre conocimiento innato. Transcript: Speaker 1 What cognitive science has discovered is that learning is made possible because they are innate domains of knowledge, core knowledge that we capitalize on in order to go beyond in Learning. So, in the book, for instance, I described these remarkable experiments not just about number. This was also a topic of a previous book of mine, the number sense, but also about probabilities in babies and babies have a sense of probability. If you can believe it, they know how to anticipate events that have a certain probability of occurring. If they see three balls that are blue and another one which is yellow and they are shaken in an earl and then one of them comes out, they think it’s much more likely to be the blue one and they Are surprised if they see the yellow one come out because this was less likely. Experiments show that the amount of looking, the looking time is directly related to the probability of the event or the improbability, the less likely it is the longer they look. And what experiments have shown is that the longer they look, the more they learn, it’s really part of the learning algorithm. It’s a device for yenting towards errors, towards things that you don’t yet know and that you have to learn. So this is all starting in the baby and even the few days of age. (Time 0:15:53)
2min aprendizaje evidencia experimento homo_sapiens probabilidad
aprendizaje homo_sapiens evidencia experimento 2min probabilidad
Definición de algoritmos bayesianos. Transcript: Speaker 1 Important. Essentially Bayesian algorithms provide a sort of optimal way to learn. If you know enough about a domain, you cannot do better than by using this Bayesian statistician algorithms. They essentially say exactly how you should combine your higher knowledge of a field and new evidence that you are getting from the environment. And there is quite a lot of evidence that that’s what babies do. In a sense, they act as scientists, the statisticians. They probe the external words for experiments. They look at the result and interpret what they’ve seen in external world in statistical manner as probabilistic evidence in favor of this or that model. So I follow Alisson Gopnik in thinking that the baby is a scientist in the crib. A little scientist who already reasons very rationally about the external word. Essentially Bayesian statistics are a big word to say that the baby is acting as a rational being, trying to draw optimal inferences from the external world. (Time 0:17:30)
2min bayesiano ciencia definición homo_sapiens infancia predicción
bayesiano predicción infancia definición homo_sapiens ciencia 2min
El cerebro es un predictor, está constantemente anticipando lo que va a suceder. Transcript: Speaker 1 We are not very good when probabilities are conveyed by decimal numbers. So if I tell you there’s a 0.001 probability, you don’t understand that very well. But if I tell you about integers, I’ll tell you out of 10,000 cases, there was only one where it came out this way, you’re much better. And in fact, there is a growing field of Bayesian statistics applied to the adult brain showing that really that’s what our brains do. Of course, it’s an approximation. It’s not perfect calculation. But our brains are actually pretty good at estimating the probabilities of events. And our brain is a predictor. The brain constantly tries to anticipate and what’s going to happen generates a prediction, which is probabilistic, and compares then what it gets with that prediction. So that’s a very important set of theories called predictive coding that I think capture a lot about what the baby is doing already from the start. In-nate categories provide some predictions. There’s, of course, still a lot to learn on top of this innate core knowledge. And so we learn by detecting the error between what we predicted and what we got. (Time 0:18:48)
2min aprendizaje cerebro definición error predicción
definición predicción aprendizaje 2min
Por qué los cerebros humanos son más potentes que la inteligencia artificial. Transcript: Speaker 1 So I am impressed by these progress, but I think there are still ways in which the human brain is much better. First of all, we require much fewer data. If you think of a baby acquiring language, the baby is actually getting very little data about the target language. And yet it’s able to converge, in fact, better than any machine we know so far on the regularities, the grammar, the whole and the lexicon of the language. We can prove that it is using very few words, sometimes just a single instance of a single word. It’s using all of the appropriate information about the context and in particular the social context in a way that machines cannot do so far. So it’s in the sentence reading the mind of the speaker. This social component is another aspect which is completely missing from machines. We don’t just learn, we educate ourselves. So we have a theory of mind that tells us that the teacher in front of us is actually trying to help us by giving us examples that are appropriate for us to learn. By using this sort of information, we can go much further and much faster. Again, experiment show that babies do that. They attend precisely to the attention of the adult in front of them and they use that in order to learn. And I want to mention the third thing in which the brain is superior to other machines. This is the ability to abstract information in a symbolic form. If you look at neural networks, artificial neural networks these days, they learn a lot, but they have no idea what they have learned. They are totally unable to formulate it as a symbolic formula like scientific expression. And that means that they’re also totally unable to transmit it to another machine. But we do that routinely. We synthesize the information in a conscious form in a way that we can speak about. And this is remarkable. We are perhaps not as good as learning Go. The machines are better than us these days. But when we learn Go, we can speak about it. We can say, you know, you shouldn’t play beyond the fourth row until some moment in the game. And this helps us to learn. We are a species that was able to bootstrap itself because we share the information. This ability to share the information through language, to formulate it explicitly, sometimes in the form of equations is absolutely suntore and it’s not imitated by machines at The moment. (Time 0:20:25)
aprendizaje cognición enseñanza homo_sapiens IA símbolos teoría_de_la_mente
teoría_de_la_mente símbolos ia aprendizaje cognición enseñanza homo_sapiens
La plasticidad cerebral es literal y tiene periodos críticos. Transcript: Speaker 1 And this is of course an important chapter in the book. I try to describe what we know, but plasticity about how synapses modify themselves and how this eventually becomes the substrate of learning. It’s an area that is still moving, but we really begin to know a lot and I think it’s useful to know. So for instance, did you know that when you look at a little baby for just one second, probably something like two or three million synapses have been changing, either disconnecting Or reconnecting. The baby is until the age of about one or two, they have many more synapses than we do. There is an overproduction of synapses, and that probably explains why they are so extraordinary plastic. Then the number of synapses will decrease. We select, we pwn the synaptic trees of the neurons until we reach adulthood where we have a sort of pwn-down system that is much more efficient, but less able to learn. And another thing which I found as I was reading about this very intriguing is that we begin to understand why we lose plasticity. First of all, we’ve lost some synapses, but second, the neurons are also surrounding themselves, the pyramidal neurons with a sort of matrix of rigid molecules that prevent synapses From being mazed. It’s like the neurons are in a cage. They cannot move anymore. The child’s brain is a much more liquid state. It has a lot of water. There’s a lot of movable parts that are still growing. At the other state, it’s a much more rigid system that physically cannot move. So when you think of brain plasticity, you have to think it completely literally. They are moving elements in your brain. In particular, on the dendrites of the neurons, they are little spines, minuscule ones, 10,000 bernions that move around. They retract, they contract, they multiply, and that’s the substrate of learning. (Time 0:24:42)
2min aprendizaje cerebro homo_sapiens neuronas plasticidad poda sinapsis
La plasticidad tiene distintos dominios, con sus respectivos períodos sensibles. Transcript: Speaker 1 The first thing is to recognize that there are sensitive periods and that these depend very much on the domain. There are some domains where really plasticity is extremely early and then closes down. One case in point is phonological learning. When we acquire the sounds of language, this goes on in the first year of life. The baby has not even started to speak. But he has already converged onto the categories of vowels and consonants. Vowels about six months, consonants after about 12 months. At this point, you are no longer hearing the differences between phonemes that are not useful in your language. So Japanese speakers no longer hear the difference between R and L. It can still reverse for a few years and that’s also a subject of research, not fully known. But let’s say after the age of nine or 10, it begins to freeze completely and cannot reverse anymore. The same is true for grammar. The abutiful experiment is one experiment with millions of subjects. It was run off the internet by Steve Pinker and his colleagues in Harvard. And what they found was that there is a slowly decreasing curve for the ability to acquire your grammar and then it collapses dramatically. So if you are trying to learn a second language, you should really learn that before puberty because after that, it’s really much, much more difficult. But as you say, there are other domains like we can learn new words. So the lexicon is not a big problem. We remain plastic. We don’t quite understand why some brain areas remain plastics and others lose (Time 0:26:44)
2min aprendizaje cerebro desarrollo lenguaje plasticidad
La hipótesis del reciclaje neuronal. Transcript: Speaker 1 It’s a very simple idea. We are the only primate species that is able to learn completely new tricks. One of them is reading the use of our eyes in order to convey information that normally should come through spoken language. There is mathematics, manipulation of symbols in order to do calculations in very abstract domains. So my claim is that when we acquire these domains, we have to recycle to reorient to repurpose circuits that had evolved for another purpose. Maybe a slightly related one. So in the case of reading, we understand a lot about this. We know that some visual circuits of the brain that used to be able to recognize objects and faces, reorient in order to recognize the shapes of words and the strings of letters. And they become tuned to the particular letters that we have learned to read. So we have seen this occurring. We did some recent experiments, for instance, where we scanned some children several times as they were acquiring reading. And so we could see their brain change. And this part of the object recognition circuit is being reoriented towards the acquisition of letter shapes. So I think this is a general phenomenon. We have in our brain a number of circuits that are partially pre-wired, but because of plasticity, they can be reoriented to another activity. And for the teacher, I think this is important. It means that we never start with a tabular hazard. The brain is never a tabular hazard, a blank slate. We have some initial abilities, and we have to grow something new out of it. (Time 0:28:19)
2min aprendizaje cerebro circuitos definición homo_sapiens lectura plasticidad tabula_rasa
2min aprendizaje definición homo_sapiens
El periodo crítico para la literalidad es en la infancia. A adultos les cuesta mucho más. Transcript: Speaker 2 This has implications, I guess, if somebody is trying to learn how to read as an adult. Speaker 1 Well, unfortunately, what we are finding is that it’s much, much more difficult. So what you have to imagine is the ventral part of the brain. There is a sheet of cortex. It’s covered with specialized regions. There is one region in particular that’s well-studied, which is specializing for faces. But there are other regions that specialize for words, some other specialized for numbers. The specialization can be amazing. So it’s not the same area for letters and for numbers, because they play different purposes. And these areas have to be laid down early on, probably before puberty again. Trying to learn as an adult, well, you use whatever you’ve learned when you were young, but you are not going to be able to completely reorient these patches of cortex. You are going to have to deal with whatever state they are in at the moment. So I think that explains the difficulty. As adults that were illiterate, they have the hardest time in the world just figuring out that the B and the D are different. (Time 0:31:15)
2min adultos aprendizaje desarrollo infancia lectura
Primer pilar del aprendizaje: Atención. Transcript: Speaker 1 So attention, the first one. The brain is a filter. At the highest level of processing in the brain, you only can process a small amount of information, perhaps one information at a time at the highest level, at the conscious level. So it means that there is a gigantic filter that takes the million of bits of information that reaches our retina and decides that only a small part is important. If attention is misdirected, you’re not learning. There is very little learning in the absence of attention. So we need to learn to attend. We need to focus. We need to avoid destruction. On the other hand, it’s a very helpful thing. If you attend, you dramatically amplify information. Which is really on the corner of your eye and somewhere on the retina can become dramatically amplified and access this global workspace where it becomes the focus of conscious attention. So my claim is that the teacher’s most important talent is to direct children’s attention to the appropriate level. They must decide what is important, what is not. And this translates into extremely practical information. For instance, there is experimental research, which I described in the book that says that you should not accidentally decorate your classroom. That the classroom, which is too decorated, distracts children. Lots of things that you could attend, you get distracted. The same of course applies to the phone or to tablet computer. You want focus, not a distraction. So that’s the first pillar. It’s a big one. There really is very little learning in the absence of attention. So when a child tells you, for instance, I don’t see, this is something that happens in mathematics. Do you see it? No, I don’t see it. It could be very abstract concept. What does it mean? It’s very literal. The children are able to attend to the appropriate concept. And therefore, they are simply blind. We can be completely blind in the absence of attention. I think you all know about the invisible gorilla experiment. You can have a gorilla in front of you and not see it because you’re not attending. So I think it’s exactly the same in schools. Some children are completely honest when they say, I don’t see, even though it’s in front of their eyes, because they haven’t learned how to focus their attention on the proper level. And we need to guide them. (Time 0:35:12)
aprendizaje atención cognición
cognición aprendizaje atención
Segundo pilar del aprendizaje: Involucramiento activo. Transcript: Speaker 1 So active engagement, that’s pillar number two. What do I mean by that? I don’t mean that the child must be pedaling or doing physical exercise. It’s been one of the misinterpretations. What I mean is that the child’s brain must be actively engaged. Acting as a scientist, we started with those, formulating hypothesis, performing experiments, mental experiment, projecting actively on the external world hypothesis and selecting Among them as opposed to being passive. So we know that there is very little learning in a passive organism. If you just bombard a child with information passively, not much is learned. For instance, there are beautiful experiments. Actually you put an object in front of the child and at the same time, there’s a loudspeaker in the room that gives the name of the object, zero learning. There is a perfect correlation between the name and the object. There’s essentially no learning whatsoever. There is learning when the child is engaged, typically by another person with whom he shares attention and when there is curiosity. So I speak a lot about curiosity in the book. I think it’s one of the major systems that we need to study at the brain level. We begin to know a little bit about it. The human brain in particular has a motivation system, which is not just driven by food or by sex. It’s the dopamine system, of course, which is universal in mammals at least. But in our case, it’s driven by the urge to learn. We have a system that will give us a little discharge of dopamine and reward when we can explore a new domain, when we can ask a new question and get the answer. So curiosity is a fundamental property of the learning algorithm. And we need to trigger it and to foster it. We need to try to avoid to kill it. Of course, unfortunately, school can kill curiosity if it’s the impression that the teacher knows everything and all I have to do is sit there. So this is the idea of the child as a curious, playful individual who actively engages in the learning. So again, it translates into very direct applications. There is beautiful research in the classroom, all sorts of research actually showing that you shouldn’t do a magistrate lecture, but it’s much better to have the child actively interacting With the teacher, for instance, being able to ask questions. There are some teachers that say you should spend half of the lesson with a question from the learner. There are many ways, of course, to actively engage the child, but that’s a general idea. (Time 0:37:26)
aprendizaje esfuerzo motivación
Tercer pilar del aprendizaje: Feedback sobre el error. Transcript: Speaker 1 But you want the third pillar error feedback in order to select the hypothesis. In order to learn properly, we want to generate a hypothesis and immediately know whether it’s true or false, whether it works or not, whether it fits the world or not. So the brain makes a prediction and then must receive an error signal that tells it whether this prediction is correct or not. What we found at the brain level is that a lot of the messages that are exchanged between binary as error signals. I anticipated something about music and then suddenly I hear a sour note, it’s not in the proper place and I have to correct my mental model of the melody. I didn’t have the proper model of the melody. So this is something also that the teachers must know. Send precise, informative error feedback signals. There’s nothing to do with punishment. It has nothing to do with grades, which are summary statements that are not informative enough. It is a precise feedback about what went right and what went wrong and how you can correct. (Time 0:40:50)
aprendizaje feedback interés predicción
Para aprender, es más efectivo probarse que estudiar. Transcript: Speaker 1 So there is in fact, beautiful experimental research again on this. So suppose you have two hours to learn something. Should you just spend the two hours revising in the textbook, studying the textbook, maybe you put some color on it, you know? Or should you have some periods where you study and some periods where you test yourself. And it’s not completely trivial because when you test yourself, of course, you don’t study. So you lose time in a certain sense. So when the experiment was done, the experimenters asked teachers what were the predictions and they all say, well, you should study. You know, the more you study, the better you are. Well, it’s not true. What is true is that if you test yourself, you are better. Why? Because when you test yourself, you get this error feedback. You know that there are some items that were right and you don’t need to revise them again, but you know also that there are some that were wrong. It’s important to test yourself not immediately, but after a little period, because when you just studied, the information is still in your working memory or in your recent memory. So you have the impression of knowing. But this is not the right memory system. It’s not the one that you are going to need in the long run the next day. So if you wait just five minutes or 10 minutes and you test yourself, you realize that in fact, you don’t know. And then you can correct yourself. You get the error feedback that allows you to correct your mental model. I try to explain to teachers that this has nothing to do with bad grades, that there’s something to avoid which is punishment. It’s very negative effect on learning. We all freeze and the learning algorithm stops if we are under stress on the negative feedback. So you want to have just a non-punitive assessment of what was right and what was wrong. (Time 0:42:11)
2min aprendizaje ejercicio esfuerzo estudio evaluación evidencia experimento feedback
2min aprendizaje evidencia experimento
Cuarto pilar del aprendizaje: Consolidación (a través del sueño) Transcript: Speaker 1 So this is leading actually to our fourth pillar, which is consolidation. It’s not enough to just learn. You need to consolidate. You need to automatize. Learning must be transferred from the conscious compartment, the conscious global workspace to a non-conscious compartment. This is absolutely essential because you need to free the resources of the cortex, in particular, the prefrontal cortex, for other learning. So when you begin to learn something, your mental resources are absolutely stretched. Maybe I can give an example. I remember very well when I started to learn to drive a car. And maybe you remember as well. It’s a catastrophe, my God. And of course, in Europe, we have to learn to drive with a stick shift. So there is also the hands are doing a lot of things as well as the feet and you have to watch in the mirror and the guy next to you is speaking to you. And it’s horrible. The control cortex is completely overloaded. Just wait a few hours, few lessons, 20 lessons perhaps, and you come back and it’s completely different. The learning of how to move the feet and the hands become automatized and you feel much more at ease. You’ve learned essentially to automatize a lot of activities. And then you can think about something else like listening to the radio perhaps. So it’s just an example to show you are important it is to automatize. What is fantastic is that we begin to know how it works. And one important aspects of how it works is sleep. Sleep is a key moment of learning. This is absolutely non trivial. I think we all think that when we sleep, we just trust, but that’s not true. Our brain doesn’t rest. Our brain learns very actively. And we begin to see it in brain imaging also at the neural level. (Time 0:44:19)
2min aprendizaje sueño
Evidencias sobre la importancia del sueño en la consolidación del aprendizaje. Transcript: Speaker 1 Sleep is a key moment of learning. This is absolutely non trivial. I think we all think that when we sleep, we just trust, but that’s not true. Our brain doesn’t rest. Our brain learns very actively. And we begin to see it in brain imaging also at the neural level. So when we sleep, what has been discovered? Well, first of all, the psychologist made a discovery. The psychologist discovers that you could learn something at Nozia during the day, maybe for three or four hours, maybe it’s a video game, for instance. And there’s a point where you don’t get better. You reach an asymptote and you just stay there. You go to sleep and the next morning you haven’t practiced anything, but your next morning you are better. That’s just an experimental finding. So it looks like the period of sleep somehow made you better. And then the neuroscientist made a discovery. The neuroscientist found that in fact the brain has been rehearsing during the night. This was fun in the rat, first of all. They found that the neurons that were firing during the day actually fired again during the night in precisely the same sequence and the same neurons that had been activating during The day in a particular period of learning. So for instance, the rat maybe was exploring a maze. There was a very precise pattern of activity of the play cells that, you know, each neuron has a certain place. And so when you move in a maze, these places get activated in a sequence. Well during the night, the very same neurons were activated in the very same sequence with one difference, whether they were activated much faster and therefore many more times. The brain was replaying the activation perhaps 15 times or 20 times faster. And this is a way to rehearse during the night without even being conscious of it. There have been many, many studies of this, including in humans. We can for instance enhance this replay and then the brain learns better. If you have deeper sleep and more periods of these dis sleeps when there’s more replay than you have better learning, (Time 0:45:47)
aprendizaje evidencia experimento sueño
aprendizaje evidencia experimento sueño
Más evidencias de la importancia del sueño en el aprendizaje. Transcript: Speaker 2 Does it matter what time of day we do the original learning? Speaker 1 It doesn’t seem to. It also does not matter when we sleep. So for instance, the experiments have had naps and naps are also important and they have a consolidation effect. It does matter at what age you do it. So once again children are advantaged, there is one study where the effects of the same amount of sleep one night of sleep is three times larger in young children. So you see how important it is and we all know of course that babies and young children must sleep longer. It’s much more important for them because they are learning more. In fact there is evidence for sort of homeostasis. So when you do a lot of learning during the day, you have to sleep more during the night. It’s a wonderfully organized system which you just need to be more aware of it because many of us lose sleep, we get distracted by TV or by series or by the computer and when we lose sleep, We also lose some ability to learn and to consolidate. Speaker 2 I think that that’s one piece of practical information we can all use. Speaker 1 Oh yes. And also I want to say we make discoveries during sleep. It’s not just reserved for scientists. So first of all this has been proven by experiment. It’s true that the brain will make discoveries during sleep. But we all do. In fact children make discoveries for instance about the meaning of words during sleep. They have experience of a few words during the day but they generalize, they place the word in a broader context, they understand the multiple meanings of some words during the night And when they wake up the next day they have consolidated their learning of the world. (Time 0:48:01)
2min aprendizaje evidencia experimento sueño
2min aprendizaje evidencia experimento sueño
La importancia de proveer un ambiente cultural, vincular y biológicamente rico a los niños Transcript: Speaker 1 The period of childhood is one of the most important periods for shaping a human mind. It is a period where you need to enrich the environment because the amount of enrichment of the environment shapes who we are basically. And the human brain especially in this respect it’s absorbing much more from others in particular from the social environment. I think this means that we have a special responsibility for our children. We need to provide them with the best possible environment during this blessed period where they are learning. I say that because I think we can still improve a lot, you know. First of all, we must provide children with the appropriate language. There is a beautiful experiment that I explained in the book that showed that the amount of child directed language that is received by a child has a direct impact on the organization Of the brain networks for language. And we know that there are huge differences as a function of social economic status when children are preparing to learn to read. And this will have a determinant effect on their ability to learn to read fast and on the rest of their scholarly. So we have a big responsibility here. It’s not just the environment as providing stimuli for learning, but it’s also providing the appropriate biological materials. Nutrition, vitamins, oxygenation, sleep, the lack of poison such as alcohol for instance. There is a whole environment for healthy brain to learn optimally. I mentioned alcohol because we know that exposure to alcohol in utero as dramatic effects on the ability of the child brain to develop properly and this will have lifelong effects. (Time 0:50:01)
2min aprendizaje desarrollo infancia salud