Exploring the Connection Between Brain, Mind and Body

[Music] the brain consists of over 80 billion neurons our thoughts emotions and actions are all encoded in the activity of these neurons you look at the single synapse we know it's part of a circuit and we know that those circuits are processing many thoughts what is thought what is consciousness brain is the most complex organ in the body and understanding its inner workings is the biggest engineering challenge of the century just like we can use machine learning and data science to make sick sense of physical signals like wireless signals that come to a cell phone we can also use the same principles to make sense of much more complex signals that come from the brain how does the brain function and how can we harness that knowledge for making better interfaces between brains and machines it's not only just an organ but at the same time it performs a lot of computation you're hallucinating what's out there from the bits of information coming in the brain circuits actually fill in what's missing so your perception is an active perception it's not a passive perception the largest concern that we've found that patients have is keeping their cognitive abilities intact remaining independent and it's all about the brain state of mind is absolutely crucial in our health state there are no diseases that are focused on one organ in the body so everything is connected so the brain is no exception the great opportunity here from a engineering and science perspective is to study the brain not in an isolated fashion but also how it interfaces with the body of course and the environment without the body the brain wouldn't be having any connection with the world and it's not just a conduit it is part of the computation itself have you ever wondered why robots are so stiff the controllers can't control all the joints at the same time the way that we can control all of our limbs at the same time i mean compare a bird with a jet fighter for example you know the jet fighter is very stiff just goes in one direction it can turn slowly but birds you know they can flap their wings they can a light on branches they could do u-turns i mean it's just astonishing control is going on in parallel many different parts of the brain at the same time and somehow combining all of that information and at different time scales and we hope that we can reverse engineer the control system in the brain so that we can adopt that in artificial bodies [Music] the problem with the brain is it's very heterogeneous there are thousands of different types of cells and there's an incredible diversity in terms of how they're connected and mental disorders are circuit diseases there are diseases of the interactions between the cells you've got to understand something about how the different parts of the brains interact with each other and rebalance it and the brain initiative has created the neurotechnologies that are going to allow us to go in and make those changes my lab focuses on building plantable systems to study the computation the neural circuit function in the brain so in order to understand brain circuits we need spatial information and temporal information spatial information represents the location of the neurons they are wiring the synapses they form and the temporal information represents their functional activity our goal is to develop neural technologies which enable us to record the neural activity both high spatial and high temporal resolution simultaneously so we developed neural implantable electrodes which are optically transparent so they are made of a material called graphene graphene is a 2d material so it's only one atomic layer thick and it's a hexagonal carbon lattice graphene is flexible which means implantable electrodes made of graphene can actually conform the brain and graphene electrodes because they are optically transparent they don't absorb the light and when you combine them with optics it does not generate light induced artifacts and finally graphene is a material which actually can block electrochemical reactions preventing corrosion of metals or corrosion of other materials that are used in implantable systems so it has potential to improve the longevity of implantable electrodes so these transparent sensors can be used in multimodal experiments where we can combine electrophysiological recordings of neural activity simultaneously with optical methods there's so much we do not know about basic things like how our emotions are represented in the brain where they're represented in the brain or um how how we control our movements so our lab basically tries to develop machine learning algorithms that can decode brain signals have these signals represent our actions and our emotions our senses and then also come up with brain machine interface technologies to treat various neurological and neuropsychiatric disorders we build mathematical and machine learning algorithms that can look at brain activity and then based on that track say mood symptoms under depression with the goal to then provide therapy in a tailored manner meaning provided only then and as needed rather than giving you therapy without knowing your current disease symptoms so in these cases the idea of the brain machine interface is to actually regulate abnormal brain activity patterns through some sort of intervention like deep brain stimulation i don't want to always have the dbs on i want to only turn it on when the patient needs it we actually record data from each individual patient and we train our models to that patient's data so that we can come up with a personalized model for them and therefore personalize their therapy whether it's a brain machine interface to restore movement in paralysis or whether it's a brain machine interface that tries to regulate mood symptoms in depression we're trying to build these closed-loop systems that provide tailored therapies [Music] every time you remember something you edit it so that what you remember may not actually be what happened your brain is in the business of keeping track of what's important in the world and what you might need to know in the future in order to survive a single neuron can be used to represent many different things that are similar it's a very efficient way of doing it if you had to store one thing one idea one the thought in one neuron you you know even though you have a lot of neurons it would be very messy that's what a computer does right it stores one piece of information in one memory location but the brain and the cortex have an associative way of relating information and and this allows us to come up with very fast answers because we can access all the memories at the same time in parallel one of the focuses of my research is not on how does memory form during the day when you have an experience but rather how is it consolidated at night when you fall asleep we've known for some time that that sleep was important for restoring brain circuits but also we now know for reorganizing the the brain circuits so that they can recover information that you experience during the day and what we discovered was that during the night there are thousands of something called sleep spindles bursts of activity electrical activity that occur a couple of seconds when you're going between dream sleep and what's called deep slow wave sleep and they occur every 5-10 minutes and this is a trigger for a whole cascade of biochemical reactions that then occur which then can change not just the strength of the connection the synapse but actually the physical size can get bigger and smaller that's why you go to sleep at night and we know that the more sleep spindles you have the better you remember things but what was surprising is that these sleep spindles which are thought to be synchronous across the whole brain at the same time we now know that these sleep spindles actually are traveling ways that are going in a circular motion and if if you've ever seen star wars you know that why we call these princess leo waves good brain health throughout the lifespan we're finding that it is the number one health concern cognition can be remedied and improved and maintained through sleep management physical exercise stress management addressing health behaviors in the home the interesting aspect of home monitoring is that there are so many opportunities to make tiny inferences from unobtrusive measures we've measured walking speed routinely in hallways where there are very few distractions and we're able to infer cognitive health from daily repeated measures on walking speed in fact walking speed is a standard neuropsychological test there are some confounding factors like arthritis and things like that but we can look for within subject trends when we measure this every single day so there's a variety of conditions that we're able to monitor in the home one is just general cognitive ability overall most of this is done with our cognitive computer games we started out with focus groups doing needs assessments understanding what they like to play so we needed to make the games fun so people would want to play them every day then we embedded algorithms within those games so we have games that focus on inferences for divided attention then we have others that look at executive function what we call memory buffer length with our memory game and verbal fluency with word games we typically will rely on commercially available sensor technology we add the ai and the infrastructure for data processing and inferences to help tailor the best possible intervention to the individual in the home [Music] there's a convergence occurring between computing in the brain and computing in artificial intelligence artificial intelligence and natural intelligence can come together by taking inspiration from how the brain works not at this higher level of abstraction but really get at deeper levels of the biophysics of what the brain is doing brain only consumes energy similar to a light bulb computation is performed in such an efficient way that the same neural circuits are involved in both processing the information and storing the information the computational power of the brain was way vastly beyond anything that we had up until recently but now that is coming closer we can actually take advantage of that we can actually build hardware that is coming closer it's still far away but coming closer to the kind of parallel hardware we see in the brain in my lab we also developed neural inspired or brain inspired programmable nanoelectronic devices our goal here is basically to emulate some of the programming functions that are performed by the synapses and neurons in the brain in our approach we actually perform computation in the memory and that can actually improve the energy efficiency orders of magnitude because most of the energy is consumed by that data transfer between memory and processor in conventional digital architecture and this parallel architecture is also particularly suitable for processing neural information so my lab is also trying to use this system this neuro inspired hardware for building faster arm and more efficient brain computer interfaces [Music] the workshop and symposium on brain mind and body really encompasses neuroengineering neuroscience cognitive science machine learning different from traditional means for doing neuroengineering or neuroscience typically are very invasive and restrictive this workshop is about harnessing unobtrusive nerve technologies to probe and modulate brain function in order to solve these more difficult problems we need many different parts of engineering to come together we need them to talk to each other we need them to work together and and that's happening here in embs as engineers we can make progress before scientists really figure out the entire function of the brain we take a systems approach of understanding how parts of the brain function without a need to know all the detail i find it fascinating how complex it is i find it fascinating how we can use machine learning and mathematical tools to understand it potential ways to interface with it to come up with alternative therapies for millions of patients who are suffering from brain disorders trying to understand these very complex networks and how they are able to learn it's just a wonderful time to to be around and doing research [Music] you

Channel: Institute for Neural Computation

Views: 20,797

Keywords:

Id: Ec_EXSV4qJQ

Length: 15min 29sec (929 seconds)

Published: Thu Jul 28 2022