About Neuroscience. Physics, biology and IT together
We have already written that one of the results of study at the university (from our point of view) should be the breadth of knowledge. Today we publish the post of one of our graduates, who decided to change the direction of work after university, and try to see what came of it.
I want to tell you about interdisciplinary research (a very fashionable phrase, isn’t it) that I am doing now. While studying at the university (Bachelor of St. Petersburg Polytechnic and Master of Academic University in the field of nanotechnology), I worked in several laboratories of the Physicotechnical Institute. Ioffe dealt with semiconductors in various manifestations (molecular beam epitaxy, deep levels, semiconductor lasers). Now my field of research has changed dramatically and is called Neuroscience. Geographically - Stanford, California .
Let's try to figure it out.
')
The task of this science is quite ambitious - to understand how the brain works. Historically, this issue has been studied in the field of biology, medicine, psychology and a bit of philosophy. Now, genetics, physics, mathematics and computer science are leading. In general, neuroscience has established itself as a separate science, the relevant faculties have even begun to appear in universities. There are about 100 billion neurons in the human brain and 150 trillion connections between them - the scale is, of course, incomprehensible. The brain at the moment remains, perhaps, one of the most poorly studied phenomena of nature.
Many have probably heard that information in the nervous system is transmitted in the form of electrical impulses. This video demonstrates the activity of several dozen individual neurons in the mouse brain:
A mouse (genetically modified with fluorescent neurons) is alive and well (which is very important, since there is especially nothing to look at in the dead brain). In the video, the level of fluorescence of neurons changes when an electrical impulse is generated in them, and as a result, the cell bodies “flash”. The electrical impulse is further transmitted along the axon (output) of one neuron to the dendrites (inputs) of the others. The structure of each neuron separately has been studied quite well. It is also known which areas of the brain are responsible for what. However, at the level of neural cell ensembles - complete darkness in scientific terms.
How does such an ensemble of neurons perform certain information processing functions (and which ones)? How is memory encoded in the activity of neurons and how are sensory signals presented? What are the general properties of the ensemble of neurons as a dynamic system, as well as many other interesting questions - this is what many scientists in this science are fighting over.
Many agree that it is scientists in physics, mathematics and engineering that can advance neuroscience forward, build more coherent theories based on abstract mathematical models. Without biological "magic", however, is indispensable. Such is the symbiosis of science. By the way, the vast majority of neural networks studied in computer science to networks of real neurons still have very little to do. Real networks are strictly non-linear, and their dynamics, even on small scales, is rather poorly described by existing mathematical models. But absolutely, if one succeeds in constructing a theory, then information technologies will borrow many algorithms from the brain in the future. For example, Google’s pattern recognition algorithms work somehow, but, of course, not as energy efficient and not as reliable as your brain (for a piece of chocolate, of course).
Finally, about what exactly I have to do. For two years, I designed and created an installation for optical activity detection on the scale of several thousand neurons. Installation is quite complicated, with all the optics, electronics, mechanics and software were made, in general, from scratch. As a result, data can be obtained - the same as in the video only in a much larger area of ​​the brain, by 1-2 orders of magnitude.
If we draw an analogy with semiconductors, then it is unlikely to make a semiconductor laser without a scanning electron microscope. So here - we need tools for observation, detection and manipulation of the object of study. Now that the instrument has appeared, the most interesting period is ahead - the experiments themselves, and after understanding the results and, I hope, some progress in understanding the structure of the brain.
Many, I think, are interested in the question of how nanotechnology and education at the Academic University is connected with what I am doing now. I will cite a few considerations. Whatever you decide to do in the future, you still have to constantly learn new skills, knowledge and skills, and no master’s degree can teach you everything that is possible (and, I hope, does not set such goals). This is especially evident when you are doing research, because nobody can tell about the fact that no one has yet discovered or done, and you have to think about new ideas yourself. In AU, an environment has been created in which not only knowledge is provided, but skills to solve new problems appear, to deal with scientific and technical problems. This is in my opinion very important. Of course, if you decide to pursue semiconductor physics or nanotechnology after graduating from the master's program, AU is the best place I know of. But even if you later want to change the direction in part or even dramatically, then, firstly, all the doors are open, and, secondly, your knowledge will work for you.
Oleg Rumyantsev
I want to tell you about interdisciplinary research (a very fashionable phrase, isn’t it) that I am doing now. While studying at the university (Bachelor of St. Petersburg Polytechnic and Master of Academic University in the field of nanotechnology), I worked in several laboratories of the Physicotechnical Institute. Ioffe dealt with semiconductors in various manifestations (molecular beam epitaxy, deep levels, semiconductor lasers). Now my field of research has changed dramatically and is called Neuroscience. Geographically - Stanford, California .
Let's try to figure it out.
')
The task of this science is quite ambitious - to understand how the brain works. Historically, this issue has been studied in the field of biology, medicine, psychology and a bit of philosophy. Now, genetics, physics, mathematics and computer science are leading. In general, neuroscience has established itself as a separate science, the relevant faculties have even begun to appear in universities. There are about 100 billion neurons in the human brain and 150 trillion connections between them - the scale is, of course, incomprehensible. The brain at the moment remains, perhaps, one of the most poorly studied phenomena of nature.
Many have probably heard that information in the nervous system is transmitted in the form of electrical impulses. This video demonstrates the activity of several dozen individual neurons in the mouse brain:
A mouse (genetically modified with fluorescent neurons) is alive and well (which is very important, since there is especially nothing to look at in the dead brain). In the video, the level of fluorescence of neurons changes when an electrical impulse is generated in them, and as a result, the cell bodies “flash”. The electrical impulse is further transmitted along the axon (output) of one neuron to the dendrites (inputs) of the others. The structure of each neuron separately has been studied quite well. It is also known which areas of the brain are responsible for what. However, at the level of neural cell ensembles - complete darkness in scientific terms.
How does such an ensemble of neurons perform certain information processing functions (and which ones)? How is memory encoded in the activity of neurons and how are sensory signals presented? What are the general properties of the ensemble of neurons as a dynamic system, as well as many other interesting questions - this is what many scientists in this science are fighting over.
Many agree that it is scientists in physics, mathematics and engineering that can advance neuroscience forward, build more coherent theories based on abstract mathematical models. Without biological "magic", however, is indispensable. Such is the symbiosis of science. By the way, the vast majority of neural networks studied in computer science to networks of real neurons still have very little to do. Real networks are strictly non-linear, and their dynamics, even on small scales, is rather poorly described by existing mathematical models. But absolutely, if one succeeds in constructing a theory, then information technologies will borrow many algorithms from the brain in the future. For example, Google’s pattern recognition algorithms work somehow, but, of course, not as energy efficient and not as reliable as your brain (for a piece of chocolate, of course).
Finally, about what exactly I have to do. For two years, I designed and created an installation for optical activity detection on the scale of several thousand neurons. Installation is quite complicated, with all the optics, electronics, mechanics and software were made, in general, from scratch. As a result, data can be obtained - the same as in the video only in a much larger area of ​​the brain, by 1-2 orders of magnitude.
If we draw an analogy with semiconductors, then it is unlikely to make a semiconductor laser without a scanning electron microscope. So here - we need tools for observation, detection and manipulation of the object of study. Now that the instrument has appeared, the most interesting period is ahead - the experiments themselves, and after understanding the results and, I hope, some progress in understanding the structure of the brain.
Many, I think, are interested in the question of how nanotechnology and education at the Academic University is connected with what I am doing now. I will cite a few considerations. Whatever you decide to do in the future, you still have to constantly learn new skills, knowledge and skills, and no master’s degree can teach you everything that is possible (and, I hope, does not set such goals). This is especially evident when you are doing research, because nobody can tell about the fact that no one has yet discovered or done, and you have to think about new ideas yourself. In AU, an environment has been created in which not only knowledge is provided, but skills to solve new problems appear, to deal with scientific and technical problems. This is in my opinion very important. Of course, if you decide to pursue semiconductor physics or nanotechnology after graduating from the master's program, AU is the best place I know of. But even if you later want to change the direction in part or even dramatically, then, firstly, all the doors are open, and, secondly, your knowledge will work for you.
Oleg Rumyantsev
Source: https://habr.com/ru/post/218573/
All Articles