|
|
- 3Q: Robert Desimone on the federal BRAIN Initiative 2013-04-23
- Mapping the human brain, with its billions of neurons, is one of science’s most elusive projects. But a new federal program — the $100 million Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative — could help neuroscientists at MIT and other institutions unlock some of the brain’s mysteries.
How will MIT contribute to the initiative’s goals? How will the initiative impact research already being done at MIT and in the Boston area? How will science benefit? Robert Desimone, the Doris and Don Berkey Professor of Neuroscience and director of MIT’s McGovern Institute for Brain Research — and one of four MIT researchers selected to attend the initiative’s White House announcement on April 2 — discussed these questions with MIT News.
Q. What types of new technologies will be developed to achieve the goals of the initiative? And what are the benefits of the initiative, in terms of better understanding the human brain and treatment of neurological disorders?
A. Among the technologies being discussed are large arrays of nanoscale electrodes; robotic devices for massively parallel whole-cell recordings; new optical methods for imaging activity deep within the brain; and tools from molecular genetics that would allow neurons to store records of their own activity and which we could read out at a later time. We also need better ways to assess biological measures beyond electrical activity — gene expression, for example.
Much of this work will be done in animal models, but we must also develop noninvasive methods so we can relate what we learn from animals to what can be measured in human subjects. And finally, we will also need new analytical methods, including a lot of computing power, if we are to make sense of all these new data and to understand how 100 billion neurons can work together as a system.
When neurons interact with each other in large numbers, new phenomena emerge — much as new social phenomena emerge when large numbers of people interact in groups. Understanding these large-scale interactions will be important if we are to understand the basis of both normal behavior and the altered behaviors seen in many brain disorders. There is evidence that both autism and schizophrenia, for example, involve abnormal synchronous activity across widespread neural populations.
Q. What will MIT’s role be in the initiative? How will MIT collaborate with other Boston-area institutions (and institutions around the nation) to achieve the initiative’s goals?
A. MIT is well-positioned to contribute to the BRAIN Initiative, as many of our researchers are already leaders in developing new technologies for neuroscience. One example is optogenetics, a method for controlling brain activity with light that is already revolutionizing the field.
At the McGovern Institute, we have established a neurotechnology program that provides seed funding for neuroscientists to work with engineers, computer scientists, materials scientists and so on, both within and beyond MIT. We’ve already supported more than 20 such projects, some of which have now turned into major research programs.
We’re very fortunate to have so many top research and clinical institutions in Boston, and we have strong collaborations with many of them. The Martinos Center at the McGovern Institute, where we do human neuroimaging, shares strong ties with its sister center at Massachusetts General Hospital, and also has collaborations with many other local hospitals and universities. Some of us are also members of a Boston-wide initiative to understand the activity of large neural populations, funded by a grant from the National Science Foundation. We also have faculty affiliated with the Broad Institute, and with the Stanley Center for Psychiatric Research, which is tremendously helpful for our work on psychiatric disease.
Q. How will the initiative affect the research already being done at MIT and other facilities in the Boston area, including MIT’s neighbors in Kendall Square?
A. MIT labs will apply for funding through the BRAIN Initiative as soon as the funding mechanisms are established. Many of us hope that President Obama’s support will encourage private foundations and individuals to contribute.
MIT has a strong track record of working with industry, and we will certainly need to do that if our discoveries are to lead to new therapies. Kendall Square has a great concentration of biotech and high-tech companies, and several large pharmaceutical companies also have a strong presence here. I see them as natural collaborators on the BRAIN Initiative, especially given the huge unmet need for new treatments for brain disease.
Beyond new therapies, I believe the new technologies developed through BRAIN will lead to many other spinoff products, from new optical devices to intelligent machines. The president said it well: Scientific research has had a great return on investment.
- 2008-05-15
-
- The Interplay of Analogy-Making with Active Vision and Motor Control in Anticipatory Robots 2007-04-22
-
Kiryazov, K., Petkov, G., Grinberg, M. , Kokinov, B., and Balkenius, C. (2007). The Interplay of Analogy-Making with Active Vision and Motor Control in Anticipatory Robots. In Butz, M. et al. (Ed.) Anticipatory Behavior in Adaptive Learning Systems: From Brains to Individual and Social Behavior. LNAI, 4520, Springer-Verlag
- Training Recurrent Neural Networks by Evolino 2007-04-17
-
J. Schmidhuber, D. Wierstra, M. Gagliolo, F. Gomez:
Training Recurrent Neural Networks by Evolino. To appear in
Neural Computation.
- Modeling Top-Down Perception and Analogical Transfer with Single Anticipatory Mechanism. 2007-05-09
-
Georgi Petkov, Kiril Kiryazov, Maurice Grinberg, Boicho Kokinov (2007); Proceedings of the Second European Cognitive Science Conference, Greece
- Studying XCS/BOA learning in Boolean functions 2007-03-13
-
Butz, M.V., & Pelikan, M. (2006). Studying XCS/BOA learning in Boolean functions: Structure encoding and random boolean functions. Proceedings of the Genetic and Evolutionary Computation Conference (GECCO-2006). 1449-1456.
- Surprise-Driven Belief Revision 2007-04-10
-
Lorini, E., Castelfranchi, C. (2006); In Proceedings of Second Biennial Conference on Cognitive Science, St. Petersburg, 9-13 June, 2006.
- In 2.007, slow and steady wins the race 2013-05-10
- Tim the Beaver, MIT’s mascot, is on an operating table and in need of several different medical procedures.
That scenario, a variation of the board game “Operation,” was the basis for this year’s final competition in 2.007 (Design and Manufacturing I), in which students design, build and operate small robots that compete in a variety of tricky tasks.
This year’s challenge, which featured several ways to earn points during two-minute matches, led to a wide variety of robot designs and strategies. In the end, the victory went to sophomore Aleksya Aguirre, whose “Hot Dog” robot took a slow-and-steady approach to carrying out two of the tasks — and succeeded in those almost every time.
For the first time in the course’s 43-year history, the final matchup in the single-elimination tournament, which started with 92 competitors, featured two women. The finalists, Aguirre and sophomore Sophie Seidell, are both mechanical engineering majors.
At the beginning of the semester, each 2.007 student who wishes to compete receives an identical kit of materials, motors, gears and controllers for use in creating a robot for the competition. Others in the 180-student class have the option of spending the semester designing an electric vehicle or an underwater vehicle.
With the help of several instructors and teaching assistants, each student develops a strategy and designs a robot; some — such as Seidell — design two robots to attempt different strategies in the competition.
Competition tasks varied in difficulty, and in point value. Removing a metal wrench from a rectangular pit at Tim’s ankle was worth 25 points. If the wrench was then placed on a shelf, that earned extra points — the higher the shelf, the higher the total. Performing “angioplasty” by inflating a balloon inside a stent scored points based on the diameter to which the balloon was successfully inflated.
Removing a “butterfly” from Tim’s stomach — a tricky operation, simulating laparoscopy — garnered 13 points. But these points carried a bonus: Only after retrieving a butterfly was a robot allowed to cross to the opponent’s side of the operating table to interfere with the other student’s robot, or to score extra points by grabbing the opponent’s wrench, butterfly or other parts.
That strategy was key in one quarterfinal match. Sophomore Jose Smith’s “magic man” robot was able to grab both his own wrench and butterfly, and then proceeded to block sophomore Ryan Fish’s “Loki” robot, which had already picked up a wrench, from scoring further.
The easiest task was retrieving one of Tim’s ribs and dumping it into a “biowaste pit” — worth just one point per rib. Two other tasks turned out to be so difficult that few students even attempted them, and none succeeded during the finals: stretching a rubber band between two posts, and performing an “fMRI” to identify a pattern of magnets beneath Tim’s head.
During most of the competition, students controlled their robots using TV remotes or controllers for either a videogame or a radio-controlled toy car. But each student also had the option of operating the robot autonomously for 30 seconds, and any points scored during this autonomous period were doubled.
While some had succeeded in earlier elimination matches, during the finals only one robot was able to score points in the autonomous mode: Sophomore David Flamholz’s robot snatched the wrench autonomously, gaining him 50 points and a wild round of applause and cheering from the audience.
As always, unexpected issues arose: A wheel came loose, a controller stopped working, or (quite often) one of the several pits on the board snagged a robot’s wheel.
Aguirre’s simple, focused strategy ultimately prevailed. Unlike many robots that were flexible enough to attempt several tasks, hers was designed to do just two things reliably: Pick up the wrench and one rib, and place them on a shelf. It worked almost every time.
“I decided on a simple design that would be consistent,” she said after her victory in her very first robotic contest. “I came into the class with no experience, and I was very intimidated.”
She’s not intimidated anymore. “I really enjoy the hands-on work,” she said.
- Brains,anticipations, individual and social behavior: an introduction to anticipatory behavior systems. 2007-04-16
-
Butz M., Sigaud O., Pezzulo G., Baldassarre G. (in press). Brains,
anticipations, individual and social behavior: an introduction to
anticipatory behavior systems. In Butz M., Sigaud O., Pezzulo G.,
Baldassarre G. (Eds.), Anticipatory Behavior in Adaptive Learning Systems:
From Brains to Individual and Social Behavior. Berlin: Springer Verlag.
|
|
