Prime Time Replay:
James Bower on
Project Genesis & The Digital Brain
MsgId: *brain_storm(1)
Date: Fri Mar 28 21:51:18 PST 1997
From: Rob_Killheffer At: 205.198.116.70
Welcome to Brainstorms! We'll be starting our show soon -- at 10 pm ET. Our guest tonight will be neuroscientist Dr. James M. Bower of the California Institute of Technology. Dr. Bower's work focuses on understanding how our brains process information -- specifically, how the cortexes of both the cerebrum and the cerebellum do so. Dr. Bower comes at the question from both a anatomical and a computational point of view, and he's made excellent use of increasingly complex computer models of neurons and neuronal activity. In fact, May's Live Science feature will be coming to us from Dr. Bower's lab, so you can view tonight's chat as a kind of preview. Join us!
MsgId: *brain_storm(2)
Date: Fri Mar 28 21:53:04 PST 1997
From: Rob_Killheffer At: 205.198.116.70
Please note that for the first half-hour or so we'd like to keep this chat closed while I question Dr. Bower about his work; I'll make an announcement when we're ready to take audience questions. So please hold back until then. Thanks!
MsgId: *brain_storm(3)
Date: Fri Mar 28 22:00:25 PST 1997
From: Jim At: 131.215.15.202
Hi there Rob -- seems as though everything is working properly -- so I assume here we go /ga
MsgId: *brain_storm(4)
Date: Fri Mar 28 22:04:45 PST 1997
From: Rob_Killheffer At: 205.198.116.61
Hi, Jim. Let me first ask, if I might, about one of your main areas of interest: the cerebellum, which has often been a relatively little-studied part of the brain, it seems to me. What's the cerebellum's role, so to speak, in the brain? What sorts of functions is it involved with? /ga
MsgId: *brain_storm(5)
Date: Fri Mar 28 22:08:28 PST 1997
From: Jim At: 131.215.15.202
The cerebellum is really a mystery. We have essentially known its anatomical organization and structure for 100 years ------ in fact we know more about the anatomical organization of the cerebellum than any other part of the mammalian brain ----- yet there are still very wide ranging debates about what it does. ---------- Most neurobiologists think that it is involved in coordinating movement ---- I don't think so at all ----- It think that it is involved in coordinating the acquisition of sensory data of all types. ------ but the essential point is that no body really knows./ga
MsgId: *brain_storm(6)
Date: Fri Mar 28 22:12:01 PST 1997
From: Rob_Killheffer At: 205.198.116.61
Your point of view on it, then, sounds somewhat like that of some earlier researchers who saw a role for the cerebellum in the development of the young brain -- that it was vital in the infant's acquisition of sensory data, and that sensory deprivation could hinder the development of several other brain functions. I'm thinking specifically of Dr. James Prescott's work, for instance. What's your take on that? /ga
MsgId: *brain_storm(7)
Date: Fri Mar 28 22:24:39 PST 1997
From: Jim At: 131.215.15.202
I am not able to get the messages in to you/ga
MsgId: *brain_storm(8)
Date: Fri Mar 28 22:25:45 PST 1997
From: Jim At: 131.215.15.202
Hmmmm -- I'm impressed -- usually nobody knows much about the cerebellum and its history..... which is quite interesting............. One of the interesting aspects of cerebellar theory, is that, over the last 30 years, the options that people have been thinking about have actually become more limited rather than more open. in other words, the tremendous growth in neuroscience has resulted in this case (and perhaps in others as well) in a kind of group conservatism. If you look at ideas more than that time ago, there is much more breadth of thought. One particularly interesting example is that the person most modern papers cite as providing clear evidence for the role of the cerebellum in motor control, Gordon Holmes, actually had a much more complex view of the cerebellum than he is credited with.
MsgId: *brain_storm(9)
Date: Fri Mar 28 22:25:52 PST 1997
From: Jim At: 131.215.15.202
Two things to say then .......... one, yes, the work of Prescott is very interesting from the prespective of our theories, and largely consistent, and two, the recent boom in the use of imaging techniques in humans is finally blowing the lid off of the more conservative views of cerebellar function./ap
MsgId: *brain_storm(10)
Date: Fri Mar 28 22:26:34 PST 1997
From: Jim At: 131.215.15.202
Sorry -- there seems to be a limit on the length of text on your server........ probably not a bad idea...../ga
MsgId: *brain_storm(11)
Date: Fri Mar 28 22:29:51 PST 1997
From: Rob_Killheffer At: 205.198.116.61
I'm very interested in how the new technology -- brain imaging and computer modelling -- has advanced or changed ideas about the brain. You use computer modelling a lot in your work, so why not tell us a little about GENESIS and about how "computational" neuroscience differs from the traditional variety. /ga
MsgId: *brain_storm(12)
Date: Fri Mar 28 22:32:26 PST 1997
From: Jim At: 131.215.15.202
This question is actually directly related to the previous -- in that models built in computers hold theorists to much higher standards than do imagined models. In other words, models of brain function that are only thought about can be adjusted to account for almost any data ----- REAL models in computers are much less forgiving...
MsgId: *brain_storm(13)
Date: Fri Mar 28 22:34:19 PST 1997
From: Jim At: 131.215.15.202
The kind of models we build with GENESIS are what we have termed "realistic models". The idea is that we take all the known anatomical and physiological properties of a particular brain structure, and code them into a simulator. We test the model onl data, and then we use the model to look for possible relationships between the structure and it sfunction. In other words, the model is used as a source of ideas -- not as a representation of some preexisting idea...
MsgId: *brain_storm(14)
Date: Fri Mar 28 22:35:43 PST 1997
From: Jim At: 131.215.15.202
The assumption, is that there is a very close relationship between the structure of the brain and its function. Probably much closer than for any machine we have ever constructed ourselves. Anyway, so far, so good. Many things have popped out of our modeling work that we did not imagine when we built the models to begin with. Some of them have been critical for our rethinking what the cerebellum does...
MsgId: *brain_storm(15)
Date: Fri Mar 28 22:38:16 PST 1997
From: Jim At: 131.215.15.202
one more quick point -- this kind of modeling has only realy been possible over the last several years because computers are only starting to be powerful enough to run models of these types. And I am not talking about PCs- but the biggest machines yet built -- which we use in our research. In fact, we can make a model of single neuron complex enough that it can bring the worlds biggest computer to its knees. this may give you some sense of how remarkable the brain is (there are 6 million of those neurons in your brain, for example). /pa
MsgId: *brain_storm(16)
Date: Fri Mar 28 22:43:31 PST 1997
From: Rob_Killheffer At: 205.198.116.61
Right - I noticed that you recently moved GENESIS (which stands for something like GEneralized NEural SImulation System, yes?) over to an Intel Paragon machine, very powerful. I was struck by something you said there about the brain's function being more closely related to its structure than in any machine we've built, and I thought how appropriate that would seem, in that nature tends to "build" structures very closely to their functions, unlike human engineers. ...
MsgId: *brain_storm(17)
Date: Fri Mar 28 22:47:10 PST 1997
From: Jim At: 131.215.15.202
Nature tends to "build" structures very closely to their functions, unlike human engineers. In fact, I would claim that the brain is so much more efficient and remarkable, that it is not even clear that we know what kind of machine it is. If you look through history, again -- we have always used whatever is the most sophisticated technology of the day for the brain. Greeks and romans -- acquiducts ........... Descartes -- machinary ........ in this century --- railroads -- telegraph systems --- digital computers ---- and now "parallel analog computing devices" ---- Are we getting any closer ??? I am not so sure./pa
MsgId: *brain_storm(18)
Date: Fri Mar 28 22:48:46 PST 1997
From: Rob_Killheffer At: 205.198.116.61
And I also thought back to one of the books I was reading to brush up on the cerebellum -- I didn't have its history down from memory! -- and how it made a kind of knee-jerk juxtaposition between "the physical structures of a child's brain" and "the hazy contours of the ego" -- but in light of ongoing research, it seems less and less likely that there's that big a difference between the brain's physical structure and what we describe as personality or ego, don't you think? /ga
MsgId: *brain_storm(19)
Date: Fri Mar 28 22:52:02 PST 1997
From: Jim At: 131.215.15.202
I am an extream reductionist (I have to be to do what I do)...... In my view, all there is is the brain --- ego, consciousness, feelings, emotions, everything is just slimy gook, when you get down to it. What our models already tell us, is that there is more than enough complexity within the nervous system to support amazingly complex, computational powers. but it is all just that - - computation with carbon./pa
MsgId: *brain_storm(20)
Date: Fri Mar 28 22:56:07 PST 1997
From: Rob_Killheffer At: 205.198.116.61
I agree with you pretty thoroughly on that, the material reductionism. It's hard for many to believe because it's so hard to get a real picture (I don't think I have one) of the order of magnitude of the brain's complexity. It's hard to imagine. At any rate, I wanted also -- since, believe it or not, we're running out of time already -- to ask specifically about the Purkinje cells that you're so interested in (you even named your web site "Purkinje Park"). What's so interesting about those cells, and what do you think they'll tell us about the cerebellum's function? /ga
MsgId: *brain_storm(21)
Date: Fri Mar 28 22:57:05 PST 1997
From: Jim At: 131.215.15.202
I don't think that anyone does ...
MsgId: *brain_storm(22)
Date: Fri Mar 28 23:00:47 PST 1997
From: Jim At: 131.215.15.202
Purkinje cells are the central neurons in the cerebellum -- just a few facts, and the you can decide for yourself if they are interesting--------- 1) they are the largest neuron in the brain; 2) they are continuously active at night (awake asleep or anesthetized) at rates of between 10 and 150 outputs per second; 2) the received inputs from 200,000 other neurons -- that is each cell receives that many inputs; and 4) most of all they are physically stunning -- absolutely beautiful; and last, but not least, the ones in your brain are the same shape as the ones in the brain of a fish - or a bird - or a reptile -- or the earlist vertebrate -- so what do we all need these cells for -- no one knows -- we are trying to find out /pa
MsgId: *brain_storm(23)
Date: Fri Mar 28 23:06:28 PST 1997
From: Rob_Killheffer At: 205.198.116.61
I'd have to say they sound pretty interesting. 200,000 inputs -- that's a lot, even for a neuron. We're over our time already, so let me apologize for not having time to take any audience questions. Never fear, though: you'll hear lots more about the work in Dr. Bower's lab when Live Science takes you there next month! For now, thank you, Dr. Bower, for appearing here on Brainstorms. I'll be tuning in with interest to our Live Science coverage of your lab./ga
MsgId: *brain_storm(24)
Date: Fri Mar 28 23:08:12 PST 1997
From: Jim At: 131.215.15.202
happy to be here -- and I hope that everyone learned something more than the fact that computational neurobiologists type too fast, and can't spell very well -- thanks for the interesting conversation. Jim /pa
MsgId: *brain_storm(25)
Date: Fri Mar 28 23:10:38 PST 1997
From: Rob_Killheffer At: 205.198.116.61
I think we learned some things -- enough to whet our appetites, at least. Thanks again for coming! For anyone who's interested, stop by Dr. Bower's web site at: http://www.bbb.caltech.edu/bowerlab for a more thorough look at what he and his colleagues are doing. Until next week, that's all for Brainstorms. Good night!
Home || Prime Time || Live Science || Machine Dreams || Project Open Book || SF-Fantasy-Horror
Continuum || Antimatter || Mind-Brain Lab || Interactive IQ || Gallery || OMNI ToonsQuestions, comments and suggestions can be mailed to the webmaster.
Copyright (C) 1997 by Omni Publications International, Ltd. All Rights Reserved.