Database Reference
In-Depth Information
It takes sometimes a decade to get these sorts of ideas out into the scientific
community, because in sciences there is never a kind of transparent ROI, so
it's harder to expect people to be really eager to get onboard. But the hope
is that if we can start building the right models to find the right patterns using
the right data, then maybe we can start making progress on some of these
complicated systems.
Gutierrez:
When did you start wanting to study the brain? And what
motivates your current work?
Jonas:
When I was a kid, I used to build circuits and taught myself how to
program a computer. I remember sitting in a 7
th
-grade class and wondering
what the assembly language was for my brain, because clearly there's this
computing thing going on there. When I arrived at MIT, I decided to double-
major in EECS and BCS sciences partly because I just couldn't give up the
engineering side and partly because I had some smart faculty members tell me
that if you actually want to do this sort of work, you really need to have the
hardcore quantitative background.
Then when I began graduate work at MIT building instrumentation to actually
record from the cells, I realized partway through that what we really needed
were the tools to understand the data that was being generated. I've come
to the slow realization over the past several years that I'm not really the
kind of scientist who has the patience to sit there and methodically explore
a given system with existing tools. Usually, after a few weeks of working, I get
frustrated with current tools, and I'm like—no, let's just build a better tool.
And it's this frustration that's part of what is really motivating me to help build
better tools.
The other part that drives me is being able to understand the brain faster. The
area of the brain that we study is called the hippocampus: it's kind of like the
RAM for your brain. There was a postdoc in a lab I worked in who discovered
a phenomenon called “reverse replay”
3
basically by staring at the data for five
years using traditional methods. I remember thinking I could have written an
algorithm five years ago that would have just found that. It had me thinking
of how much understanding is sitting on the hard drives of people in various
labs just needing these sorts of analytic techniques. So that really continues
to motivate me when I look at these data sets. I'm like, “Look, the answer is in
here somewhere!” The challenge for us is to actually build the tools to find it.
It's no different than trying to build a telescope to find something far away or
build a microscope to see something really small.
3
David J. Foster and Matthew A. Wilson, “Reverse replay of behavioural sequences in
hippocampal place cells during the awake state,”
www.nature.com/nature/journal/
