Schools What are the best schools for Neurophysics/Comp. Neuroscience

[fader]

Hey guys,

I'll be applying for Physics and Comp. Neuroscience PhD programs in the Fall. Any good schools worth applying to? Which schools consider "neuro-related" physics their strength? All I know is on the East Coast, UNC has a Neurophysics lab, and on the West Coast, Berkeley & UCSD are supposed to be good for this specialization. Any others worth noting?

Thanks

 

[cmos]

On the East Coast, I have a friend who is looking into the neuroengineering track at the University of Rochester. I'm note sure how well this school ranks with the ones you mentioned though.

 

[mathcompsci]

I'm pretty certain UC San Diego has an amazing school for this.

Here is the link to their program.

http://neurograd.ucsd.edu/

My wife is wanting to go into neuroscience. Can I ask you what you background is out of curiosity. What kind of bachelors did you get?

 

[mathcompsci]

I just found (seconds ago) a gold mine for you. My wife will be quite pleased also.

There is a massive list of all the Neuroscience programs in what appears to be all of North America at the following site.

I did a quick search for the state MA and it saved me a lot of time from going to each and every single university's web page and looking at the details.

Check it out.

http://www.andp.org/

 

[Cincinnatus]

I am a graduate student in theoretical neuroscience. I feel sort of uncomfortable making such a list of good schools since it would inevitably neglect so many good people with positions elsewhere. So, rather than give you a list of all the schools at which I think one could get a very good education in theoretical neuroscience, I will instead give you the list of schools that a recent applicant cohort (mine) were all talking about the most. That is, the "hot" schools for theoretical/computational neuroscience right now seem to be:

MIT
Columbia
University College London (Gatsby Unit)
Princeton
UC San Diego
UC San Fransisco
UC Berkeley
CalTech
New York University

A generally good way to look for information on where interesting research is happening is to follow the funding. For theoretical neuroscience (especially on the more systems/cellular side of things) a good place to look is the website of the Schwartz foundation. They've established "initiatives" at most of the schools where there is a significant amount of theoretical neuroscience work going on. Here's the link: http://www.theswartzfoundation.org/

I can give more specific recommendations if you are more specific about your interests. Saying you want to do theoretical or quantitative work in neuroscience really just tells me about your choice of methodology. It says nothing about any actual research questions you might be interested in...

 

[fader]

Thanks guys for all the recommendations and links!

mathcompsci:
I did a double major in Biochemistry (B.S.) and Physics (B.A.). I worked in a neurochemistry lab in the chemistry department and neurosurgery department at a hospital. That was when I was premed.

Once I graduated, I realized that med school really wasn't my cup of tea. My interest in Biochem proper waned, and instead I got really interested in Neuroscience & Quantum Computation. I'm sure Cincinnatus would agree with me that although the brain isn't a quantum computer, the subjects are related in their interdisciplinary approach. Quantum Computation takes the best from math, physics and computer science, while neuroscience is an even bigger hodgepodge of biochemistry, physics, computer science, math, etc.

I see neurophysics as relevant to adapt to current technological evolution, while at the same time improve the quality of human life. You could say I'm one of those guys who wants something like the Matrix to occur (of course, the AI shouldn't enslave us). Quantum Computation/Neural Computation are emerging technologies that will be widely utilized in the future. So, I would say that my interests live at the interface of these sciences.

Cincinnatus: I think you're the right person to keep contact with as far as my career goes. Being a theoretical neuroscience grad student, what could you say about my current status? I'll be applying to just physics programs with an interest in condensed matter and also be applying to comp. neuroscience programs. I don't mind which program I get into as long it'll give me a flexibility to move around and try new things. Although I did find out from a few friends that a lot condensed matter people having been going into neuroscience. Have you run into a lot of physics people? Has your program let you take courses outside of neuroscience, namely in physics or math?

So far my list for programs includes:
Berkeley
UCSD
CMU
UNC
Rochester

I think I'll have to check out UCSF and NYU. I'm currently finishing up some core courses (Mechanics, E&M) at Cornell.

Also, while I'm applying for these programs, I need to find a good internship to keep myself busy next year. Any recommendations? So far, I've just put down NIST, IBM, NRL or LANL for internships.

Thanks for all your help guys!

 

[Cincinnatus]

There are lots of people with physics backgrounds in neuroscience programs. I know some people who are technically in a physics department also but are studying neuroscience, though this is a bit more rare. I'd say that physics is the most common undergraduate major for people studying theoretical neuroscience in graduate school, more common than math and computer science majors. As a double major in physics and biochemistry you'll be in a good position to apply to neuroscience graduate programs, they look for people with that kind of background.

You'll probably want to emphasize the part of your biology background which is "larger" than the molecular/cellular questions that biochemistry majors typically focus on. That is, it'd be good to learn some neurophysiology / systems neuroscience or cognitive neuroscience since these are the scales that most computational people work on.

 

[fader]

that's good to know.

I have a request:

Can you send me a couple of places to apply for an internship next year? I have a job working for the patent office, but would like something a bit closer to the field I'd be going into. I presume you went straight from college to grad school? What kind of experience would suit me best? I'd love to do something with neural networks or just general computational work. You know, places to work for as a research assistant. I know there's NIH, but I'm sure your grad school also has something to offer, since you're doing theo. neuro.

Thanks for the help again!

 

[Cincinnatus]

Well if you're an undergraduate now then you can apply to REUs and such programs. If you've already graduated then such positions are harder to find I think, I don't know much about trying to do that. I imagine that you would need to convince a particular professor to give you a job.

Try to find a theoretician who is willing to let you work on a computational project. That's probably more important if you don't already have that experience. If you do already, then it might be a good idea to do an internship in another area of neuroscience (almost any will do). You need to have some knowledge base and interests in a particular area of neuroscience, just saying you're interested in "theoretical" work doesn't mean much and makes you sound less serious. You need to be able to say "I'm interested in applying theoretical/computational methodology to the study of <whatever>".

 

[fader]

I've already graduated. Are there any good internships to look for? I know LANL has a neural computation group. I'm sure I could do a one year thing with them. I had to give a presentation in a quantum computation course about using neural networks for image processing, except only applying those principles to an adiabatic quantum computer.

So, image processing is one possibility. Also, I keep hearing that people model "this part" or "that part" of the brain with NEURON software, the brainchild of people at Yale. Do you work with NEURON?

I worked in a neurochemistry lab during my undergrad primarily focusing on Cyclic Voltammetry. I could just ask the same prof I worked with, but that's more analytical chemistry than computational neuroscience.

 

[Cincinnatus]

I don't know anything about the group at LANL. You probably already know, but I'll say it anyway. Neural computation is a huge umbrella term that includes a lot of areas of research many of which have little or nothing to do with neuroscience. I'm not saying that this work isn't interesting (because I certainly don't think that) but that the study of artificial neural networks is more properly an area of computer science completely divorced from the biology of the brain. The focus there is on using neural networks for computational purposes not trying to do things the way the brain "does it".

You asked about NEURON, I've used it... It's not very difficult to use. That's about all I have to say about it.

You should read David Marr's book Vision, or alternatively if you don't want to hunt down an old book, you could probably just find a summary someplace on the internet which would explain the relevant point. Marr has a philosophical point about the methodology of computational neuroscience which is still very influential.

This page seems to explain Marr's philosophical point very concisely: http://www.rci.rutgers.edu/~cfs/472_html/CogArch/DM.html

Warren McCulloch also wrote a lot of interesting (though now extremely dated) philosophical articles about computational neuroscience. I recall one I particularly liked called "What is a number that a man may know it and a man that he may know a number?" Of course there's also the famous paper: "What the frog's eye tells the frog's brain". I think both of these were reprinted in a collection of McCulloch's papers called "Embodiments of mind".

It's my humble opinion (disagreed with by almost everyone) that everyone in this field should read these classics.

 

[Asphodel]

http://depts.washington.edu/pbiopage/people_faculty_area_cn.php

 

[fader]

thanks for the washington link. I see there's a lot of physicists there.

Cincinnatus: Yeah neural computation is a big umbrella for other stuff. I actually prefer the divorce of neural networks from the biology. But at the same time, I'd like to see some kind of biological application in the future. I don't have a degree in CS, but I know quite a few programming languages. I'm just looking for something that's massively interdisciplinary.

So far, the Da Silva group looks fun to work with and so do the people at Washington. They have a theorist working on ANNs there, which sounds really interesting.