Switching from Electrical Engineering to Theoretical Physics

[neuroantenna]

Hi all,

I'm going to graduate this May with a degree in Electrical Engineering. While I've done pretty well, my passion was never in EE and I'd like to aim for graduate school in Computational Neuroscience. When I look at the background and methodology employed by many favorite (current) researchers, I'm realizing that the best preparation would have been a double major in Math and Physics (the common Theoretical route, I assume?).

My specific interests involve interpreting biologically-plausible artificial neural networks as nonlinear dynamical systems and interpreting their activity to understand their function. While there are many researchers in this field, two great examples are Surya Ganguli (former Theoretical Physicist) and Eric Shea Brown.

I'm surely not the first engineering major to switch their focus towards Theoretical Physics, so I was hoping I could get some general advice on those who have transitioned. While neuroscience in practice doesn't look much like physics, the problem solving strategy of applying an abstract mathematical framework to uncover a system's first principles is still required. Beyond gaining fluency with the necessary math (mathematical statistics/statistical mechanics, nonlinear dynamics, topology/abstract algebra relevant to nonlinear dynamics), I would really benefit from developing that aforementioned problem solving strategy.

Hoping anyone has some general advice or can point me in the right direction for engineers with a newfound passion for physics/mathematics!Edit: I should note that "theoretical physics" isn't really the proper term for my future direction, but the undergraduate/master's level training in terms of math fluency, problem solving and critical thinking ability is quite similar. So I'm looking for advice to transition and obtain such skills.

 

[mitochan]

I'm surely not the first engineering major to switch their focus towards Theoretical Physics, so I was hoping I could get some general advice on those who have transitioned.

Recently I listened the radio program https://www.bbc.co.uk/programmes/m000fw0p on Paul Dirac. His case might be of your interest.

 

[neuroantenna]

So you're saying I'm going to win the Nobel?

 

[Joshy]

I do think the title can be a bit confusing about your goals.

I'm not sure if you're meaning to change your major before graduating? There's a very similar thread to this question if I'm understanding right (Double Major in Electrical Engineering and Physics?) I think most people agreed (or at least didn't outright reject it) that you don't need to finish your bachelors same major as the intended graduate studies- take a look at the requirements for the program that is catching your attention or talk to a counselor. I think it's fair to apply that to this question given the similarities and more so for your question because you are so close to finishing. I don't know anything about Computational Neuroscience... my personal guess based on the few universities I visited, that's it not a very common undergraduate major and so most of the people applying for graduate studies in that program have very mixed backgrounds (including electrical engineers).

What made you come to this conclusion about Computation Neuroscience? Did you take any relevant classes or worked on some project? Heard about it at a talk or seminar? Do you have any research experience at all? I would imagine if the more experienced/knowledgeable readers knew more (I'm very inexperienced myself) they might be able to offer you better insight.

 

[neuroantenna]

What made you come to this conclusion about Computation Neuroscience? Did you take any relevant classes or worked on some project? Heard about it at a talk or seminar? Do you have any research experience at all? I would imagine if the more experienced/knowledgeable readers knew more (I'm very inexperienced myself) they might be able to offer you better insight.

My current research experience is computational and involves neuroscience but isn't in the vain of research currently dominating mainstream comp neuro. I came to the conclusion by hunting down every single lab I would love to do grad studies in and looking at the background of the PIs and grad students. It's also quite clear given current research (this paper is perhaps the clearest example of how a physics/applied math background helps). There is an overwhelming indication that those from an applied math/physics background have a leg up. It's not that EE is irrelevant or inherently prohibitive, it's just my undergrad experience was far removed from stuff like state-space representation of dynamical systems (despite taking multiple control systems classes), nonlinear dynamics, stat mech, mathematical statistic etc...

However, after taking some time from posting OP, I've realized that regardless of my background, I'm always going to have to learn more theory. Such is the life of an academic. My present issue, however, is that it's hard to find a lab that I'm interested into gain relevant experience before applying to grad school because... I don't have prior experience. Caught in the millennial death-trap of not having prior experience. y33t.

Question: Should I (1) get a master's and take more applied math courses and more importantly get placed in a theory-driven lab (more preferable) that positions me well when applying to grad school, OR (2) just work in a more experimental lab and hope I can transition to a more theory-driven lab either in grad school or for my postdoc?

(1) is more expensive, but gets me closer to my eventual goal. (2) is less expensive but might make it more difficult to transition to my eventual goal.

 

[DaveE]

I would focus more and what you already know versus what you want/need to know in the future. The degree listed on your diploma, EE vs. physics, just isn't going to be nearly as important as your specific experience. An EE degree can represent such a wide range of study that I don't think people will use it alone as a screening tool.

 

[neuroantenna]

I would focus more and what you already know versus what you want/need to know in the future. The degree listed on your diploma, EE vs. physics, just isn't going to be nearly as important as your specific experience. An EE degree can represent such a wide range of study that I don't think people will use it alone as a screening tool.

I see. I get the general gist of "stay in your wheelhouse" but this is a bit difficult as my EE programs didn't focus on the specifics that would normally make EE fit well with Comp Neuro. However I do get the idea of sticking to one's guns.

Recommendation on whether I should work in a more experimental lab (unpaid intern, and transition later) vs do a masters program ($$) to work in a theoretical lab (closer to my goals)?

 

[Joshy]

I would apply for both and see what you could get; if you can get into a masters program without the internship, then go for it.

If you cannot get into the masters program, then you might need the experience to support your application the next time you apply. I would not turn away the masters program for an internship even if it's paid.

Look at the program requirements for the university you are interested. You probably will need recommendation letters, and I would start writing a statement of purpose... have it reviewed.

 

[StatGuy2000]

To the OP:

I see that you are interested in pursuing graduate studies in computational neuroscience. Based on what I know (from people who I know personally who have entered the field), the most important skills to have in pursuing this interdisciplinary field would be someone with a strong quantitative background. Such a background could include mathematics or physics, but also include people backgrounds in engineering fields, such as electrical engineering, as well those with backgrounds in computer science. Or those whose backgrounds combine a background in, say, psychology or biology with math/CS/physics/engineering.

So your education in electrical engineering should actually give you a more than adequate (in fact excellent) background in pursuing a PhD in computational neuroscience. I would suggest that you speak to your academic advisor and to those responsible for admission programs in computational neuroscience to discuss your background and interests.

Just my 2 cents worth.

 

[DaveE]

"stay in your wheelhouse"

No! Create your own wheelhouse.
Don't be too constrained by how your institution defines things. Figure out what you want/need to know and go get it. The names aren't as important as the knowledge. EE vs. physics just doesn't matter very much in the real world.
Sometimes you need to adjust the aim of your guns at a new target.

 

[neuroantenna]

Sometimes you need to adjust the aim of your guns at a new target.

I like it