1. PF Contest - Win "Conquering the Physics GRE" book! Click Here to Enter
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Masters doubt regarding computational neuroscience career

  1. Apr 30, 2013 #1
    I'm a biomedical engineering student interested in computational neuroscience, "the neural code" and systems neuroscience to be more concrete. The problem is in my country Mexico, master's degrees either focus in the mathematical side or the biological. What would be better to study a master's in dynamical systems and control or in Probability and statistics (studying the biology on my own) and then try to get into a Ph.D in neurobiology or a master's and Ph.D in neurobiology and study the math on my own.
  2. jcsd
  3. Apr 30, 2013 #2
    That would be my suggestion. The key word here is "computational" neuroscience, which includes neuroinformatics. That means you're most likely going to be dealing largely with abstract neural network architectures rather than experimental brain "wet-ware" physiology. So the maths are going to be primary.

    If you want an idea of what kind of maths your going to need for that effort, try reading some of Karl Friston's work. I don't think you really want to try learning that on your own if you can avoid it. I went the second route you talk about and am finding it more difficult than I thought to catch up on the math with self-study.

    You can learn most of the neuro-bio stuff on your own, although I would certainly take neuroanatomy lab and physiology lab. That stuff you can't really learn very well on your own from a textbook. It needs to be hands-on.

    Good luck.
  4. Apr 30, 2013 #3
    Thanks a lot, I just got myself started with non-linear dynamics I'm reading Strogatz I've looked at some papers and the math looks quite daunting at first. For what I've seen so far it looks like those trying to model neural networks are people involved with non-linearity and chaos, and those trying to "decode the neural code" are people more involved with statistics, probability and information theory, so I think that what is left is to decide whether to pursue dynamical systems or probability and statistics.

    About the biology it seems that they don't get involved a lot with al the biochemistry and biological issues, am I wrong?

    I have taken a course on neurophisiology in which I was able to do an immunomicroscopy, and I spoke with the professor to do a dissection later.

    Thanks a lot
  5. Apr 30, 2013 #4
    You will need both approaches, and I know many in the field. Coursera covers many of the different facets in a top up manner, so just pick the various bits there. you will find computational neuroscience is constantly shifting right now.... so its better to recruit whatever university facilities to get a solid grounding in physics, maths for just about any area. If they have a big computer center get hands on just writing programs on it. lab skills are if you want to work in the human brain project type comp-neuro or other projects which require you to acquire biological data.

    In regards to the knowledge of the bio stuff, comp neuro at the cutting edge is all over the place, but you wont hear about that at the general level, where they want to fixate on spiking neurons. There was a recent study where human level neuron to astrocyte ratios were transplanted into mice and the mice memory went right up a gear. I was privvy to the confusion in the cutting edge of the comp neuro community. In that one day everybody was scrabbling about for depth on glial cells, how they worked, who were the prominent researchers etc...

    The same for ephatic fields in neurons (find Markrams recent paper.. human brain project)..etc etc. Another flurry of interest in Calcium diffusion resulted. So the point is try to keep a very broad neuro overview despite a lot of fixation in building spiking systems right now. An idea would be to check out the IBM tissue simulator and the human brain projects depth. They realize there are many unknowns which are going to happen, so they are building the computational systems to try and be as general as possible.
  6. Apr 30, 2013 #5
    Thanks a lot I didn't know about the ephaptic coupling quite interesting, it seems like a pretty challenging field, with a lot of approaches and "impossible" to gather all the knowledge needed in just a master and a PhD. I like that :P I think I would go with the probability and statistics master's
  7. May 1, 2013 #6
    I think everything you said there is basically correct. Looks like you got a good handle on it. I am active in this field and therefore am a little biased, but I think the dynamical systems framework embodies the best approached to modeling brain networks (even decoding the "neural code," as you say), and eventually the only autonomous and intelligent systems constructions will utilize chaotic neurodynamics in their implementation. Again, that is a personal bias, so take it for what you will.

    I recommend visiting the CLION site at U Memphis,


    and reviewing some of the papers by Freeman, Bressler, Kozma, and others they reference for further perusal. I am actually trying to get into that program myself as a PhD candidate as I personally know a few of the guys there. Again, I'm feverishly trying to "recharge" my math skills because the department I'd be going into is "applied math," and my skills are nowhere close to applied math PhD caliber. However, my background is in evolutionary neurobiology, physical anthropology, and related fields, which most of the guys there are relatively weak in, so its a good complement (I hope). We'll see. One goal of the CLION group is to develop autonomous robotic systems such as planetary rovers, etc. So the challenges there are diverse and complex and reach outside of basic neural netwrok interactions into sensors, interfaces, cognitive mapping, kinesthetics, etc. So, a lot of answering your question relates specifically to what you want to do in computational neuroscience. What were your plans?
  8. May 1, 2013 #7
    Well about my plans that's kind of a big question, I got interested in a lot of things from computational neuroscience to cybernetics, this particular research blew my mind , I visited the CLION site and it looks like an amazing lab with a lot of interdisciplinary work good luck!!!

    In what refers to computational neuroscience studying "the neural code" and systems neuroscience to further applications in BCI and helping paraplegic people.

    And cybernetics my thoughts were more of the sci-fi style trying to make robots like the one I posted and to keep alive neuron cells outside the human body.

    I still have a long way to go and plenty of reading to do in order to clear my mind and decide what my main focus will be so far I think that a master's in dynamical systems and control would span more of my interests.
    Last edited by a moderator: Sep 25, 2014
  9. May 1, 2013 #8
    Thanks, good luck to you too. If you're interested in BCI, here's an article that combines neurodynamics and BCI that you may be interested in:


    Its a free article, just follow the link, and you can also access part two from the same page (upper right link).
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook