Page 2 of 2

Re: Help with Procedure!

Posted: Sat Jun 29, 2013 12:17 am
by Martenzi
Dear Ted,

for the sake of time and effort put into this procedure (for both of us) I wanna ask and reassure a couple of things.

Computer Modeller / Neuroscientist / Student of what?
It came to me that the use of NEURON and this forum may be completely different from how I perceive the intentions of NEURON. I have asked for a couple of things and made some suggestions whereby your response has been very helpful on the programming side and quite confusing on the feedback/suggestions side. (With "side" I refer to the replies to my posts on this forum).

As I understand from my supervisor and your replies, NEURON is a computational engine where anyone may built their theoretical experiment with full control and model different outcomes to their liking.The learning curve is somewhat high for someone like me, no background in computer science except intro to Java and the everyday hobby use of computers.
The learning curve for someone with a mindset for programming is thus likely to be short. However, is NEURON created with the intention of teaching students how to program or learn neuroscience or both?

Now I have learnt some HOC and understand the basic mindset. My project at ANU was to explore a particular cell and play around with NEURON and use some of my ideas to generate some useful data. That project had to be revised to only learn NEURON for the sake of time and effort within my degree. As I dug into NEURON I became quite confused the more I learned. There is so much potential but, in my view, so much time spend by you and students of neuroscience like myself, to teach me the skill to program a tool in order to carry out what I intended to do. The reinventions of the wheel comes to mind. I believe that the traditional view of the new area of computational neuroscience may take this approach since it relies heavily on the skills from computer science. But it seems overkill to me that students are assumed to be or become programmers to carry out computational research. Wouldn't it be enough to turn NEURON into a user friendly Neuroscientist calculator? We dont have to know exactly how to calculate statistics in order to use it? Maybe it is just me.

When I asked myself this I figured two things
(1) either I´m missing something vital in the use of NEURON or
(2) It takes to much time and money to create such a software (if you teach someone how to fish, you dont need to feed them anymore principle)

Thus I could only conclude that Computational Neuroscience requires scientists to explore data through programming. Am I correct?
I could also conclude that NEURONs forum is for computer modellers whereby it is culturally assumed that students has intentions to learn programming in order to carry out research.

Now that is the mistake I made. I was not aware of this. Maybe I´m still wrong about my conclusions.

But then I realised that you are spending loads of time helping out on the forum, which Im grateful for. And fast response considering you have other work to do. I found myself asking: why not turn NEURON into a more user friendly software where you can load a cell and play around with the most common computational features that neuroscientists are using practically and expand as we go along.
I cant see the problematics with this, hence Im asking out of curiosity. It cant be that Im the only seeing this potential?

To my final questions: I asked for a procedure that could generate common data used by neuroscientist to explore synaptic input. I wanted one procedure so I only have to type it once with the numbers and locations that I have. This could be shown directly in a graph and saved to a file automatically. This to me seems like the most obvious thing to do at first sight. Forgive me Ted, I´m really gratefully for all help, but please correct my assumptions and relieve my curiosity.

(You would make my day if you would write a procedure for me that did anything like above described. Why would you bother with that, you think? - Because the time you spend on helping me with code is likely to be more consuming than the time it would take for you to write the hoc. Again, I present my thoughts openly. This is simply because my project is done but I still wanna play with Neuron, but dont have the time to learn how to write the hoc myself). I have suggested before a thread where people are encouraged to post fully functional procedure with description so that others dont need to reinvent the wheel. Im sorry if I step on someones toes or miss the whole idea of NEURON but please have in mind, Im new to all of this.

Since I have graduated my incentive to learn more HOC is entirely based on whether I will proceed with this type of research. However, NEURON is actually one reason why I might avoid it since Im not interested in programming everything to obtain control. Im interested in playing around with the ideas about the cell which unfortunately are lost due to impracticality of making the software work. The GUI is useful when dealing with precise instrumentation, few locations and little mouse effort needed. When I need to change parameters and locations fast to obtain overview of electrophysiological properties, it takes to much time.

I hope you understand my point of view, I may have naive assumptions and taking water of my head but I gotta know why I´m wrong. I am aware that there is nothing so great as a puzzle for a programmer but my puzzle is from the view of neuroscience, to solve it I need the computational data. I am also aware that you are a computer scientist in your background. Since this forum deals with a software for neuroscientists but seems more into programming I might be in the wrong place.

Thank you immensely for you time!

you paper from MIT was a good read , as much as I understood from it.