+ Reply to Thread
Results 1 to 1 of 1

Thread: A couple of questions about results in computational neuroscience

  1. #1
    Points: 6, Level: 1
    Level completed: 11%, Points required for next Level: 44

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    A couple of questions about results in computational neuroscience




    Hi guys,

    This is my first time posting here.

    I am a medical student and I have been conducting research in computational neuroscience over the past few months and I'm not all that certain what I can do with my results apart from simply describing them.

    Effectively what I've done is run computer simulations and looked at the power of oscillations of a certain frequency that they have produced. What I did was change one particular variable in the simulation to 17 different values and then read out the oscillatory power. So basically one set of simulations would produce a data set of 17. I then did this about 10 times to obtain 10 sets of 17 data points.

    The results look a little something like this http://imgur.com/8EjcY54 (also attached just in case imgur is over capacity)

    Now I basically have a couple of questions

    1. For a data set where there is no obvious transformation that can be done to linearise it, what kind of regression is appropriate? Or is it a matter of trying out different functions to find one with the highest R^2 value? I used the Curve Fitting Tool in MATLAB to fit a 3 term gaussian function in the example above.

    2. If I didn't do regression, what kind of things could I do with my data to say that there is a statistically significant difference between the y values I obtain for an x value of 1 versus an x value of 2? I was considering plotting the mean and then error bars but that leaves me with the questions - what kind of error bars? At each of those x values, the y values are not normally distributed so I can't do something like standard error of the mean.

    I would appreciate any help. I can gain access to things like STATA and SPSS, but ideally methods which I could implement in Excel or MATLAB would be preferred.

    Thanks all
    Attached Images  
    Last edited by ilzwalrus; 07-09-2015 at 02:01 PM.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats