Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Bird eyes - and how to model them!

  1. Oct 15, 2013 #1
    OK so I am not sure if this should be on a physics forum, a maths forum or a biology forum. All I know is I am stuck.

    A little back ground: Researchers have modelled how various animals can see depending on their visual sensor cells, namely cones (colours vision) and rods (luminance).

    What I understand is that I am supposed to use my light spectrophotometer reflectance data and run a discrimination model to get JNDs (Just Noticeable Differences), the measurements used to discriminate weather or not a bird can tell the difference between either colours (using the models used for comes) or luminance (rod models). This is usually based on a discrimination factor (often Weber fraction 0.05), and relative proportions of vision cone types in predator's retina. The results from discrimination model are JND's.

    As far as I understand Vorobyev & Osorio 1998, Osorio & Vorobyev 2005 (Attached :)) are probably the most influential papers defining foundations of vision modelling. But seriously, having read them through approximately 4 or 5 times over the last 3 weeks I am no further forward in actually opening excel (or the software needed) and actually making a model. I don't even know if 'running a model' is metaphoric or I need to open R or similar and run something!!!

    Any help would be greatly appreciated.

    Kind Regards

  2. jcsd
  3. Oct 15, 2013 #2

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper

    Start with the end in mind: what is your model supposed to be telling you?
    What will it be useful for - specifically?

    What is your level of physics and math here?

    Note: excell is unlikely be up to the task.
  4. Oct 15, 2013 #3
    The model is supposed to tell me what the JNDs are for my two treatments, from this I can tell if a bird can discriminate between two visual signals.

    It will be and has been used for putting spectrophotometer data into an ecologically relative context.

    My personal level? not great but I need to complete this analysis so I really have no choice but to figure it out. I know people who are less mathematically literate than I am have completed this analysis and I am finding it incredibly frustrating. Equations scare, I will not lie. I do however want to learn how to solve them..and not be scared any more!!
  5. Oct 15, 2013 #4
    Point taken on excel!!
  6. Oct 15, 2013 #5

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper

    Have you done this kind of modelling before - say for different sorts of eyes?
    i.e. what is the holdup? Where do you get stuck?
  7. Oct 15, 2013 #6
    No, I have not done this before or anything like it. I do not have a clue where to start!

    I have my Photospec. data that is it. I just don't get the concept of running a model or where to get the numbers to put into an said model. Just imagine someone with zero confidence in maths and equations.

    But with a health, almost unhealthy appetite to learn!
  8. Oct 15, 2013 #7


    User Avatar

    Staff: Mentor

    Is this for some kind of research project? What is the end goal? How did you end up with this project, if you don't have the modeling experience yet?
  9. Oct 15, 2013 #8
    Fair question!

    I am a PhD. student learning my trade so to speak! I guess as in all research, the end goal is to publish! I am a biologist, working with iridescent insects (see some other posts). I would like to model my light spectro. data with their potential predators! i.e. birds and specifically bluetits as most research on avian visual modelling has been done with this species.

    I was not being vague on purpose, I just didn't think the information was needed. Apologies, I have no mysterious motives.

    Kind Regards

  10. Oct 15, 2013 #9

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper

    So you know how eyes work and you have some standard model for rods and cones?
    iirc different species of bird have different vision characteristics.
  11. Oct 16, 2013 #10
    Yes, there is a standard model for blue tits (and various other species) for both rods and cones. The birds have 4 cone types, as ours plus U.V. sensitive. The models are in the attached papers!
  12. Oct 16, 2013 #11

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper

    So what's the problem?
  13. Oct 16, 2013 #12
    How do I implement the (any) model?
    Where do I start to understand it enough to get out what I want?
    What is a discrimination model?
  14. Oct 16, 2013 #13

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper

    Oh I see - you have the model but you have no idea what it means, and your math is not up to the task.

    It's not trivial. You will need to bring yourself up to speed in a new field and get familiar with the kid of math needed.

    You need to look up the citations in the paper you get your model from to start with - explore the literature.
    This is part of the PhD work you need to learn - there is no royal road here. It's also much bigger than what we can deal with in these forums. Since you know others who have used these models, why not ask them for pointers? Networing and consultation are an important part of the scientific process.

    You also need a clearer idea of what the end result is supposed to be.

    Per your posts:
    "running the model" is, indeed, a metaphor for using the equations - the math - that is used to describe the model. You need to implement the model equations in some way that makes sense for your desired end result.
    The authors should have provided examples or references to examples.

    You should have experience with other examples of mathematical models - like predator-prey models, or cell growth models, stuff like that, in your undergrad work?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook