Bayesian Inference for Psychometric Functions. Kuss, Jakel, Wichmann. Journal of Vision, 2005

  1. Main goal is to do Bayesian Inference on subject data to estimate parameters of their psychometric functions (relationship between stimulus and ability of subject to discriminate between stimuli)
  2. Because the Bayesian inference here cannot be performed analytically, MCMC is used to sample from the posterior distribution over the parameters to the function describing the psychometric curve
  3. Also allows  for confidence intervals and other things to be extracted from the posterior
  4. Traditional methods uses ML methods and bootstrapping
  5. Because psychophysical experiments are tiring for subjects, methods exist that find only a single pt on the psychometric function somewhere between 50% and 90% performance, called the threshold
    1. Called adaptive methods, exist as parametric and nonparametric
    2. In theory, the data from this could be used to estimate the entire psychometric function, but it generally doesn’t work well because the methods only try to make an estimate for a point, as opposed to the whole curve
  6. There are also methods that attempt to estimate the entire function, but relatively few that also build confidence intervals
  7. The psychometric function is commonly assumed to have a sigmoidal shape
  8. <Not sure if I finished this paper>

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: