Skip to content

vlt.neuro.vision.contrast.indexes.fitindexes

  FITINDEXES - compute contrast fit index values from a response structure

  FI = vlt.neuro.vision.contrast.index.fitindexes(RESPSTRUCT, FITTERMS)

   RESPSTRUCT is a structure  of response properties with fields:
  |------------------------------------------------------------------|
  | Field     | Description                                          |
  |------------------------------------------------------------------|
  | curve     |   4xnumber of contrasts tested,                      |
  |           |     curve(1,:) is contrasts tested                   |
  |           |     curve(2,:) is mean response at each contrast     |
  |           |     curve(3,:) is standard deviation at each contrast|
  |           |     curve(4,:) is standard error at each contrast    |
  | ind       |   cell list of individual trial responses for each   |
  |           |     contrast                                         |
  | blankresp |   1x3 vector of mean control response, mean control  |
  |           |     response standard deviation, and mean control    |
  |           |     response standard error.
  |------------------------------------------------------------------|

  FITTERMS is the number of terms to include in the Naka-Rushton Fit
   (R = RM * C.^N/(C.^(N*S)+c50^.(N*S))

  |------------------------------------------------------------------|
  | FITTERMS  | Description                                          |
  |------------------------------------------------------------------|
  |    2      | Uses only RM and C50 (N and S are 1)                 |
  |    3      | Uses only RM, C50, and N (S is 1)                    |
  |    4      | Uses RM, C50, N, and S                               |
  |------------------------------------------------------------------|

  Returns a structure FI with fields:

  |------------------------------------------------------------------|
  | Field                  | Description                             |
  |------------------------------------------------------------------|
  | fit_parameters         | [RM C50 N S]  Naka Rushton fit params   |
  |                        |   N and S are left off if FITTERMS<3,4  |
  | fit                    | 2x101 matrix; first row is contrasts    |
  |                        |   0:0.01:1; second row is fit of        |
  |                        |   responses for those contrasts         | 
  | r2                     | Fit r^2 value                           |
  | empirical_C50          | Empirical fit contrast that gives 50% of|
  |                        |   response to 100% contrast (differs    |
  |                        |   from C50 of the fit b/c max response  |
  |                        |   is not usually at 100% contrast)      |
  | relative_max_gain      | Relative maximum gain (Heimel et al.    |
  |                        |   2005)                                 |
  | saturation_index       | Saturation index (only meaningful for   |
  |                        |   fits including S but provided always) |
  | sensitivity            | 1x10 vector of contrast sensitivity with|
  |                        |   criteria of [1 ... 10] standard       |
  |                        |   deviations of the control response.   |
  |                        |   If the standard deviation of the      |
  |                        |   control response is 0, then it is     |
  |                        |   recalculated assuming the response is |
  |                        |   1 unit on one trial.                  |
  |------------------------------------------------------------------|