MOCS: Method of Constant Stimuli (MOCS)

In OPI: Open Perimetry Interface

Method of Constant Stimuli (MOCS)

Description

MOCS performs either a yes/no or n-interval-forced-choice Method of Constant Stimuli test

Usage

MOCS(
  params = NA,
  order = "random",
  responseWindowMeth = "constant",
  responseFloor = 1500,
  responseHistory = 5,
  keyHandler = function(correct, ret) return(list(seen = TRUE, time = 0, err = NULL)),
  interStimMin = 200,
  interStimMax = 500,
  beep_function,
  makeStim,
  stim_print,
  ...
)

Arguments

params	A matrix where each row is x y i n correct_n ll1 ll2 ... llm where x is X coordinate of location. y is Y coordinate of location. i is a location number (assigned by caller). n is Number of times this location/luminance(s) should be repeated. correct_n is the index i of the luminance level (lli) that should be treated as a “correct” response (the correct interval). For a standard MOCS, this will be 1; for a 2AFC, this will be 1 or 2. This number will be in the range [1,m]. lli is the i'th luminance level to be used at this location for interval i of the presentation in cd/\mbox{m}^2. For a standard MOCS, i=1, and the params matrix will have 5 columns. For a 2AFC, there will be two lli's, and params will have 6 columns.
order	Control the order in which the stimuli are presented. "random" Randomise the order of trials/locations. "fixed" Present each row of params in order of 1:nrow(params), ignoring the n (4th) column in params.
responseWindowMeth	Control time perimeter waits for response. "speed" After an average of the last speedHistory response times, with a minimum of responseFloor. Initially responseFloor. "constant" Always use responseFloor. "forceKey" Wait for a keyboard input.
responseFloor	Minimum response window (for any responseWindowMeth except "forceKey").
responseHistory	Number of past yeses to average to get response window (only used if responseWindowMeth is "speed").
keyHandler	Function to get a keyboard input and returns as for opiPresent: list(err={NULL|msg}, seen={TRUE|FALSE}, time. The parameters passed to the function are the correct interval number (column 4 of params), and the result of opiPresent. See Examples.
interStimMin	Regardless of response, wait runif(interStimMin, interStimMax) ms.
interStimMax	Regardless of response, wait runif(interStimMin, interStimMax) ms.
beep_function	A function that takes the string 'correct', the string 'incorrect', or a stimulus number and plays an appropriate sound. See examples.
makeStim	A helper function to take a row of params and a response window length in ms, and create a list of OPI stimuli types for passing to opiPresent. This may include a checkFixationOK function. See Example.
stim_print	A function that takes an opiStaticStimulus and return list from opiPresent and returns a string to print for each presentation. It is called immediately after each opiPresent, and the string is prepended with the (x,y) coordinates of the presentation and ends with a newline.
...	Extra parameters to pass to the opiPresent function.

Details

Whether the test is yes/no or forced-choice is determined by the number of columns in params. The code simply presents all columns from 5 onwards and collects a response at the end. So if there is only 5 columns, it is a yes/no task. If there are 6 columns it is a 2-interval-forced-choice. Generally, an nIFC experiment has 4+n columns in params.

Note that when the order is "random", the number of trials in the test will be the sum of the 3rd column of params. When the order is "fixed", there is only one presentation per row, regardless of the value in the 3rd column of params.

If a response is received before the final trial in a nIFC experiment, it is ignored.

If the checkFixationOK function is present in a stimulus, then it is called after each presentation, and the result is “anded” with each stimulus in a trial to get a TRUE/FALSE for fixating on all stimuli in a trial.

Value

Returns a data.frame with one row per stimulus copied from params with extra columns appended: checkFixation checks, and the return values from opiPresent() (see example). These last values will differ depending on which machine/simulation you are running (as chosen with chooseOpi().

• column 1: x

• column 2: y

• column 3: location number

• column 4: number of times to repeat this stim

• column 5: correct stimulus index

• column 6: TRUE/FALSE was fixating for all presentations in this trial according to checkFixationOK

• column 7...: columns from params

• ...: columns from opiPresent return

References

A. Turpin, P.H. Artes and A.M. McKendrick. "The Open Perimetry Interface: An enabling tool for clinical visual psychophysics", Journal of Vision 12(11) 2012.

See Also

dbTocd, opiPresent

Examples

# For the Octopus 900
# Check if pupil centre is within 10 pixels of (160,140)
checkFixationOK <- function(ret) return(sqrt((ret$pupilX - 160)^2 + (ret$pupilY - 140)^2) < 10)

# Return a list of opi stim objects (list of class opiStaticStimulus) for each level (dB) in
# p[5:length(p)]. Each stim has responseWindow BETWEEN_FLASH_TIME, except the last which has
# rwin. This one assumes p is on old Octopus 900 dB scale (0dB == 4000 cd/m^2).
makeStim <- function(p, rwin) {
  BETWEEN_FLASH_TIME <- 750   # ms
  res <- NULL
  for(i in 5:length(p)) {
    s <- list(x=p[1], y=p[2], level=dbTocd(p[i],4000/pi), size=0.43, duration=200,
              responseWindow=ifelse(i < length(p), BETWEEN_FLASH_TIME, rwin),
              checkFixationOK=NULL)
    class(s) <- "opiStaticStimulus"
    res <- c(res, list(s))
  }
  return(res)
}

################################################################
# Read in a key press 'z' is correct==1, 'm' otherwise
#    correct is either 1 or 2, whichever is the correct interval
#
# Return list(seen={TRUE|FALSE}, time=time, err=NULL))
#        seen is TRUE if correct key pressed
################################################################
## Not run: 
  if (length(dir(".", "getKeyPress.py")) < 1)
    stop('Python script getKeyPress.py missing?')

## End(Not run)

keyHandler <- function(correct, ret) {
  return(list(seen=TRUE, time=0, err=NULL))
  ONE <- "b'z'"
  TWO <- "b'm'"
  time <- Sys.time()
  key <- 'q'
  while (key != ONE && key != TWO) {
    a <- system('python getKeyPress.py', intern=TRUE)
    key <- a # substr(a, nchar(a), nchar(a))
    print(paste('Key pressed: ',key,'from',a))
    if (key == "b'8'")
      stop('Key 8 pressed')
  }
  time <- Sys.time() - time
  if ((key == ONE && correct == 1) || (key == TWO && correct == 2))
    return(list(seen=TRUE, time=time, err=NULL))
  else
    return(list(seen=FALSE, time=time, err=NULL))
}

################################################################
# Read in return value from opipresent with F310 controller.
# First param is correct, next is 1 for left button, 2 for right button
# Left button (LB) is correct for interval 1, RB for interval 2
#    correct is either 1 or 2, whichever is the correct interval
#
# Return list(seen={TRUE|FALSE}, time=time, err=NULL))
#        seen is TRUE if correct key pressed
################################################################
F310Handler <- function(correct, opiResult) {
  z <- opiResult$seen == correct
  opiResult$seen <- z
  return(opiResult)
}

################################################################
# 2 example beep_function
################################################################
## Not run: 
  require(beepr)
  myBeep <- function(type='None') {
    if (type == 'correct') {
      beepr::beep(2)  # coin noise
      Sys.sleep(0.5)
    }
    if (type == 'incorrect') {
      beepr::beep(1) # system("rundll32 user32.dll,MessageBeep -1") # system beep
      #Sys.sleep(0.0)
    }
  }
  require(audio)
  myBeep <- function(type="None") {
    if (type == 'correct') {
      wait(audio::play(sin(1:10000/10)))
    }
    if (type == 'incorrect') {
      wait(audio::play(sin(1:10000/20)))
    }
  }

## End(Not run)

################################################################
# An example stim_print function
################################################################
## Not run: 
  stim_print <- function(s, ret) {
    sprintf("%4.1f %2.0f",cdTodb(s$level,10000/pi), ret$seen)
  }

## End(Not run)

OPI documentation built on Nov. 7, 2023, 9:06 a.m.