fourTwo: 4-2 Staircase

In OPI: Open Perimetry Interface

4-2 Staircase

Description

fourTwo is a 4-2 dB staircase beginning at level est terminating after two reversals. The final estimate is the average of the last two presentations. It also terminates if the minStimulus is not seen twice, or the maxStimulus is seen twice.

Usage

fourTwo.start(est = 25, instRange = c(0, 40), verbose = FALSE, makeStim, ...)

fourTwo.step(state, nextStim = NULL)

fourTwo.stop(state)

fourTwo.final(state)

Arguments

est	Starting estimate in dB
instRange	Dynamic range of the instrument c(min,max) in dB
verbose	True if you want each presentation printed
makeStim	A function that takes a dB value and numPresentations and returns an OPI datatype ready for passing to opiPresent
...	Extra parameters to pass to the opiPresent function
state	Current state of the fourTwo returned by fourTwo.start and fourTwo.step
nextStim	A valid object for opiPresent to use as its nextStim.

Details

This is an implementation of a 4-2 1-up 1-down staircase. The initial staircase starts at est and proceeds in steps of 4 dB until the first reversal, and 2dB until the next reversal. The mean of the last two presentations is taken as the threshold value. Note this function will repeatedly call opiPresent for a stimulus until opiPresent returns NULL (ie no error occured). If more than one fourTwo is to be interleaved (for example, testing multiple locations), then the fourTwo.start, fourTwo.step, fourTwo.stop and fourTwo.final calls can maintain the state of the fourTwo after each presentation, and should be used. See examples below.

Value

Multilple locations

fourTwo.start returns a list that can be passed to fourTwo.step, fourTwo.stop, and fourTwo.final. It represents the state of a fourTwo at a single location at a point in time and contains the following. * name, fourTwo. * startingEstimate=est, input param. * currentLevel, the next stimulus to present. * minStimulus=instRange[1], input param. * maxStimulus=instRange[2], input param. * makeStim, input param. * lastSeen, the last seen stimulus. * lastResponse, the last response given. * stairResult, The final result if finished (initially NA). * finished, "Not" if staircase has not finished, or one of "Rev" (finished due to 2 reversals), "Max" (finished due to 2 maxStimulus seen), "Min" (finished due to 2 minStimulus not seen). * verbose, number of reversals so far. * numberOfReversals, number of reversals so far. * currSeenLimit, number of times maxStimulus has been seen. * currNotSeenLimit, number of times minStimulus not seen. * numPresentations, number of presentations so far. * stimuli, vector of stimuli shown at each call to fourTwo.step. * responses, vector of responses received (1 seen, 0 not) received at each call to fourTwo.step. * responseTimes, vector of response times received at each call to fourTwo.step. * opiParams=list(...), input param

fourTwo.step returns a list containing * state, the new state after presenting a stimuli and getting a response. * resp, the return from the opiPresent call that was made.

fourTwo.stop returns TRUE if the staircase is finished (2 reversals, or maxStimulus is seen twice or minStimulus is not seen twice).

fourTwo.final returns the final estimate of threshold (mean of last two reversals). This issues a warning if called before the staircase has finished, but still returns a value.

See Also

dbTocd, opiPresent, FT

Examples

# Stimulus is Size III white-on-white as in the HFA
makeStim <- function(db, n) {
  s <- list(x=9, y=9, level=dbTocd(db), size=0.43, color="white",
            duration=200, responseWindow=1500)
  class(s) <- "opiStaticStimulus"
  return(s)
}
chooseOpi("SimHenson")
if (!is.null(opiInitialize(type="C", cap=6)))
  stop("opiInitialize failed")

##############################################
# This section is for multiple fourTwos
##############################################
makeStimHelper <- function(db,n, x, y) {  # returns a function of (db,n)
  ff <- function(db, n) db+n
  body(ff) <- substitute({
    s <- list(x=x, y=y, level=dbTocd(db), size=0.43, color="white",
              duration=200, responseWindow=1500)
    class(s) <- "opiStaticStimulus"
    return(s)}, list(x=x,y=y))
  return(ff)
}
# List of (x, y, true threshold) triples
locations <- list(c(9,9,30), c(-9,-9,32), c(9,-9,31), c(-9,9,33))

# Setup starting states for each location
states <- lapply(locations, function(loc) {
  fourTwo.start(makeStim=makeStimHelper(db,n,loc[1],loc[2]),
                tt=loc[3], fpr=0.03, fnr=0.01)})

# Loop through until all states are "stop"
while(!all(st <- unlist(lapply(states, fourTwo.stop)))) {
  i <- which(!st)                         # choose a random, 
  i <- i[runif(1, min=1, max=length(i))]  # unstopped state
  r <- fourTwo.step(states[[i]])               # step it
  states[[i]] <- r$state                  # update the states
}

finals <- lapply(states, fourTwo.final)    # get final estimates of threshold
for(i in 1:length(locations)) {
  cat(sprintf("Location (%+2d,%+2d) ",locations[[i]][1], locations[[i]][2]))
      cat(sprintf("has threshold %4.2f\n", finals[[i]]))
}

if (!is.null(opiClose()))
  warning("opiClose() failed")

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