R: Full Threshold

FT {OPI}

R Documentation

Full Threshold

Description

FT begins with a 4-2dB staircase beginning at level est. If the final estimate (last seen) is more than 4dB away from est, a second 4-2 staircase is completed beginning at the estimate returned from the first

Usage

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

FT.start(est = 25, instRange = c(0, 40), makeStim, ...)

FT.step(state, nextStim = NULL)

FT.stop(state)

FT.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 FT returned by `FT.start` and `FT.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 as implemented in the first Humphrey Field Analyzer. The initial staircase starts at est and proceeds in steps of 4 dB until the first reversal, and 2dB until the next reversal. The last seen stimulus is taken as the threshold value. If, after the first staircase, the threshold is more than 4 dB away from the starting point, then a second staircase is initiated with a starting point equal to the threshold found with the first staircase.

Note this function will repeatedly call opiPresent for a stimulus until opiPresent returns NULL (ie no error occurred)

If more than one FT is to be interleaved (for example, testing multiple locations), then the FT.start, FT.step, FT.stop and FT.final calls can maintain the state of the FT after each presentation, and should be used. If only a single FT is required, then the simpler FT can be used. See examples below

Value

Single location

Returns a list containing

npres Total number of presentations
respSeq Response sequence stored as a list of (seen,dB) pairs
first First staircase estimate in dB
final Final threshold estimate in dB

Multiple locations

FT.start returns a list that can be passed to FT.step, FT.stop, and FT.final. It represents the state of a FT at a single location at a point in time and contains the following.

name FT
A copy of all of the parameters supplied to FT.start: startingEstimate=est, minStimulus=instRange[1], maxStimulus=instRange[2], makeStim, and opiParams=list(...).
currentLevel The next stimulus to present.
lastSeen The last seen stimulus.
lastResponse The last response given.
firstStairResult The result of the first staircase (initially NA).
secondStairResult The result of the first staircase (initially NA, and could remain NA).
finished TRUE if staircase has finished (2 reversals, or max/min seen/not-seen twice).
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 FT.step.
responses Vector of responses received (1 seen, 0 not) received at each call to FT.step.
responseTimes Vector of response times received at each call toFT.step.

FT.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.

FT.stop returns TRUE if the first staircase has had 2 reversals, or maxStimulus is seen twice or minStimulus is not seen twice and the final estimate is within 4 dB of the starting stimulus. Returns TRUE if the second staircase has had 2 reversals, or maxStimulus is seen twice or minStimulus is not seen twice

FT.final returns the final estimate of threshold based on state, which is the last seen in the second staircase, if it ran, or the first staircase otherwise

FT.final.details returns a list containing

final The final threshold.
first The threshold determined by the first staircase (might be different from final).
stopReason Either Reversals, Max, or Min which are the three ways in which FT can terminate.
np Number of presentation for the whole procedure (including both staircases if run).

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.

H. Bebie, F. Fankhauser and J. Spahr. "Static perimetry: strategies", Acta Ophthalmology 54 1976.

C.A. Johnson, B.C. Chauhan, and L.R. Shapiro. "Properties of staircase procedures for estimating thresholds in automated perimetry", Investigative Ophthalmology and Vision Science 33 1993.

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)$err))
  stop("opiInitialize failed")

result <- FT(makeStim=makeStim, tt=30, fpr=0.15, fnr=0.01)
if (!is.null(opiClose()$err))
  warning("opiClose() failed")

##############################################
# This section is for multiple FTs
##############################################
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) {
  FT.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, FT.stop)))) {
  i <- which(!st)                         # choose a random,
  i <- i[runif(1, min=1, max=length(i))]  # unstopped state
  r <- FT.step(states[[i]])               # step it
  states[[i]] <- r$state                  # update the states
}

finals <- lapply(states, FT.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()$err))
  warning("opiClose() failed")

[Package OPI version 3.0.2 Index]