# Method of Constant Stimuli (MOCS) MOCS performs either a yes/no or n-interval-forced-choice Method of Constant Stimuli test ```r 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/$m^2$. 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`: `err`, `seen` and `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. ## Returns 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 ## 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. ## Examples ```r # 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) ``` ## 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`

MOCS function