Millisecond Forums

Version 5.0.0.10 mouseover issues

https://forums.millisecond.com/Topic22814.aspx

By AKrishna - 11/6/2017

Hi all,

I just updated from version 5.0.0.5 to the latest version and an experiment that previously worked is now unfortunately broken. I have already found a workaround, so this issue isn't urgent, but I thought I'd post anyway in case this is a bug of interest to the devs.

The issue centers around a task in which participants click a cross in the center of the screen, then move the mouse to the left or right (free choice). Based on their choice, a picture is then displayed. If they don't move the mouse within 2000ms, they get a "too slow" feedback.

I coded this by creating a fixation trial:

<trial FixationToTrain>
/ stimulustimes = [0 = fixation]
/ validresponse = (fixation)
/ branch = [trial.InstrTraining]
/ inputdevice = mousekey
/ showmousecursor = true
/ recorddata = false
</trial>

.. which branches to the trial that collects the mouseover response:

<trial InstrTraining>
/ stimulustimes = [0 = respondnow, LEFT, RIGHT]
/ isvalidresponse = [(trial.InstrTraining.response == values.negdirection || trial.InstrTraining.response == values.posdirection)]
/ iscorrectresponse = [(trial.InstrTraining.response == values.posdirection)]
/ branch = [if (trial.InstrTraining.latency == 2000) trial.timeouttrialInstrTrain]
/ branch = [if (trial.InstrTraining.response == values.negdirection) trial.negpic]
/ branch = [if (trial.InstrTraining.response == values.posdirection) trial.pospic]
/ timeout = 2000
/ inputdevice = mouseover
/ showmousecursor = false
</trial>

<text respondnow>
/ items = (" ")
/ fontstyle = ("Arial", 10%)
</text>

<shape LEFT>
/ shape = rectangle
/ size = (10%,100%)
/ position = (values.stimwidth,50)
/ color = white
</shape>

<shape RIGHT>
/ shape = rectangle
/ size = (10%,100%)
/ position = (100-values.stimwidth,50)
/ color = white
</shape>

... which then branches to a picture display trial when the participant hits one of the two shapes with the hidden mouse cursor (or "too slow" feedback if they don't). values.negdirection and values.posdirection are assigned earlier in the experiment via subject ID to be LEFT or RIGHT (case sensitive).

As written here, the code worked in v.5.0.0.5, but in 5.0.0.10, the trial fails to register any response and consequently always moves to the "too slow" message. I've already verified that the mouse is moving correctly (by unhiding the mouse cursor) and that the shapes are where they should be (by making them red and looking). Replacing the /isvalidresponse with /validresponse = (values.negdirection, values.posdirection) doesn't help, but /validresponse = (LEFT, RIGHT) does (which is my current workaround). From what I can tell using the monkey, the /iscorrectresponse attribute resolves correctly.

Anyone have any ideas? Thanks for your time!