Hi,
I am trying to use Quest in my 2AFC intact/scrambled task where I am trying to find the shortest possible display duration that the image can be identified accurately. I’m running this using a 75Hz CRT monitor and built my experiment on code. I noticed that the next “intensity” (In my case that would be the next stimulus duration which I specify in frames) that Quest suggests based on the previous trials’ responses are minuscule, i.e if I get a trial right at a duration of 30 frames, the duration on the next trial is 29.7 frames. Because I am using frames, my code rounds the suggested duration up so 29.7 frames effectively becomes 30 frames. Then in the trial after that it goes to 29.5, for example. This makes it so that the quest is sort of “stuck” at a specific duration because the duration it suggests is too small to be functionally different from the previous trial. I would like for the durations that quest puts out to be integers but I’m not sure how to go about doing that. From what I gather on how Quest works, I thought changing the grain parameter to 1 would help, but I realized that it doesn’t change the output of Quest. I then tried using Quest+ where I was able to specify a valid list of durations in integers as intensity_vals and that did not seem to work as expected either.
I guess my question is, is there a way I can specify a step size for Quest or even Quest + like I would if I was using a 1 up 2 down staircase?
Here are the arguments I use for my quest.
start_val=30 (in frames)
start_val_sd=10,
p_threshold=0.71,
gamma=0.5,
n_trials=100,
min_val=1, (in frames)
max_val=30, (in frames)
grain=1
Happy to provide any extra information/clarification!