timed task loop, regardless of read time

kessler
New Member

Posts: 2

timed task loop, regardless of read time Sept 9, 2020 2:12:37 GMT

Quote

Post by kessler on Sept 9, 2020 2:12:37 GMT

I am hoping to create an asynchrony test, wherein the participant will tap along with a stimulus (visual, auditory, vis+aud, melody).

I want the stimulus to repeat every 500 ms, and I want to record the RT but not allow the RT to affect the duration of the loop. I only see "delay" and "readkey" with optional timeout. Is there a way to time the loop, rather having it depend on results?

Thank you.

PsyToolkit
Administrator

Posts: 545

timed task loop, regardless of read time Sept 9, 2020 11:32:40 GMT

Quote

Post by PsyToolkit on Sept 9, 2020 11:32:40 GMT

Yes, that can be done in different ways. Let's say, you want the loop to be 30 seconds (30000 ms), then you could do something like the code below (which is actually a working example, you can just copy and paste, compile, and run it).

There is no while loop at all. I think it can be much easier. Note code below is in blue and the comment in grey (in PsyToolkit script, comments are texts followed by hash mark).

In this code, I assume that you want to show the stimulus less than 500 ms, for example, only 200 ms. In this case, what we do here is that we do two instructions of "readkey"

First for the first 200 ms, and if the participant has not pressed the required key (in this instance the key 'b'), then you wait longer.

The trick here is to use time stamps. If 30000 ms have passed, the last list ends. Thus, the number 10000 is just a high number (it could be lower, because realistically, a fast tapper would be able to tap maybe 60 times, never 10000 of course).

task tapping
   keys b
   show text "keep tapping with key b for 30 seconds" 0 -200 100 100 100
   show rectangle 0 0 50 50 0 255 0 # show a green rectangle at screen center
   readkey 1 200 # maximum time is 200 ms, it is actually difficult to press so quickly
   if STATUS != TIMEOUT
     clear -1
     set $reaction_time RT
   fi
   if STATUS == TIMEOUT
     readkey 1 200
     clear -1
     set $reaction_time expression RT + 200 # we add 200, because people did not press first time
   fi
   set $remaining_time expression 500 - $reaction_time
   timestamp the_time_now # update a time stamp every tap
   set $time_so_far timestamp-diff this_is_start_time the_time_now
   save $time_so_far $reaction_time STATUS
   delay $remaining_time
   if $time_so_far > 30000
   end tasklist # this jumps out of the tasklist when time limit of 30 seconds passed
   fi

block test
timestamp this_is_start_time # set a time stamp start point
tasklist
    tapping 10000
end

kessler
New Member

Posts: 2

timed task loop, regardless of read time Sept 16, 2020 0:53:10 GMT

Quote

Post by kessler on Sept 16, 2020 0:53:10 GMT

Thank you for your speedy reply. I've played a bit with the above code, and modified it (see below), a bit, but there are a several problems.

1. The loop loses time, each cycle. If it was regular, I could take an average and shorten the loop time, but it is not.

2. In related news,the flashes do not seem perfectly regular. It would probably be more obvious with tones. Further, if I "misbehave" and tap very fast (multiple times per cycle), it very obviously affects the flashing stimulus.

3. I cannot capture "anticipatory" clicks.

4. I cannot capture all of the clicks.

I'm not entirely certain that you understood what I want, and I'm beginning to wonder whether this is possible. I hope I'm missing something.

I want to be able to catch response time before or after the stimulus (before = anticipating, because the stimulus is at a regular frequency). Because the readkey function only captures the RT after, and then won't capture a second click, if it's within the same timeout period (which would be an anticipatory click), there are "timeouts" that aren't really missed clicks.

I'm attaching spreadsheets with a couple of data sets. In the second, I ran the script with 1000-ms cycles, and counted clicks. In each of these, I clicked 28 or 29 times during the 30 seconds of flashing, but the data only shows ~21. I believe that these cases are when I was wrapping around from anticipatory to responsive, and therefore there were more than 1000 ms between my clicks. This is borne out by the often short RT values immediately following a timeout, but does not capture all of the clicks. Also, it's not always a short RT.

I started to theorize that the more timeouts there were, the greater would be the looptime delay, but that also is not always the case.

I am pretty stumped as to how to proceed, and whether this will even work. Your input is greatly appreciated.

task tapping
keys space
show text "Use the spacebar to tap along with the beat for 30 seconds" 0 -200 100 100 100
show rectangle 0 0 50 50 0 255 0 # show a green rectangle at screen center
timestamp Stim
set $stim_time timestamp-diff this_is_start_time Stim
readkey 1 50 # maximum time is 50 ms
if STATUS != TIMEOUT
clear -1
set $reaction_time RT
fi
if STATUS == TIMEOUT
clear -1
readkey 1 450
set $reaction_time expression RT + 50 # we add 50, because key not pressed before first timeout
fi
set $remaining_time expression 500 - $reaction_time
timestamp the_time_now # update a time stamp every tap
set $time_so_far timestamp-diff this_is_start_time the_time_now
save $stim_time $time_so_far $reaction_time STATUS $remaining_time
delay $remaining_time
if $time_so_far > 30000
end tasklist # this jumps out of the tasklist when time limit of 30 seconds passed
fi

block test
timestamp this_is_start_time # set a time stamp start point
tasklist
tapping 10000
end

Attachments:

ptk_asynchrony_debuggung.xlsx (25.7 KB)

ptk_asynchrony_debuggung_1Hz.xlsx (35.35 KB)

PsyToolkit
Administrator

Posts: 545

timed task loop, regardless of read time Sept 16, 2020 7:38:56 GMT

Quote

Post by PsyToolkit on Sept 16, 2020 7:38:56 GMT

Hi, this is probably going too deep detail for the forum, I will be in touch with you via email to solve this problem. It should definitely be possible. It is great you have such detailed data, let's try to make this work.