It’s nothing special- it runs in Ruby Shoes and I can email you the source code if it would be at all interesting. I wrote it after reading a number of papers describing sleep dep experiments that used it.
It shows a stimulus (increasing time in red in the center of a black background), then once you click in the upper left corner it waits 2-10 seconds (chosen at random) and then shows you another stimulus. This goes on for ten minutes. You get a host of reaction times (and so can watch that change) but the real value (and the reason it’s boring) is that your ability to sustain concentration drops with sleep deprivation, and so lapses (where it takes you at least twice as long as your baseline, typically .5 to .8 seconds) start showing up. They become more frequent and longer the more sleep deprived you get (at my worst, it took me about fifteen seconds to respond to a stimulus; I didn’t code it but the paper I took the parameters from had a buzzer sound if you went 30 seconds without a response). It outputs the times (absolute times of stimuli, responses, and false starts) to a text file, and I never wrote any analysis code because I considered the affiliated sleep experiment a failure.
It’s also a very good proxy for driving- it involves a lot of staring at boring, barely changing views, and then when one gets into an accident, it’s typically something that noticing it coming a half-second earlier could have helped a lot with. And so when you find that a half-second of inattention goes from a <1% of the time thing to a 5% of the time thing, you swear off the heavy machinery.
There are definitely benefits of a simple arithmetic test, I just didn’t want you to underestimate the benefits by measuring the wrong thing.
Tell me more about your PVT program. I should probably try something like that.
It’s nothing special- it runs in Ruby Shoes and I can email you the source code if it would be at all interesting. I wrote it after reading a number of papers describing sleep dep experiments that used it.
It shows a stimulus (increasing time in red in the center of a black background), then once you click in the upper left corner it waits 2-10 seconds (chosen at random) and then shows you another stimulus. This goes on for ten minutes. You get a host of reaction times (and so can watch that change) but the real value (and the reason it’s boring) is that your ability to sustain concentration drops with sleep deprivation, and so lapses (where it takes you at least twice as long as your baseline, typically .5 to .8 seconds) start showing up. They become more frequent and longer the more sleep deprived you get (at my worst, it took me about fifteen seconds to respond to a stimulus; I didn’t code it but the paper I took the parameters from had a buzzer sound if you went 30 seconds without a response). It outputs the times (absolute times of stimuli, responses, and false starts) to a text file, and I never wrote any analysis code because I considered the affiliated sleep experiment a failure.
It’s also a very good proxy for driving- it involves a lot of staring at boring, barely changing views, and then when one gets into an accident, it’s typically something that noticing it coming a half-second earlier could have helped a lot with. And so when you find that a half-second of inattention goes from a <1% of the time thing to a 5% of the time thing, you swear off the heavy machinery.