Stop-Signal Task (SSRT) — Free Impulse Control & Response Inhibition Test

The stop-signal task measures your ability to cancel an action already under way, that is, response inhibition. The basic task is to press the matching button within 1.5 seconds whenever a left or right arrow appears; on some trials a red border flashes around the screen shortly after the arrow, and then the job is to hold back the press you were about to make.

The key index is stop-signal reaction time (SSRT), an estimate of the internal processing time between seeing the stop signal and actually cancelling the action. It puts a number on the speed of an inhibitory process that is otherwise invisible. The theoretical basis is the race model of Logan and Cowan (1984), in which a go process and a stop process race and whichever finishes first wins.

The moment the stop signal appears (the stop-signal delay, SSD) is not fixed but tracks your performance. After a successful stop the next signal comes 50 ms later; after a failed stop it comes 50 ms earlier (starting at 200 ms, within a 50 to 1,100 ms range). This staircase pushes your stop success rate toward about 50%, which is the condition under which SSRT can be estimated reliably.

After 8 practice trials (2 of them stop trials) come 64 main trials, 16 of which (25%) are stop trials. The stop success rate at the top of the results is by design expected to sit near 50%; that is neither good nor bad, it simply means the difficulty tracking worked. If it strays far from 50% (outside 25 to 75%), the report flags the SSRT estimate as low-reliability.

SSRT is computed with the integration method. All go-trial reaction times are sorted fastest to slowest (trials with no response are entered as 1,500 ms), the RT at the rank matching your stop-failure proportion is taken, and the mean SSD is subtracted from it. If you stopped on every stop trial, or responded on every one, SSRT cannot be computed and is not shown.

The bands read as follows: under 200 ms very fast inhibition, 200 to 279 ms the typical adult range (roughly 200 to 280 ms), 280 to 349 ms somewhat slow, and 350 ms or more on the slow side. Go accuracy, mean go reaction time, omissions and mean SSD are shown too; keep in mind that slow go responses or many omissions can distort SSRT.

The most common distortion is strategic waiting, that is, deliberately slowing down in anticipation of a stop signal. That lengthens go reaction times and makes SSRT look better than it really is. The right approach is to press immediately when an arrow appears and hold back only when the red border shows. Sleep loss, the day after drinking and fatigue slow inhibition, while caffeine can push either way depending on the person.

Touch-input latency is baked into both go RT and SSRT, so switching devices can shift the numbers. SSRT tends to lengthen with age, and ADHD research repeatedly reports longer SSRT at the group level, but individual variation is wide enough that a single person's value cannot settle anything on its own.

The basic use is trend watching. Repeat every 3 to 4 weeks under the same conditions and compare only the results without a low-reliability flag. The output is a self-check, not a diagnosis, but if SSRT repeatedly comes out at 350 ms or more even on good days, and you often find it hard to stop impulsive slips in speech or action, it is worth raising at a psychiatric appointment.