A two-choice arrow task in which a stop signal sometimes follows the arrow. Measures action cancellation: how quickly a response that is already under way can be stopped, the stop-signal reaction time (SSRT).
Configure parameters and run an interactive preview exactly as participants will experience it. No data is recorded.
Adjust parameters below, then start the preview on the right.
Include practice trials
Shown with feedback before the main task
Task parameters
2 blocks × 64 trials = 128 trials, 32 of them stop trials (16 per block), about 5 min, plus about 1 min of practice. A rest screen follows each block. Stop trials are never the first trial of a block and never more than two in a row.
The consensus guide (Verbruggen et al., 2019) recommends about 50 or more stop trials for a stable SSRT. With 32, add trials or blocks, or pool short runs across sessions before estimating SSRT.
Stop-signal delay (SSD) staircase
The delay rises by one step after a successful stop and falls by one step after a failed one, so that participants stop on about half of the stop trials. One staircase runs through all blocks; practice has its own. Highest left empty: the response window − 100 ms (1400 ms).
Trial timing
Each trial: fixation cross, then the symbol until a response or the end of the response window, then a blank drawn from a normal distribution cut to the shortest–longest range (RDoC: mean 1000 ms, SD 292 ms).
The stop signal is visual: the symbol turns red and a thick ring appears around it, so it is visible without colour vision. There is no auditory stop signal, because audio output latency on Android phones varies too much for 50 ms SSD steps. Stimulus and stop-signal onsets are taken after the screen has been drawn and responses at touch-down, but this timing has not yet been validated on real devices.
Outputs: SSRT by the integration method with replacement of go omissions, p(respond | signal), mean SSD (measured when available), mean and median go RT, go accuracy, wrong-button and no-response rates, mean RT of failed stops, and flags for p(respond | signal) outside .25–.75 and for failed stops that are not faster than go responses (SSRT then not interpretable).
This is a researcher preview. No participant data is recorded.
Simulated participant view
A two-choice arrow task in which a stop signal sometimes follows the arrow. Measures action cancellation: how quickly a response that is already under way can be stopped, the stop-signal reaction time (SSRT).
No data is recorded
An arrow pointing left or right (or a circle or a square) in the centre, and two large response buttons at the bottom within thumb reach. On a quarter of the trials the symbol turns red and a thick ring appears around it after a short delay: the stop signal. The ring makes the signal visible without colour vision. The signal is visual only.
Useful when a study needs action cancellation, stopping a response that has already been started, as in ADHD, addiction and impulsivity research. Go/No-Go measures a different form of inhibition: action restraint, withholding a response to a no-go cue before it starts. Best suited to baseline or occasional sessions. In daily-life designs, short runs repeated across sessions can be pooled before SSRT is estimated.
Not a diagnostic instrument, and not a quick EMA probe: a stable SSRT needs many stop trials (about 50 or more). Not a substitute for Go/No-Go when the question is about withholding a response to a no-go cue.
Keep practice on, the default 25% stop trials and the SSD staircase. Keep the built-in instruction not to wait for the stop signal: waiting slows go responses and biases SSRT. Plan at least 50 stop trials per SSRT estimate, e.g. 2 blocks of 100 trials, or pool short runs across sessions. Before analysing SSRT, check p(respond | signal) and the race-model flag.
A longer SSRT means slower stopping. SSRT is estimated from the independent race model (Logan & Cowan, 1984): the go and the stop process race, and the response is stopped when the stop process finishes first. The estimate is not interpretable when failed stops are not faster than go responses, and unreliable when p(respond | signal) is outside .25–.75; both cases are flagged. An auditory stop signal is not offered, because audio output latency on Android phones varies too much for 50 ms SSD steps. SMAAT places stimulus onsets after the screen has been drawn and takes response times from the touch-down timestamp, but this timing has not yet been validated on real devices (photodiode or high-speed video), so compare SSRT within participants and devices rather than across them. SSRT reliability is moderate; aggregate trials across sessions where possible.
Add Stop-Signal Task to your ESM study on SMAAT — free to get started.