The +50 ms start change stays at trial 16 in 495 of 532 ds000030 version 1.0.0 scanner staircase replays.

A fixed-response replay changes a stop-signal staircase start delay while keeping the recorded STOP responses. Using all 266 scanner runs (532 staircases) from OpenNeuro ds000030 version 1.0.0, new standard-library code found that a +50 ms start change remained complete at the last presented trial (trial 16) in 495/532 staircases; it was zero in 37/532. After the final update, 492/532 retained the full difference. Both staircases retained it in 240/266 runs. The exact mean change across run means was 51025/1064 ms (47.955827067669176 ms), with median 50 ms and range 9.375–50 ms. The implementation agreed with a direct update calculation in 81,430 state checks and matched 67,881 numeric or empty target cells and all 1,302 flagged row identifiers. All 17,056 presented STOP plans and 15,990 transitions passed the recorded-rule checks. The source contained 267 training runs; the last training update matched scanner starts in 530/532 paired staircases, with two unexplained exceptions for sub-10855. The result is an arithmetic replay on fixed recorded responses, not a measured response to changed delays, an independent sample, a causal training effect, or an estimate of stopping ability. Each run had 32 STOP trials; the analysis did not estimate SSRT or calculate population confidence intervals.

stop-signal task · fixed-response replay · staircase · starting delay · response inhibition · OpenNeuro · ds000030 · reproducibility · prefix sums · scanner data