Fig. 4: iPad error convergence across trials for the PS, AS, and MGS tasks.

Convergence for a latency and b amplitude. For each subject and measure, trial-level (i.e., saccade-level) errors were first averaged across tasks at each trial index to produce a single per-trial error sequence. The curve shows the mean absolute deviation from the full subject-level error as a function of the number of trials used. Vertical lines mark the trial counts at which the average error reached 50% (blue) and 90% (green) convergence toward the full subject-level error. Convergence was quantified as \(1-\frac{{E}_{n}}{{E}_{1}}\), where En is the mean deviation after n trials and E1 is the deviation using only the first trial.