Identifying steep psychometric function slope quickly in clinical applications

Knowledge of an observer's psychometric function slope is potentially useful in clinical visual psychophysics (for example, perimetry), however, the short test times necessary in a clinical setting typically prevent slope estimation. We explore, using computer simulation, the performance of several possible procedures for estimating psychometric function slope within limited presentations (aiming for approximately 30 or 140 trials). Procedures were based on either adaptive staircase or Bayesian techniques, and performance was compared to a Method of Constant Stimuli. An adaptation of the Ψ algorithm was best performing, being able to reliably identify steep from flat psychometric functions in less than 30 presentations, however reliable quantification of shallow psychometric functions was not possible.

Research highlights► We present clinically viable algorithms for finding psychometric function slope ► An adaption of the Psi algorithm of Konstevich and Tyler works well ► Steep slopes can be reliably identified in less than 30 presentations.