Purchase this article with an account.
FA Ennis, AJ Anderson, CA Johnson; A New Staircase Method, 2A-Non-FC, in Resolution Acuity Threshold Recording . Invest. Ophthalmol. Vis. Sci. 2002;43(13):2137.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: In laboratory-based resolution acuity experiments, 2-alternative forced-choice (2AFC) staircase paradigms are often used because the 2AFC is less criterion-dependent than yes/no paradigms and staircases are more efficient than method of constant stimuli (MOCS) methods. However, these staircases require a 3up/1down paradigm and the recording of multiple reversals in order to reduce threshold measurement variability caused by correct guessing beyond the true resolution limit. Hence these staircases may be too long for clinical usage in resolution perimetry, where threshold measurements at many locations are required. Therefore the evaluation of other, potentially faster, thresholding techniques is clinically desirable. We report on a new modification of the 2AFC paradigm, the 2-alternative non-forced-choice. Methods: Resolution acuity thresholds were recorded at 200 in the periphery (nasal field, right eye) for eight normal subjects using four staircase paradigms: (1) 3up/1down 2AFC, (2) 1up/1down 2AFC, (3) 1up/1down Yes/No and (4) the new paradigm, 1up/1down 2A-Non-FC. The new technique differs from normal 2AFC techniques by allowing the subject to not respond, the non-response being interpreted as an «incorrect» response. Thresholds thus recorded were compared to those recorded using a MOCS paradigm. Results: The 1up/1down 2AFC consistently over-estimated (by a factor of 2-4) and the Yes/No method frequently under-estimated (by a factor of 0-2) resolution performance. Hence both seem inappropriate for usage in the current context. The 3up/1down agreed most closely with the MOCS results, but the new method also showed good agreement with a significant (≷50%) reduction in presentations required for the same number of reversals. The threshold variability for the new method also compared favorably with the conventional 3up/1down method. Conclusion: The new 1up/1down 2A-Non-FC paradigm performs as accurately as the conventional 3up/1down paradigm but requires less than half the number of stimulus presentations. It is more accurate than the 1up/1down 2AFC and less criterion dependant than the Yes/No paradigm, without loss of efficiency. Hence this method may prove useful in the clinical setting in the development of resolution perimetry.
This PDF is available to Subscribers Only