Abstract
Comprehensive visual function assessment is a critical stage of basic (screening, inclusion/exclusion) and clinical (diagnosis, monitoring) science, but requires tests that can be frustrating, esoteric and prohibitively long, which can lead to inaccurate or incomplete testing. We introduce FInD (Foraging Interactive D-prime), a general-purpose method that rapidly and easily measures threshold functions. Subjects search a series of (typically 3) grids composed of (typically 4*4) cells. A random number of the cells contain targets of signal intensities ranging from difficult (d’=0.1) to easy (d’=4.5), thus a clear exemplar is always present. In a self-administered test that requires little or no explanation and is suitable for special populations, observers click or touch cells containing a target. The response in each cell is classified as a Hit, Miss, False Alarm or Correct Rejection, to calculate d’ as a function of signal intensity, which is then fit with a decision function. The signal range in each grid is determined from the posterior estimates of slope and threshold (d’=1) from previous grids (or experimenter estimate on grid #1) to maximize the efficiency of the test (typically over 3*4*4=48 trials). We measured contrast sensitivity functions (CSF), threshold versus contrast functions (TvC) stereoacuity, spatial coherence and motion coherence thresholds in up to 14 normally-sighted observers with FInD and with standard temporal 2 Alternative-Forced-Choice (AFC) paradigms. Threshold and function parameter estimates for CSF and TvC were not significantly different for FInD and 2AFC paradigms, but took significantly less time (36.99+7.0s vs146.30±14.86s, p<0.01) to measure. Bland-Altman analyses showed no significant bias or test-retest differences for FInD. FInD solves multiple screening problems for comprehensive vision assessment and rapidly delivers accurate and precise estimates of multiple visual functions in an easy-to-learn, self-administrable, general paradigm.