Bayesian adaptive estimation of threshold versus contrast external noise functions: the quick TvC method.

Area of Science:

• Vision science
• Perceptual psychology

Background:

• External noise paradigms are crucial for studying perceptual mechanisms by measuring contrast threshold as a function of external noise contrast (TvC function).
• Traditional methods for measuring TvC functions at multiple performance criteria are data-intensive, often requiring over 2000 trials.

Purpose of the Study:

• To develop a novel Bayesian adaptive procedure, the quick TvC (qTvC) method, for rapid estimation of multiple TvC functions.
• To reduce the demanding data collection associated with traditional TvC function measurements.

Main Methods:

• The qTvC method adapts a Bayesian adaptive strategy to estimate three key parameters of TvC functions: optimal threshold (c(0)), critical noise level (N(c)), and common slope (eta).
• It employs a one-step-ahead search to select stimuli that minimize the expected entropy of the posterior probability distribution.
• Evaluated using Monte-Carlo simulations and a psychophysical orientation identification task.

Main Results:

• Simulations demonstrated that qTvC requires fewer than 300 trials to estimate TvC functions at three performance criteria with high accuracy (bias<5%) and precision (RMSE<1.5 dB).
• Psychophysical testing showed excellent agreement (r(2)>.95) between qTvC and the method of constant stimuli, with qTvC using only 12% of the trials (240 vs 1920).

Conclusions:

• The qTvC method offers a significant reduction in data collection time for estimating TvC functions.
• It provides accurate and precise estimates, making it valuable for studying observer state changes and special populations.
• The adaptive strategy may be applicable to estimating other psychophysical functions, such as contrast sensitivity and sensory memory decay.