Purchase this article with an account.
Luis Andres Lesmes, Sergei Gepshtein, Zhong-Lin Lu, Thomas Albright; Rapid estimation of the spatiotemporal contrast sensitivity surface. Journal of Vision 2009;9(8):696. doi: 10.1167/9.8.696.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose. The spatiotemporal contrast sensitivity surface (CSS) describes visual sensitivity (1/threshold) to moving or flickering gratings as a function of spatial and temporal frequency1. The CSS provides a fundamental characterization of the visual system in both normal and clinical populations. Many neuro-ocular diseases exhibit characteristic frequency-specific deficits on the CSS2. To overcome the long testing times typical needed to measure the CSS, we develop a family of adaptive methods for its rapid estimation.
Method. The CSS is typically studied in orthogonal and diagonal slices through its surface: spatial contrast sensitivity functions (CSFs) at fixed temporal frequencies, temporal CSFs at fixed spatial frequencies, or constant-speed CSFs at co-varied spatial and temporal frequencies. We estimated these contrast sensitivity slices by combining Bayesian adaptive inference with a trial-to-trial information-gain strategy3. To estimate the entire CSS, our novel procedure combined the information gained from adaptive runs dedicated to individual slices. Before each trial, the procedure evaluated expected gain within individual slices (6 spatial, 6 temporal, and 7 speeds) and selected a stimulus maximizing the information gain expected among all the slices. The final CSS estimate combined the surface estimates from all slices. In psychophysical experiments, we measured human sensitivity for motion direction discrimination over a large range of spatial (0.5–8 cycles/deg) and temporal frequencies (0.25–24 Hz).
Results. Simulation and psychophysical results suggest accurate CSS estimates are possible within 300–500 trials (15–25 minutes) with an average precision of 2–3 dB. Monte Carlo sampling of posteriors provides confidence regions for the CSS based on single adaptive runs.
Conclusion. This procedure offers a useful tool for clinical and practical applications that require a rapid but comprehensive evaluation of visual sensitivity.
This PDF is available to Subscribers Only