Abstract
Purpose :
Previously, we derived the signal gain and internal noise profiles in spatial vision as functions of spatial frequency through modeling the contrast sensitivity functions (CSF) measured in a range of external noise levels with the method of constant stimuli using a multi-channel perceptual template model (mPTM) (Chen et al., 2014; Hou, Lu, & Huang, 2014). Here, we extended and validated the qCSF method (Lesmes, et al., 2010) originally developed in zero external noise to measure CSF in multiple external noise conditions. The mPTM was used to extract the signal gain and internal noise profiles of spatial vision.
Methods :
The CSFs of five normal observers in zero and high external noise conditions were measured in a 4AFC sinewave grating orientation identification task with the qCSF procedure and Ψ method (Kontsevich & Tyler, 1999). External noise was generated by filtering high contrast Gaussian white noise with a one-octave-wide raised cosine filter centered at the test grating spatial frequency. The mPTM, with both signal gain and internal noise profiles described by log parabola functions, was fit to the data.
Results :
We found that (1) Consistent with previous studies, the CSFs in the high external noise condition were virtually flat; (2) The CSFs obtained with the qCSF and Ψ methods were highly correlated in both external noise conditions (r = 0.95 ± 0.03); (3) The standard deviation of the CSFs obtained with 100 qCSF trials was 0.06 ± 0.01 and 0.07 ± 0.003 decimal log unit in the zero and high external noise conditions, respectively, with no significant difference between the two (p > 0.55). (3) The bias of the CSFs obtained with 100 qCSF trials was 0 ± 0.017 and 0.006 ± 0.041 decimal log unit in the two external noise conditions, with no significant difference (p > 0.75). (4) The mPTM accounted for 96.3 ± 2.0% of the variance in the CSF data. The estimated signal gain profile was relatively flat. The magnitude of internal noise elevated with increasing spatial frequency.
Conclusions :
The qCSF method can be extended to provide efficient, precise, and accurate measures of CSF in different external noise conditions. The CSFs obtained with 200 qCSF trials in zero and high external noise conditions can be used to reliably estimate the signal gain and internal noise profiles of spatial vision.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.