Abstract
The contrast sensitivity function (CSF) is widely used to estimate individuals' visual capacities with grating spatial frequency. However, the long testing time limits its use in clinical and practical settings. Infants' CSF acquisitions are even more challenging because of their short attention span and inability to understand language. Rapid measurement of CSF (qCSF) based on Bayesian adaptive inference has been applied on adults, and here we examine whether the qCSF is applicable to preverbal infants with a preferential looking paradigm.
Infants sat facing a CRT monitor. An image for attention attraction was displayed at the center of the screen at the beginning of each trial, followed by an 8.72 degree static sinusoidal grating presented on either the left or right side of the monitor together with a tone. Infants' eye movements were monitored by a camera, and projected to another screen from which two coders made their judgments as to which side infants gazed. The contrasts and spatial frequencies (from 0.11–7.73 cpd) of gratings were estimated trial-by-trial to maximize the efficiency of predicting a CSF by Bayesian inference.
Only results with above 70% inter coder rating consistency were included. Sixty-six infants ranging from 3- to 10-months-old met this criterion. Each predicted CSF was defined by four parameters, which were compared with previous infant CSF results. Our CSF peak estimates are consistent with previous studies, only qCSF enabled finer-scaled estimations and shorter testing time – each CSF was derived from an average of 35 trials. This is a marked improvement over previous studies that required multiple visits and a minimum of 30–50 trials for a single threshold estimation. Our results show that qCSF is an efficient method that can be combined with preferential looking paradigm in obtaining a reliable CSF in preverbal infants.