CHILDREN'S CLINICAL TESTS OF CONTRAST SENSITIVITY: VALIDITY AGAINST THE PELLI-ROBSON CHART

Susan Leat

Abstract

PURPOSE. Contrast sensitivity (CS) is an important clinical measure in the visual assessment of children with low vision and special needs. Recently, two tests of CS, the Hiding Heidi and the LEA low contrast symbols, both designed for pre-literate children, have become commercially available. The LEA symbols is a matching test, and the Hiding Heidi is a preferential looking or pointing test. In this study we validated these tests against the standard of the Pelli-Robson chart.

METHOD. Thirty-one adult subjects with low vision or amblyopia, plus 3 subjects with normal visual acuity took part. These subjects were chosen to obtain a good range of CS. CS was measured with the Hiding Heidi test at 1m (HH), the LEA low contrast symbol test at 0.28m (0.28LEA) and 1m (1mLEA), and the Pelli-Robson chart at 1m (PR). Since there were visually apparent differences between the contrast levels of the Pelli-Robson chart and the stated contrast for the children's charts, each level of contrast for both the children's tests was re-calibrated subjectively against the PR chart and objectively by luminance measurements.

RESULTS. For both the HH and LEA charts, the subjective and objective calibration resulted in higher contrast values than indicated by the manufacturer. The best correlation was between the PR and the 0.28LEA (r=0.8799) and the mean of the differences was 0.57+/-0.25 log CS. The agreement improved to 0.11+/-0.25 with re-calibration of the LEA contrast levels. The 1mLEA correlated better with logMAR VA (r=0.8336) than with PR CS (r=0.8033). The poorest correlation was between the HH test and the PR (r=0.6031) and there was a ceiling effect even for low vision subjects.

CONCLUSIONS. The Hiding Heidi chart does not compare well with the PR chart, and therefore does not have great predictive value in a visual assessment (since many studies of function and CS have used the PR chart). This poor agreement is likely because of its many spatial frequency components, including very low frequencies and its relatively large area of stimulus. The LEA chart at 1m appears to be more a measure of VA than CS. The 0.28mLEA gives the most useful information, once re-calibrated for contrast, and may be useful for predicting performance of children with low vision.

Details

Year: 2001

Program Number: 4:20 pm

Author Affiliation: University of Waterloo

Co-Authors: Daniela Wegmann

Co-Author Affiliation: n/a

Room: Room 108