Figure 3

Correlation between contrast sensitivity peak value and number of mistakes made in the FM 100 hue arrangement test. Psychophysical tests were performed using monocular vision. Contrast sensitivity peak value and number of FM 100 mistakes were plotted using a log₁₀ scale. (A-C) Data points were fitted with linear regressions using the least square method: (A) controls, (B) diabetics patients without retinopathy, and (C) diabetic patients with retinopathy. There were moderately significant linear correlations between the FM 100 score and CSF peak for the three groups of subjects, but correlation was higher for controls (A) and diabetic patients with retinopathy (C) than for diabetic patients without retinopathy (B). (D) Data points were fitted with ellipses using the Khachiyan Ellipsoid Method: (a) controls, (b) diabetic patients without retinopathy, and (c) diabetic patients with retinopathy. Diabetic ellipses are higher in the diagram than the control ellipse owing to their larger FM 100 scores. The ellipse from the diabetic patients with retinopathy is displaced to the left in the diagram in comparison with the ellipses from diabetic patients without retinopathy and from controls, owing to their lower contrast sensitivity peaks.