Diagnosis of Normal and Abnormal Color Vision with Cone-Specific VEPs

Abstract

Purpose: Normal color vision depends on normal long wavelength (L), middle wavelength (M), and short wavelength sensitive (S) cones. Hereditary "red-green" color vision deficiency (CVD) is due to a shift in peak sensitivity or lack of L or M cones. Hereditary S cone CVD is rare but can be acquired as an early sign of disease. Current tests detect CVD but few diagnose type or severity, critical for linking performance to real-world demands. The anomaloscope and newer subjective tests quantify CVD but are not applicable to infants or cognitively impaired patients. Our purpose was to develop an objective test of CVD with sensitivity and specificity comparable to current tests.

Methods: A calibrated visual-evoked potential (VEP) display and Food and Drug Administration-approved system was used to record L, M, and S cone-specific pattern-onset VEPs from 18 color vision normals (CVNs) and 13 hereditary CVDs. VEP amplitudes and latencies were compared between groups to establish VEP sensitivity and specificity.

Results: Cone VEPs show 100% sensitivity for diagnosis of CVD and 94% specificity for confirming CVN. L cone (protan) CVDs showed a significant increase in L cone latency (53.1 msec, P < 0.003) and decreased amplitude (10.8 uV, P < 0.0000005) but normal M and S cone VEPs (P > 0.31). M cone (deutan) CVDs showed a significant increase in M cone latency (31.0 msec, P < 0.000004) and decreased amplitude (8.4 uV, P < 0.006) but normal L and S cone VEPs (P > 0.29).

Conclusions: Cone-specific VEPs offer a rapid, objective test to diagnose hereditary CVD and show potential for detecting acquired CVD in various diseases.

Translational relevance: This paper describes the efficacy of cone-specific color VEPs for quantification of normal and abnormal color vision. The rapid, objective nature of this approach makes it suitable for detecting color sensitivity loss in infants and the cognitively impaired.

Keywords: VEPs; color vision; electrophysiology.

Figure 1

L, M, and S cone-specific checkerboard patterns used for pattern-onset VEPs. The cone-specific colored checks appeared as increments to the gray background, which appears different on each display due to color induction effects and photographic differences. The Weber contrast values are color coded for each cone type, while the white text represents contrast to nontargeted cone types, which is near or below threshold for detection.

Figure 2

Normal cone-specific pattern-onset VEP showing VEP latency and amplitude.

Figure 3

VEP waveforms from a normal subject (left), protan (L cone deficient, middle), and deutan (M cone deficient, right). VEP amplitude is shown on the y-axis (μV) and latency on the x-axis (msec). The L cone deficient (protan) shows decreased VEP amplitude and increased latency on the L cone VEP, and the M cone deficient (deutan) shows decreased amplitude and increased latency on the M cone VEP.

Figure 4

Mean VEP amplitudes (±2 standard errors) are shown for CVN, protan, and deutan groups. Compared to CVNs, two-tailed t-tests showed a significant decrease in VEP amplitude for protans on the L cone VEP (P < 0.0000005; “*”) and for deutans on the M cone VEP (P < 0.007; “*”). There was no difference between groups on the S cone VEP (P > 0.37).

Figure 5

Mean VEP latencies (±2 standard errors) are shown for CVN and protan, deutan, and CVD groups. Compared to CVNs, two-tailed t-tests showed a significant increase in VEP latency for protans on the L cone VEP (P < 0.003; “*”) and for deutans on the M cone VEP (P < 0.000004; “*”). There was no difference between groups on the S cone VEP (P > 0.58).

Figure 6

S cone VEPs are shown for a CVN subject at field sizes ranging from the standard 30-degree field down to 5 degrees. Please see text for further details.