Objective: To compare short-wavelength automated perimetry (SWAP), a test favoring the detection of the target by the parvocellular pathways of vision, with motion automated perimetry (MAP), a test favoring detection by the magnocellular pathways, in the same eyes.
Participant: Thirty-three individuals in whom glaucoma was suspected (glaucoma suspects) and 17 patients with primary open-angle glaucoma were compared with 30 age-matched normal control subjects.
Interventions: Short-wavelength automated perimetry was done with the usual protocol (program 24-2). Motion coherence thresholds were measured with 14 random do targets that covered the 24-2 field area. Short-wavelength automated perimetry test locations corresponding to each of the 14 motion automated perimetry locations were averaged to compare 14 locations for each text.
Results: Short-wavelength automated perimetry and motion automated perimetry were correlated by visual field location (whole field r = -0.40, P < .001), especially in the superior field (r = -0.45, P < .001). Overlap for defective locations was present in 16 (94%) of the 17 eyes with glaucoma, although in the glaucoma suspect eyes each test showed the earliest deficit in a percentage of individuals with overlap in only 3 (21%) of the 14 eyes. An analysis of variance showed a significant effect of diagnosis for both tests (SWAP and MAP, P < .001); the eyes of patients with glaucoma were significantly different from those of the normal controls. The results for glaucoma suspects were significantly different on SWAP only in the superior temporal field (Tukey-Kramer test).
Conclusions: Both tests successfully identified eyes with glaucoma and a percentage of the glaucoma suspect eyes; both were correlated by field location. These results suggest that damage due to glaucoma is nonselective for either the parvocellular or the magnocellular ganglion cell axons, that there may be individual differences in which type of ganglion cell shows damage first, and that when standard visual field loss is present the results of SWAP and MAP are defective.