Spatial frequency discrimination in normal vision and in patients with multiple sclerosis

This article extends our previous reports that multiple sclerosis can cause a visual dysfunction better described as a distortion than as a blurring of vision. An earlier paper reported that multiple sclerosis spares visual acuity in some patients while reducing visual sensitivity for less fine detail. Specifically, these patients experience a loss of contrast sensitivity for low and/or intermediate spatial frequencies, while contrast sensitivity for high spatial frequencies is unimpaired. We report here that some patients also lose spatial frequency discrimination, so that these patients cannot tell which of two clearly visible gratings has the higher spatial frequency even though control subjects accurately report which grating has the higher spatial frequency. One way of regarding this discrimination loss is in terms of a deterioration of the ability to discriminate size. Contrast sensitivity was measured over the spatial frequency range 1 to 20 cycles/deg using the von Békésy tracking method for 10 patients. (20 eyes) and 16 control subjects (32 eyes). The limit of normality was taken as 2.5 standard deviations from the control mean (99 per cent confidence). Spatial frequency discrimination was measured using the criterion-free method of temporal two-alternative forced choice over the spatial frequency range 2 to 16 cycles/deg for 10 patients (20 eyes), and for 14 to 26 control eyes at each spatial frequency. Three control subjects were studied more extensively over the range 1 to 20 cycles/deg. Control subjects could discriminate two spatial frequencies that differed by more than about 5 per cent. This held for all spatial frequencies tested. Grating contrast had little effect on discrimination, provided that all test gratings were clearly visible. The normal limit for discrimination threshold was set at 2.5 standard deviations from the control mean. Seven of 10 patients have abnormal contrast sensitivity at one or more spatial frequencies. Six of 10 patients had abnormal discrimination at one or more spatial frequencies. At any given spatial frequency the correlation between the magnitudes of sensitivity loss and discrimination loss was weak, though an eye that was less sensitive than its fellow also tended to have poorer discrimination. A more subtle relationship between sensitivity loss and discrimination loss was clearly shown by one patient. Sensitivity loss was restricted to spatial frequencies below 8 cycles/deg, while discrimination loss in the same eye was restricted to spatial frequencies above 8 cycles/deg. We propose that this finding can be straightforwardly understood if discrimination is determined by the relative activities of different spatial frequency channels analogously, to the way opponent-colour mechanisms determine colour discrimination.