New type of perceptual suppression during dynamic ocular accommodation

When viewing natural scenes saccadic eye movements are used to position retinal images of interest on the fovea, while vergence eye movements act to minimise retinal disparity and maintain a single binocular view of the target. Despite these disruptions to the retinal image, we perceive a stable visual world due to suppression of the magnocellular visual pathway which processes low spatial frequency target details [1]. Retinal image clarity is maintained by the mechanism of accommodation, which changes ocular power to produce sharp retinal images of objects at different viewing distances [2]. During abrupt dynamic changes in accommodation, visual suppression seems to occur as image degradation is rarely reported. Here, we report direct psychophysical evidence that visual suppression during dynamic accommodation really does occur. This suppression is selective for luminance modulated patterns of higher spatial frequencies, implying that the suppression is occurring in the parvocellular visual pathway, which carries colour and fine detail information. Dynamic accommodation responses have a dual-mode behaviour comprising an initial fast response, followed by a slower component [3,4]. The initial fast component provides a preprogrammed accommodation response, which is ballistic in nature and proportional to the level of retinal image blur. This fast component accounts for the majority of the response amplitude and produces abrupt changes in accommodation response level required to form clear retinal images. The slow component completes the abrupt change in accommodation and employs continuous feedback control to refine the response. Visual suppression is therefore likely to occur during the fast component of the accommodation response, which does not rely upon visual feedback control. Using an adaptive staircase procedure and a two-alternative forced choice method, monocular contrast thresholds for detecting brief (43 ms) luminance modulated gratings of 1, 4 and 9 cycles per degree (cpd) presented at 1 m distance were measured during dynamic far-to-near accommodation responses (1 m to 0.33 m) at predetermined time markers after accommodation stimulus onset (see Figure 1 and Supplemental Data available on-line with this issue). Recorded accommodation responses were used to calculate the magnitude of retinal image defocus occurring at each time marker (Figure 1). Contrast thresholds were also measured during steady-state accommodation responses at viewing distances equal to the amount of defocus present in the dynamic conditions at each time marker. As expected, contrast sensitivity (the reciprocal of contrast threshold) in both dynamic and steady-state accommodation conditions was reduced as a function of stimuli defocus to a greater extent at higher spatial frequencies (Figure 2) [5]. Multiple comparisons (ttests using a Bonferroni correction) between contrast sensitivities in dynamic (Figure 2, open symbols) and steadystate (Figure 2, filled symbols) conditions revealed that contrast sensitivities for gratings of 1 cpd were not significantly different. Dynamic conditions reduced significantly (P < 0.05) the contrast sensitivity for 4 cpd gratings only for one subject (MD), while contrast sensitivities for 9 cpd gratings were significantly (P < 0.05) reduced for all subjects during the fast phase of the dynamic accommodation response (Figure 2 arrows; see Supplemental data). To counterbalance accommodation direction and exclude possible reductions in retinal image contrast during the fast component of the dynamic accommodation response we measured contrast sensitivity to brief gratings of 6 ms duration and 4 degree diameter presented at 1 m viewing distance during far-to-near and nearto-far accommodation (Supplemental Data, control experiment 1). We found a significant reduction in sensitivity during the fast phase of far-to-near accommodation (Supplemental Figure S1A) and a lack of suppression during near-tofar accommodation (Supplemental Figure S1B). This finding suggests that high spatial frequency information relating to the starting point of the accommodation response is selectively suppressed, with information about the end point being maintained. Based on this prediction, high spatial frequency gratings displayed at a near viewing distance (0.33 m) would be suppressed only during the fast phase of neartofar accommodation responses but not during far-tonear accommodation responses. This was confirmed by subsequent results (Supplemental Data, control experiment 1, Figure S1C,D). Further control experiments examined whether bias in the direction of defocus (the visual stimulus getting more blurred during the dynamic accommodation response), vergence eye movements Viewing distance (m) 1 0.8 0.6 0.5 0.4 0.33