Abstract
Current treatment of patients with central scotoma emphasizes the establishment of one preferred retinal locus (PRL). A single PRL may be suitable for viewing nearby targets, but large saccades are required to reach targets on the opposite side of the scotoma, which may cause temporary concealment of the target by the scotoma. One possible remedy is to establish a preferred retinal annulus (PRA) around the scotoma to improve target viewing. Observers with normal vision were trained to identify a blurred tumbling-E target with a 5-deg-radius simulated central scotoma. The target became clear only when it fell into a small window of clear vision contingent to the scotoma edge (Liu & Kwon, 2016). For the PRL group, the clear window was fixed at either left or right side of the scotoma. For the PRA group, it was in the radial direction to the target. Different eye-movement patterns and viewing accuracies emerged after training. The PRL group had the scotoma optimally landed between the initial fixation and target (undershots) in 73% first and 88% second saccades when the target and PRL were on the same side of the scotoma (PRL-near). However, when the target and PRL were on the opposite sides (PRL-far), only 21% first and 41% second saccades optimally landed beyond the target (overshots). Moreover, PRL-far saccades were less accurate than PRL-near saccades, resulting in more target concealments by the scotoma (41% and 43% PRL-far vs. 23% and 7% PRL-near in 1st and 2nd saccades, respectively). The PRA group had mostly optimal undershot saccades (83% first and 91% second saccades), and very few target blockages (16% first and 5% second saccades). These results demonstrated the feasibility and potential functional benefits of PRA training over PRL training. Future experiments will apply PRA and PRL trainings in AMD patients to compare accuracies and efficiencies.
Meeting abstract presented at VSS 2018