gms | German Medical Science

Objective: Psychophysical tasks have been used with simulations of artificial vision viewed by normal sighted subjects to evaluate performance and help guide device design, but with little attention paid to learning effects. We examined the effects of training when reading through a simulated visual prosthesis over multiple sessions to address this gap.

Materials and methods: Six normal, sighted subjects read MNREAD-style sentences in 20-minute sessions on a near-daily basis, over a period of eight weeks. Sentences were presented through a gaze-contingent simulation of artificial vision at five font sizes (logMAR 1.0-1.4), through three center-weighted phosphene patterns (2000, 1000, 500 phosphenes) or a control condition with normally displayed text. Novel sentences were used for each session. Task performance was assessed through reading accuracy and reading speed.

Results: Subjects markedly improved their reading accuracy equivalent to logMAR 0.3 change in acuity, with a concurrent increase of reading speed equivalent to a doubling of phosphenes. Strikingly, the hardest condition (smallest font size viewed through 500 phosphenes) was initially illegible, but supported functional reading after training.

Discussion: Our findings emphasize the importance of post-implant rehabilitation and suggest that even a pattern of 500 phosphenes can be usable for reading, an important activity of daily living.

Acknowledgement: William M. Wood Foundation, Fulbright Foundation in Greece.