Seeing 'where' through the ears: effects of learning-by-doing and long-term sensory deprivation on localization based on image-to-sound substitution.

BACKGROUND: Sensory substitution devices for the blind translate inaccessible visual information into a format that intact sensory pathways can process. We here tested image-to-sound conversion-based localization of visual stimuli (LEDs and objects) in 13 blindfolded participants. METHODS AND FINDINGS: Subjects were assigned to different roles as a function of two variables: visual deprivation (blindfolded continuously (Bc) for 24 hours per day for 21 days; blindfolded for the tests only (Bt)) and system use (system not used (Sn); system used for tests only (St); system used continuously for 21 days (Sc)). The effect of learning-by-doing was assessed by comparing the performance of eight subjects (BtSt) who only used the mobile substitution device for the tests, to that of three subjects who, in addition, practiced with it for four hours daily in their normal life (BtSc and BcSc); two subjects who did not use the device at all (BtSn and BcSn) allowed assessment of its use in the tasks we employed. The impact of long-term sensory deprivation was investigated by blindfolding three of those participants throughout the three week-long experiment (BcSn, BcSn/c, and BcSc); the other ten subjects were only blindfolded during the tests (BtSn, BtSc, and the eight BtSt subjects). Expectedly, the two subjects who never used the substitution device, while fast in finding the targets, had chance accuracy, whereas subjects who used the device were markedly slower, but showed much better accuracy which improved significantly across our four testing sessions. The three subjects who freely used the device daily as well as during tests were faster and more accurate than those who used it during tests only; however, long-term blindfolding did not notably influence performance. CONCLUSIONS: Together, the results demonstrate that the device allowed blindfolded subjects to increasingly know where something was by listening, and indicate that practice in naturalistic conditions effectively improved "visual" localization performance