Abstract
Purpose :
Congenitally blind individuals are better at voice recognition than sighted individuals (Barton, 2003), but little is known about voice recognition in individuals who lose partial vision later in life. Here we examine whether late blind visually impaired (VI) individuals have enhanced voice recognition, and whether voice recognition can be enhanced by training.
Methods :
We measured face and voice recognition abilities in five VI individuals (1 male + 4 female, age 40.0 ± 8.8) and 5 age matched sighted controls using a 2AFC same-different paradigm. For voice recognition, we successively presented two unfamiliar voice clips (~2 s) extracted from the LibriSpeech voice database (Panayotov, 2014). Each interval contained a clip from the same or a different speaker (same gender). Both intervals were masked by babble. For visual face recognition, each 1 s interval contained the face image (Lundqvist, Flykt, & Öhman, 1998) of either the same or a different individual (same gender, different facial expressions and viewing angle). We then asked VI participants to carry out voice recognition training (50 trials/day) for 3 weeks. After training, we repeated the voice recognition test with novel voices.
Results :
As expected, VI participants visual face recognition was far worse in VI individuals (d-prime = 1.10 ± 1.37) than for sighted controls (4.05±0.51). Pre-training voice recognition was slightly better in VI participants (1.11 ± 0.23) than for controls (0.57 ± 0.31). Within VI participants we did see a small effect of training - d-primes increased from 0.57 (± 0.31) to 0.89 (± 0.42).
Conclusions :
These results suggest that training may help to improve voice recognition abilities. In future work we hope to examine whether improving voice recognition mitigates the higher levels of loneliness that are associated with visual impairment.
This is a 2020 ARVO Annual Meeting abstract.