Selective Listening in Autism: The Influence of Informational Masking

Friday, May 16, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
I. F. Lin1, T. Yamada2, M. Nakamura2, H. Watanabe2, Y. Takayama2, A. Iwanami2, N. Kato2 and M. Kashino1,3,4, (1)NTT Communication Science Laboratories, Atsugi, Japan, (2)Showa University, Tokyo, Japan, (3)Department of Information Processing, Tokyo Institute of Technology, Yokohama, Japan, (4)CREST, JST, Atsugi-shi, Japan
Background: In a noisy environment, the difficulty to hear out the target speech can be attributed to energetic masking and informational masking. If the target and masker have the same properties, even when the target and masker are processed in different peripheral channels (i.e., no energetic masking), the neurotypical people are still suffered from central, informational masking.

Objectives: This study investigated how autistic people are influenced by target-masker similarity and binaural (spatial) separation when there is no energetic masking between the target and masker.

Methods: Neurotypical and autistic participants were instructed to hear out the target among maskers in a two-interval forced choice experiment. The target was composed of pure tones at random frequencies that were within a protected spectral region, and it was sent only to the right ear. While the maskers were composed of pure tones outside the protected spectral regions, these pure tones either had random frequencies (condition 1 & 3) or had the same frequencies across time (condition 2 & 4). The maskers were either sent only to the right ear (condition 1 & 2) or to both ears (condition 3 & 4). Thresholds of the sound pressure level for the target to be heard out were measured by the 3-up-1-down staircase method.

Results: In the neurotypical group, the thresholds measured in condition 2 & 3 were significantly smaller than the threshold measured in condition 1. This observation is consistent with previous studies, which reveal the benefit from target-masker dissimilarity and binaural (spatial) separation. Nevertheless, in the autistic group, the thresholds in condition 1 were not significantly larger than the thresholds in other conditions. The thresholds in all conditions in the autistic group were similar to the thresholds observed in conditions 2, 3, and 4 in the neurotypical group.

Conclusions: The results show that when the spectral regions of the target and masker were not overlapped, the ability to hear out the target in the autistic group was not degraded by target-masker similarity. This observation supported the hypothesis that when autistic people process local information, they are less influenced by the distracting global information.