Detection and localization of sounds using staircase method.

Dogs’ auditory abilities have mainly been studied using electrophysiological methods. However, in both dogs and humans, thresholds found using different methods are different, as well as their inter/intra-individual variability and biological significance. Indeed, only psychoacoustics methods may assess the entire process of sound perception and, to date, only two studies were applying these methodologies to dogs. The aim of this study was to develop a behavioral methodology to expand our knowledge regarding dog’s auditory abilities, allowing also assessment of possible improvement of thresholds’ estimation across exposures. Two procedures were developed in order to determine hearing threshold at three specific frequencies (0.5, 4.0 and 20.0 kHz), and dogs’ sound localization capability in the azimuthal plane. Dogs were initially trained to discriminate a pure tone or a white noise, respectively for hearing threshold and sound localization procedure in a two-alternative forced-choice paradigm. Then, for the assessment of hearing threshold, the intensity level of the pure tone was systematically manipulated from 70 to -3 dB SPL. For the assessment of sound localization, two identical speakers were progressively displaced from 120 to 1  of separation. For both procedures, dogs performed descending and ascending assessment using the staircase method, where the intensity level or angle of separation of a set of trials is determined according to the previous performance of the dog. Hearing thresholds were assessed in 5 dogs per each frequency. The hearing thresholds found were 19.5   2.8 dB SPL at 0.5 kHz, 14.0   4.5 dB SPL at 4kHz, 3.5   12.8 dB SPL at 20 kHz, and were similar to the ones found in the behavioral and electrophysiological studies. However, the electrophysiological methods may often not perform frequency lower than 1 kHz due to technical issues, and have bigger inter-individual variability. The Minimum Detectable Angle (MDA) of sound localization was estimated in 10 dogs, and was found at 7.6   3.4  with an improvement linked to exposure. Taken together, results indicate that a staircase method is a feasible approach for assessing hearing abilities in dogs, with low inter/intra-individual variability and with a sensitivity of 3 dB SPL for hearing threshold and 1  for MDA. Our results also highlight an improvement across time, underlining the necessity to give enough repetitions to allow a stable estimation of thresholds in a sensory discrimination task.