Learning to localise weakly-informative sound spectra with and without feedback.

How the human auditory system learns to map complex pinna-induced spectral-shape cues onto veridical estimates of sound-source elevation in the median plane is still unclear. Earlier studies demonstrated considerable sound-localisation plasticity after applying pinna moulds, and to altered vision. Several factors may contribute to auditory spatial learning, like visual or motor feedback, or updated priors. We here induced perceptual learning for sounds with degraded spectral content, having weak, but consistent, elevation-dependent cues, as demonstrated by low-gain stimulus-response relations. During training, we provided visual feedback for only six targets in the midsagittal plane, to which listeners gradually improved their response accuracy. Interestingly, listeners' performance also improved without visual feedback, albeit less strongly. Post-training results showed generalised improved response behaviour, also to non-trained locations and acoustic spectra, presented throughout the two-dimensional frontal hemifield. We argue that the auditory system learns to reweigh contributions from low-informative spectral bands to update its prior elevation estimates, and explain our results with a neuro-computational model.

Accuracy-Precision Trade-off in Human Sound Localisation.

Sensory representations are typically endowed with intrinsic noise, leading to variability and inaccuracies in perceptual responses. The Bayesian framework accounts for an optimal strategy to deal with sensory-motor uncertainty, by combining the noisy sensory input with prior information regarding the distribution of stimulus properties. The maximum-a-posteriori (MAP) estimate selects the perceptual response from the peak (mode) of the resulting posterior distribution that ensure optimal accuracy-precision trade-off when the underlying distributions are Gaussians (minimal mean-squared error, with minimum response variability). We tested this model on human eye- movement responses toward broadband sounds, masked by various levels of background noise, and for head movements to sounds with poor spectral content. We report that the response gain (accuracy) and variability (precision) of the elevation response components changed systematically with the signal-to-noise ratio of the target sound: gains were high for high SNRs and decreased for low SNRs. In contrast, the azimuth response components maintained high gains for all conditions, as predicted by maximum-likelihood estimation. However, we found that the elevation data did not follow the MAP prediction. Instead, results were better described by an alternative decision strategy, in which the response results from taking a random sample from the posterior in each trial. We discuss two potential implementations of a simple posterior sampling scheme in the auditory system that account for the results and argue that although the observed response strategies for azimuth and elevation are sub-optimal with respect to their variability, it allows the auditory system to actively explore the environment in the absence of adequate sensory evidence.

Matching Automatic Gain Control Across Devices in Bimodal Cochlear Implant Users.

OBJECTIVES: The purpose of this study was to improve bimodal benefit in listeners using a cochlear implant (CI) and a hearing aid (HA) in contralateral ears, by matching the time constants and the number of compression channels of the automatic gain control (AGC) of the HA to the CI. Equivalent AGC was hypothesized to support a balanced loudness for dynamically changing signals like speech and improve bimodal benefit for speech understanding in quiet and with noise presented from the side(s) at 90 degree. DESIGN: Fifteen subjects participated in the study, all using the same Advanced Bionics Harmony CI processor and HA (Phonak Naida S IX UP). In a 3-visit crossover design with 4 weeks between sessions, performance was measured using a HA with a standard AGC (syllabic multichannel compression with 1 ms attack time and 50 ms release time) or an AGC that was adjusted to match that of the CI processor (dual AGC broadband compression, 3 and 240 msec attack time, 80 and 1500 msec release time). In all devices, the AGC was activated above the threshold of 63 dB SPL. The authors balanced loudness across the devices for soft and loud input sounds in 3 frequency bands (0 to 548, 548 to 1000, and >1000 Hz). Speech understanding was tested in free field in quiet and in noise for three spatial speaker configurations, with target speech always presented from the front. Single-talker noise was either presented from the CI side or the HA side, or uncorrelated stationary speech-weighted noise or single-talker noise was presented from both sides. Questionnaires were administered to assess differences in perception between the two bimodal fittings. RESULTS: Significant bimodal benefit over the CI alone was only found for the AGC-matched HA for the speech tests with single-talker noise. Compared with the standard HA, matched AGC characteristics significantly improved speech understanding in single-talker noise by 1.9 dB when noise was presented from the HA side. AGC matching increased bimodal benefit insignificantly by 0.6 dB when noise was presented from the CI implanted side, or by 0.8 (single-talker noise) and 1.1 dB (stationary noise) in the more complex configurations with two simultaneous maskers from both sides. In questionnaires, subjects rated the AGC-matched HA higher than the standard HA for understanding of one person in quiet and in noise, and for the quality of sounds. Listening to a slightly raised voice, subjects indicated increased listening comfort with matched AGCs. At the end of the study, 9 of 15 subjects preferred to take home the AGC-matched HA, 1 preferred the standard HA and 5 subjects had no preference. CONCLUSION: For bimodal listening, the AGC-matched HA outperformed the standard HA in speech understanding in noise tasks using a single competing talker and it was favored in questionnaires and in a subjective preference test. When noise was presented from the HA side, AGC matching resulted in a 1.9 dB SNR additional benefit, even though the HA was at the least favorable SNR side in this speaker configuration. Our results possibly suggest better binaural processing for matched AGCs.