1,720,981 research outputs found
FrAMBI: A Software Framework for Auditory Modeling Based on Bayesian Inference
No full textResearch in hearing science often relies on auditory models to describe listener's behaviour and its neural underpinning in acoustic environments. These models gather empirical evidence from behavioural data to address research questions on the neural mechanisms underlying sound perception. Despite seemingly similar statistical methods, auditory models are often implemented for each study separately, which hinders reproducibility and across-study comparisons, thus limiting the advancement at a field level. Here, we introduce a framework for studying neural mechanisms of sound perception by employing auditory modeling based on Bayesian inference (FrAMBI), a MATLAB/Octave toolbox. FrAMBI provides a standardized structure to implement an auditory model following the perception-action cycle and enables the automatic application of statistical analysis with behavioural data. We show FrAMBI's capabilities in several examples with increasing levels of complexity within the context of sound source localisation tasks: a basic implementation for a static scenario, iterating over the perception-action cycle with a moving sound source, the definition of multiple model variants testing different neural mechanisms, and the procedure for parameter estimation and model comparison. Being integrated into the widely used auditory modelling toolbox (AMT), FrAMBI is planned to be maintained in the long term and expanded accordingly, fostering reproducible research in the field of neuroscience
Round Robin Comparison of Inter-Laboratory HRTF Measurements – Assessment with an auditory model for elevation
Full text linkRepeatability of head-related transfer function (HRTF) measurements is a critical issue in intra- and inter- laboratory setups. In this paper, simulated perceptual variabilities of HRTFs are computed as an attempt to understand if different acquisition methods achieve similar results in terms of psychoacoustic features. We consider 12 HRTF independent measurement sets of a Neumann KU-100 dummy head from the international round-robin study Club Fritz. Our analysis of HRTF variabilities focuses on localization performance in elevation within the mid-sagittal plane. A round robin evaluation is performed by means of an auditory model which is able to predict elevation errors and front-back confusion for a given pair of target and template HRTF sets. Results report comparable localization performances between four HRTF databases, suggesting that these acquisition methods led to similar performances in providing elevation cues. Such findings further emphasize the intrinsic complexity and the sensitivity of the HRTF measurement process. The final aim of this study is to certify the quality and repeatability of a measurement process at perceptual level; this findings could be extended to the acquisition of human head acoustics
On the evaluation of head-related transfer functions with probabilistic auditory models of human sound localization
Full text linkUnderstanding spatial hearing leads to implement efficient and effective auralization rendering algorithms with headphones. Two important aspects contribute to sound localization: (i) acoustic filtering of listener body, and (ii) non-acoustic factors introduced by auditory periphery. Accordingly, head-related transfer functions (HRTFs) describe users acoustics in terms of their spatial filtering. Binaural synthesis through generic HRTFs (commonly a dummy head) is the most simple solution for an auralization framework. In this scenario, a high variability in localization tasks between subjects yields to an unreliable rendering. Listener's acoustic and perceptual characterization require HRTF modeling and auditory models predictions in order to provide an effective auralization on individual basis. Systemic comparisons of HRTF approximations and different user profiles can help to predict listener's performances. We consider a case study on both vertical and horizontal localization with different HRTFs and two probabilistic auditory models. In our analysis, spatial audio rendering with non-individual HRTFs has a special attention for its commercial relevance compared to unpractical and questionable use of individual HRTFs
Action planning and affective states within the auditory peripersonal space in normal hearing and cochlear-implanted listeners
No full textFast reaction to approaching stimuli is vital for survival. When sounds enter the auditory peripersonal space (PPS), sounds perceived as being nearer elicit higher motor cortex activation. There is a close relationship between motor preparation and the perceptual components of sounds, particularly of highly arousing sounds. Here we compared the ability to recognize, evaluate, and react to affective stimuli entering the PPS between 20 normal-hearing (NH, 7 women) and 10 cochlear-implanted (CI, 3 women) subjects. The subjects were asked to quickly flex their arm in reaction to positive (P), negative (N), and neutral (Nu) affective sounds ending virtually at five distances from their body. Pre-motor reaction time (pm-RT) was detected via electromyography from the postural muscles to measure action anticipation at the sound-stopping distance; the sounds were also evaluated for their perceived level of valence and arousal. While both groups were able to localize sound distance, only the NH group modulated their pm-RT based on the perceived sound distance. Furthermore, when the sound carried no affective components, the pm-RT to the Nu sounds was shorter compared to the P and the N sounds for both groups. Only the NH group perceived the closer sounds as more arousing than the distant sounds, whereas both groups perceived sound valence similarly. Our findings underline the role of emotional states in action preparation and describe the perceptual components essential for prompt reaction to sounds approaching the peripersonal space
Classifying non-individual head-related transfer functions with a computational auditory model: calibration and metrics
Full text linkThis study explores the use of a multi-feature Bayesian auditory sound localisation model to classify non-individual head-related transfer functions (HRTFs). Based on predicted sound localisation performance, these are grouped into ‘good’ and ‘bad’, and the ‘best’/‘worst’ is selected from each category. Firstly, we present a greedy algorithm for automated individual calibration of the model based on the individual sound localisation data. We then discuss data analysis of predicted directional localisation errors and present an algorithm for categorising the HRTFs based on the localisation error distributions within a limited range of directions in front of the listener. Finally, we discuss the validity of the classification algorithm when using averaged instead of individual model parameters. This analysis of auditory modelling results aims to provide a perceptual foundation for automated HRTF personalisation techniques for an improved experience of binaural spatial audio technologies
The role of spatial perception in auditory looming bias: neurobehavioral evidence from impossible ears
Full text linkIntroductionSpatial hearing enables both voluntary localization of sound sources and automatic monitoring of the surroundings. The auditory looming bias (ALB), characterized by the prioritized processing of approaching (looming) sounds over receding ones, is thought to serve as an early hazard detection mechanism. The bias could theoretically reflect an adaptation to the low-level acoustic properties of approaching sounds, or alternatively necessitate the sound to be localizable in space.MethodsTo investigate whether ALB reflects spatial perceptual decisions or mere acoustic changes, we simulated ears that disrupted spectrospatial associations on the perceptual level while maintaining the original spectrospatial entropy on the acoustic level. We then assessed sound localization, ALB and distance ratings.ResultsCompared to native ears, these novel ears impaired sound localization in both the direction and ego-centric distance dimensions. ALB manifestation also differed significantly between native and novel ears, as evidenced by behavioral discrimination performance and early cortical activity (N1 latency). Notably, the N1 electroencephalographic response closely resembled distance ratings, suggesting a strong link between spatial perception and ALB-related neural processing. Integrating this neural marker into a hierarchical perceptual decision-making model improved explanatory power, underscoring its behavioral relevance.DiscussionThese findings suggest a strong link between the localizability of sounds and their ability to elicit ALB
Bayesian prior uncertainty and surprisal elicit distinct neural patterns during sound localization in dynamic environments
Full text linkEstimating the location of a stimulus is a key function in sensory processing, and widely considered to result from the integration of prior information and sensory input according to Bayesian principles. A deviation of sensory input from the prior elicits surprisal, depending on the uncertainty of the prior. While this mechanism is increasingly understood in the visual domain, much less is known about its implementation in audition, especially regarding spatial localization. Here, we combined human EEG with computational modeling to study auditory spatial inference in a noisy, volatile environment and analyzed behavioral and neural patterns associated with prior uncertainty and surprisal. First, our results demonstrate that participants indeed used prior information during periods of stable environmental statistics, but showed evidence of surprisal and discarded prior information following environmental changes. Second, we observed distinct EEG activity patterns associated with prior uncertainty and surprisal in both the time- and time-frequency domain, which are in line with previous studies using visual tasks. Third, these EEG activity patterns were predictive of our participants' sound localization error, response uncertainty, and prior bias on a trial-by-trial basis. In summary, our work provides novel behavioral and neural evidence for Bayesian inference during dynamic auditory localization
Localization in Elevation with Non-IndividualHead-Related Transfer Functions:Comparing Predictions of Two Auditory Models
Full text linkThis paper explores the limits of human localization of sound sources when listening with non-individual Head-Related Transfer Functions (HRTFs), by simulating performances of a localization task in the mid-sagittal plane. Computational simulations are performed with the CIPIC HRTF database using two different auditory models which mimic human hearing processing from a functional point of view. Our methodology investigates the opportunity of using virtual experiments instead of time- and resource- demanding psychoacoustic tests, which could also lead to potentially unreliable results. Four different perceptual metrics were implemented in order to identify relevant differences between auditory models in a selection problem of best-available non-individual HRTFs. Results report a high correlation between the two models denoting an overall similar trend, however, we discuss discrepancies in the predictions which should be carefully considered for the applicability of our methodology to the HRTF selection problem
- …
