This innovative framework is set to improve how researchers investigate sound perception. It tackles a key challenge in auditory science: the demand for a standardized approach to modeling how we process and react to sound.

Scientists often rely on auditory models to describe listener’s behavior and its underpinning in the acoustic environment. However, despite seemingly similar statistical methods, many research teams develop custom-built models for each study. This one-off approach makes it hard to compare results across different studies, limiting progress in the field.

FrAMBI (Framework for Auditory Modeling Based on Bayesian Inference) provides a standardized structure for implementing an auditory model. It follows the perception-action cycle and enables the automatic application of statistical analysis with behavioral data. Using Bayesian inference, a statistical method that helps refine predictions based on prior knowledge, FrAMBI allows researchers to create flexible and scientifically rigorous models.

The paper showcases FrAMBI’s capabilities with several examples of increasing complexity in sound source localization.

FrAMBI is integrated into the widely used Auditory Modeling Toolbox (AMT), ensuring long-term maintenance and expansion, therefore making it a valuable resource for researchers.

By standardizing auditory research, FrAMBI fosters collaboration, simplifies result comparisons, and accelerates scientific progress.

Read the full Neurinformatics journal article →HERE.