Predictive coding and stochastic resonance as fundamental principles of auditory phantom perception

Area of Science:

• Neuroscience
• Auditory Perception
• Artificial Intelligence

Background:

• Tinnitus serves as a model for understanding auditory phantom perceptions and healthy auditory processing.
• Hearing loss is common in tinnitus patients, but not all individuals with hearing loss experience tinnitus.

Approach:

• Reviewing interdisciplinary research at the intersection of artificial intelligence, psychology, and neuroscience.
• Developing computational models to explain the mechanisms of auditory phantom perception.
• Applying the Bayesian brain framework to formalize the emergence of tinnitus.

Key Points:

• Intrinsic neural noise amplification via adaptive stochastic resonance acts as a compensatory mechanism for hearing loss.
• The Bayesian brain framework explains how altered neural signal precision can lead to misinterpretation of sensory evidence.
• Predictive coding (top-down) and adaptive stochastic resonance (bottom-up) are key principles underlying auditory phantom perceptions.

Conclusions:

• Auditory phantom perceptions, like tinnitus, result from a complex interplay of neural noise, compensatory mechanisms, and predictive processing.
• These principles are crucial for understanding healthy auditory perception.
• The identified mechanisms hold potential for advancing neuroscience-inspired artificial intelligence and machine learning.