AI Insight
Researchers developed a computational model optimized for everyday sound tasks like localization and recognition, then tested its ability to discriminate between sounds that differ in properties like intensity, frequency, or orientation. The model's discrimination thresholds across 1632 different perceptual tests closely matched those of humans, suggesting that human perceptual limitations arise from representations optimized for real-world tasks rather than from random neural noise. This challenges traditional explanations that attribute perceptual limits primarily to internal biological noise.
Why it matters
This finding fundamentally reframes our understanding of human sensory limitations, suggesting they reflect efficient adaptation to natural environments rather than inherent flaws. The approach could inform the development of more human-like artificial sensory systems and help predict human perceptual abilities in new contexts without extensive behavioral testing.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
When asked to discriminate between two stimuli that differ along some dimension, such as intensity, frequency, or orientation, humans exhibit a smallest stimulus difference, or threshold, below which discrimination becomes impossible. These limits on perceptual sensitivity have traditionally been explained by internal noise. Contemporary theories instead link perceptual sensitivity to efficient coding of the natural environment but have not been comprehensively tested. Here, we measured the psychoacoustic sensitivity of a computational model optimized for everyday sound localization and recognition tasks. We used linear classifiers operating on the model’s learned representations to estimate 1632 thresholds spanning the perceptual dimensions of sound. These model-derived thresholds closely matched thresholds measured behaviorally in humans on the same test battery. The results show that many classical psychophysical phenomena emerge from representations optimized for natural behavior and suggest that discrimination thresholds are determined by linear separability in these optimized representations, rather than intrinsic neural noise.
Source: Perceptual sensitivity is explained by optimization for ecological tasks