Josh H. McDermott

This study shows that perceptual sensitivity thresholds in humans can be explained by a computational model optimized for ecological sound tasks, matching human behavioral thresholds.

The paper tests cross-cultural consistency in pitch perception by comparing US and Tsimane participants, finding shared patterns despite cultural differences in pitch-related behavior.

Neural networks trained to detect sound categories in synthetic multi-source scenes approach human accuracy on the EnvAudioEval behavioral benchmark and track human patterns better than cochlear or cortical filterbank baselines.

The paper shows that music-selective brain regions respond more strongly to music with intact pitch and temporal structure compared to scrambled versions, indicating their sensitivity to both note-level and pattern-level musical structure.

This study uses an artificial neural network with optimized feature gains to model human selective listening, showing it matches human performance and predicts attentional effects.

The paper shows humans can infer physical properties from impact sounds but struggle when object resonance distributions differ from real-world data.

The paper develops task-optimized models of sensory uncertainty and shows human confidence judgments align with model-derived uncertainty estimates in sound localization and pitch perception.

This paper develops a deep learning framework to evaluate cochlear implant performance by simulating auditory nerve input and testing speech recognition and sound localization.

The paper shows that precise spike timing in auditory nerve models improves sound localization and speech perception, but the required precision varies with task demands.

The paper investigates how human hearing adapts to background noise, finding that internal models of noise structure, or 'noise schemas', improve the detection and localization of foreground sounds in noisy environments.

This study uses generative models to explain auditory perception, showing they can account for illusions and real-world sounds with human-like organization.

The paper evaluates how well deep neural networks model auditory brain responses and finds that some models better predict brain activity than others, with task-specific effects on prediction quality.

This study investigates how timbre affects relative pitch judgments and finds that pitch accuracy is reduced when sounds differ in timbre, even when judgments are based on fundamental frequency.

The paper identifies privileged representational axes in both biological and artificial neural networks, showing alignment between DCNNs and perceptual cortices and benefits of axis-sensitive coding.

This study shows that harmonic frequency relations improve hearing in noise for both musicians and non-musicians, aiding detection and discrimination of sounds.

This study uses a deep neural network with human ears to model sound localization in realistic environments, showing how biological hearing adapts to real-world challenges.

The paper generates model metamers to compare invariances in neural networks and humans, finding discrepancies and proposing a new benchmark for model assessment.

The paper shows that music-selective neural populations exist in individuals without musical training and respond to various music types and rhythms.

The paper shows that harmonic sounds are better discriminated when separated in time, suggesting memory-based compression relies on fundamental frequency.

The paper uses intracranial recordings to show a neural population in human auditory cortex selectively responsive to musical song.

The paper identifies auditory grouping cues by analyzing natural sound statistics, revealing both known and novel principles of perceptual grouping.

The paper demonstrates that the brain extrapolates statistical properties of interrupted background sounds over several seconds, creating illusory textures that are perceived as real.

The paper tests a sensory model of auditory cortex by comparing neural responses to natural and model-matched stimuli, finding divergent responses in non-primary regions that suggest higher-order computations.

The paper investigates species-specific differences in pitch perception by comparing ferrets and humans, finding that cochlear filter bandwidths influence the salience of temporal envelope and harmonic cues.