Abstract
Adaptation to sound-level statistics is thought to be crucial for optimal perception, but research has focused on neurophysiological recordings, mostly in non-human mammals. Behavioral evidence for adaptation to sound-level statistics is sparse. Here we use electroencephalography (EEG) and behavioral methods to investigate how the statistics of sound-level distributions affect neural activity in auditory cortex and the detection of near-threshold changes in sound amplitude. We presented noise bursts with sound levels drawn from distributions with either a low (15 dB SL) or a high (45 dB SL) modal sound level. One group of participants listened passively to the stimulation while EEG was recorded (Experiment I). A second participant group performed a behavioral amplitude-modulation detection task (Experiment II). Neural responses to noise bursts and sensitivity to amplitude modulation depended on sound-level statistical context: Consistent with an account positing that the sensitivity of neurons to sound intensity changes with ambient sound level, neural responses and amplitude-modulation sensitivity (d’) for noise bursts at moderate intensities were larger for low ambient sound levels compared to high ambient sound levels. Neural activity appears to adapt to sound-level statistics in humans, perhaps fine-tuning perceptual sensitivity to optimize detection of subtle changes in sound amplitude.