The role of similarity in restoring missing notes in music
Year: 2015
Type: dissertation
Abstract: In noisy environments, the auditory system is often presented with incomplete information. Many mechanisms facilitate perception in these noisy environments, including auditory restoration, a mechanism that perceptually fills in gaps in sounds. The information restored with auditory restoration ranges spectral-temporally from simple (e.g., steady tones) to complex (e.g., words). This dissertation proposes a theory whereby restoration is a similarity-based mapping from a disrupted stimulus to an internal percept. It is argued that this process is biased to restore monotonic spectral-temporal patterns (i.e., a pattern that does not change from increasing to decreasing frequency or the reverse), even if the original stimulus were non-monotonic. Three experiments were conducted to test this theory. The first experiment, using a similarity judgment paradigm, demonstrated the perceptual salience of the monotonic/non-monotonic dichotomy for the tunes used in the other experiments. The second experiment trained participants to identify a subset of the tunes from the first experiment and then asked them to distinguish between a tune containing a veridical note added to noise and a tune containing only a restored note occurring with the noise. The results indicate that listeners have considerable difficulty distinguishing a veridical and a restored note and that the difficulty increases the closer the veridical and restored notes are spectrally. Additionally, this experiment provided evidence that, even when the portion of the tune replaced by noise was originally non-monotonic, participants generally restored a monotonic pattern. The third experiment also trained participants on a subset of the tunes from the first experiment and then asked them to identify a tune when a note was removed and replaced by noise or silence.
Author: Aronoff, Justin (author)
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2520697642