Aarhus Universitets segl

Harmonic and inharmonic sounds: which ones elicit greater prediction errors?

Publikation: Bidrag til bog/antologi/rapport/proceedingKonferenceabstrakt i proceedingsForskning

Our brain is constantly making predictions based on models that anticipate what will occur in the world surrounding us (Friston, 2010). In audition, pitch is important for object identification and categorization, from human voices to environmental sounds. Most of the natural sounds we experience everyday are harmonic, like animal vocalizations or musical instruments. These sounds are complex tones that follow the harmonic series, comprising a fundamental frequency (f0) and integer multiples of it called harmonics (e.g. 2f, 3f…). In contrast, inharmonic sounds deviate from the harmonic series because the harmonics are not integer multiples. In this electroencephalographic study, we investigate differences in the neural responses predicting harmonic and inharmonic sounds. We hypothesize that violations of predictions for harmonic sounds elicit stronger responses than violations of predictions for inharmonic sounds. Using a 32-electrode active system, we measured the neural responses in an auditory oddball-roving paradigm eliciting mismatch negativities. Participants freely consented to participate in the study, approved by our Institutional Review Board. They watched a silent movie and listened to a series of artificially generated complex tones presented in trains of stimuli that changed in pitch after some repetitions. There were three counterbalanced conditions: harmonic sounds, inharmonic sounds with the same jitter rates applied to the harmonics, and inharmonic sounds with distinct jitter rates (similar to McPherson & McDermott, 2018). Our preliminary analyses showed differences in the MMNs of the three conditions in line with our hypothesis. There were higher MMNs for violations of harmonic sounds. Furthermore, the MMN of violating inharmonic sounds was greater when the jitter rate was constant. In sum, these early findings suggest that harmonic sounds have higher precision-weighting of the prediction errors. We will present the full dataset in the conference and discuss the implications of a predictive coding of pitch.
TitelMMN 2022 : Mismatch negativity: neural basis and biomarker development
Antal sider2
Udgivelsesårsep. 2022
StatusUdgivet - sep. 2022
BegivenhedThe 9th Mismatch Negativity conference - Fukushima Medical University, Fukushima, Japan
Varighed: 21 sep. 202223 sep. 2022
Konferencens nummer: 9


KonferenceThe 9th Mismatch Negativity conference
LokationFukushima Medical University

Se relationer på Aarhus Universitet Citationsformater

ID: 305596999