Predictive encoding of pure tones and FM-sweeps in the human auditory cortex

Abstract

Expectations substantially influence perception, but the neural mechanisms underlying this influence are not fully understood. A prominent view is that sensory neurons encode prediction error with respect to expectations on upcoming sensory input. Although the encoding of prediction error has been previously demonstrated in the human auditory cortex (AC), previous studies often induced expectations using stimulus repetition, potentially confounding prediction error with neural habituation. These studies also measured AC as a single population, failing to consider possible predictive specializations of different AC fields. Moreover, the few studies that considered prediction error to stimuli other than pure tones yielded conflicting results. Here, we used fMRI to systematically investigate prediction error to subjective expectations in auditory cortical fields Te1.0, Te1.1, Te1.2, and Te3, and two types of stimuli: pure tones and frequency modulated (FM) sweeps. Our results show that prediction error is elicited with respect to the participants’ expectations independently of stimulus repetition, and similarly expressed across auditory fields. Moreover, despite the radically different strategies underlying the decoding of pure tones and FM-sweeps, both stimulus modalities were encoded as prediction error in most fields of AC. Together, our results provide unequivocal evidence that predictive coding is the general encoding mechanism in AC.

Cortical haemodynamic responses predict individual ability to recognise vocal emotions with uninformative pitch cues but do not distinguish different emotions

“…It is possible that these demands reflect inter‐listeners differences in prediction error—a measure of the distance between the predicted and true sensory input, as encoded by neurons (de Lange et al, 2018 ; Friston, 2005 ). Elevated cortical activity evoked by changes in the frequency of pure tones or in the frequency profile of sounds (Rubin et al, 2016 ; Stein et al, 2021 ) may reflect neural signature of a prediction error. We propose that listeners who rely more heavily on F0 cues (i.e., are less able to extract emotional meaning from other cues) exhibit evidence of increased prediction error in right STG when F0 cues are uninformative.…”

“…Although the enhanced activity in right STG for happy relative to sad speech with uninformative F0 cues was initially unexpected, closer examination of the acoustic profiles of the two emotions in the natural conditions indicates that happy speech showed more variance in F0 than sad speech, and the mean F0 for happy speech was further from the target value (217 Hz) at which the F0 was held constant than the mean for sad speech. We propose that the larger amplitudes in cortical haemodynamic responses in right STG observed for happy than sad speech when F0 cues are attenuated indicate a difference in prediction error (Friston, 2005;Stein et al, 2021), mediated by the relative distance between the original and manipulated acoustic profiles for each emotion. That is, we propose that the difference in richness of F0 cues between happy with and without attenuated F0 cues is larger than the difference between the two versions of sad, and that greater prediction error, observed in enhanced haemodynamic activity, likely reflects the relative reduction in the richness of F0 cues.…”

“…It is possible that these demands reflect inter-listeners differences in prediction error-a measure of the distance between the predicted and true sensory input, as encoded by neurons (de Lange et al, 2018;Friston, 2005). Elevated cortical activity evoked by changes in the frequency of pure tones or in the frequency profile of sounds (Rubin et al, 2016;Stein et al, 2021)…”