Predictive coding in auditory perception: challenges and unresolved questions

Abstract

Predictive coding is arguably the currently dominant theoretical framework for the study of perception. It has been employed to explain important auditory perceptual phenomena, and it has inspired theoretical, experimental and computational modelling efforts aimed at describing how the auditory system parses the complex sound input into meaningful units (auditory scene analysis). These efforts have uncovered some vital questions, addressing which could help to further specify predictive coding and clarify some of its basic assumptions. The goal of the current review is to motivate these questions and show how unresolved issues in explaining some auditory phenomena lead to general questions of the theoretical framework. We focus on experimental and computational modelling issues related to sequential grouping in auditory scene analysis (auditory pattern detection and bistable perception), as we believe that this is the research topic where predictive coding has the highest potential for advancing our understanding. In addition to specific questions, our analysis led us to identify three more general questions that require further clarification: (1) What exactly is meant by prediction in predictive coding? (2) What governs which generative models make the predictions? and (3) What (if it exists) is the correlate of perceptual experience within the predictive coding framework?

Evaluating the neurophysiological evidence for predictive processing as a model of perception

“…41 for a recent review), their precise relationships to the modulations observed in sensory processing are not yet established. Finally, it is not clear whether the formation of sensory predictions is a unitary neural process or an array of independent, task‐tailored mechanisms 161 . Expectations can be formed in relation to stimulus timing, stimulus location, or stimulus content, 162 and electrophysiological evidence suggests that these kinds of predictions are instantiated by distinct neuromodulatory mechanisms, in dissociable networks and at different latencies 44,162,163 .…”

“…Overall, this evidence is consistent with PP's contention that the brain constructs dynamic representations of regularities in the sensory stream and that these representations are intimately involved in subsequent sensory processing. 144,154,159 Hypothesis 2: Summary and conclusions Humans are capable of rapidly extracting regularities from the sensory environment 144,158,160,161 and there is strong evidence that the resultant expectations influence sensory processing. 154 It is a substantial achievement to begin to isolate this predictive activity in recordings of neural activity spanning human, monkey, and rodent research.…”

“…Finally, it is not clear whether the formation of sensory predictions is a unitary neural process or an array of independent, task-tailored mechanisms. 161 Expectations can be formed in relation to stimulus timing, stimulus location, or stimulus content, 162 and electrophysiological evidence suggests that these kinds of predictions are instantiated by distinct neuromodulatory mechanisms, in dissociable networks and at different latencies. 44,162,163 Similarly, expectations can arise from arbitrary stimulus pairings, 90 predictive cues, or higher order regularities.…”

“…Although this may reflect on the veracity of the theory, it may also be attributable to the great difficulty such a task poses. 161 It is not sufficient to demonstrate that descending connections to early areas introduce contextual modulations that are essential for perception because this is accommodated by traditional models of visual processing (e.g., Ref. 131).…”

Cross-modal Suppression of Auditory Association Cortex by Visual Speech as a Mechanism for Audiovisual Speech Perception

“…Predictive coding is a well-established principle at all levels of the auditory system (Denham and Winkler, 2017). Cross-modal suppression may result from similar mechanisms as predictive coding, with the difference that the information about the expected auditory stimulus does not come from previously-presented auditory stimuli but from early-arriving visual speech information.…”