Neurobiological Signatures of Auditory False Perception and Phantom Perception as a Consequence of Sensory Prediction Errors.

Abstract

Simple SummaryThe principle of Bayesian inference provides a theoretical framework for stable perception in numerous tasks, including sensory–perceptual tasks, sensorimotor, and motor tasks. Efference copy (EC) signals enable organisms to reduce cognitive loading by decreasing the sensory processing of their own actions. In the auditory domain, the sensorimotor prediction error is responsible for false perceptions, such as auditory hallucinations in schizophrenia, while sensory prediction error also leads to auditory phantom perception such as tinnitus. Attenuation of the N1 component in event-related potentials has been suggested as evidence for the integrity of the EC mechanism. The current study investigated the EC mechanism in auditory false perception and phantom perception as a consequence of sensory prediction errors. N1 attenuation failures were present in tinnitus patients with significant hearing impairment and in those with schizophrenia, indicating that phantom perception after sensory deprivation might lead to impairment in the EC mechanism. However, the corresponding neural representation with spatiotemporal patterns presented differently between patients with schizophrenia and tinnitus. Although the present study had several constraints, the results provide a new perspective on neurobiological aspects of abnormal auditory perception resulting from deficits in predictive coding.In this study, we hypothesized that top-down sensory prediction error due to peripheral hearing loss might influence sensorimotor integration using the efference copy (EC) signals as functional connections between auditory and motor brain areas. Using neurophysiological methods, we demonstrated that the auditory responses to self-generated sound were not suppressed in a group of patients with tinnitus accompanied by significant hearing impairment and in a schizophrenia group. However, the response was attenuated in a group with tinnitus accompanied by mild hearing impairment, similar to a healthy control group. The bias of attentional networks to self-generated sound was also observed in the subjects with tinnitus with significant hearing impairment compared to those with mild hearing impairment and healthy subjects, but it did not reach the notable disintegration found in those in the schizophrenia group. Even though the present study had significant constraints in that we did not include hearing loss subjects without tinnitus, these results might suggest that auditory deafferentation (hearing loss) may influence sensorimotor integration process using EC signals. However, the impaired sensorimotor integration in subjects with tinnitus with significant hearing impairment may have resulted from aberrant auditory signals due to sensory loss, not fundamental deficits in the reafference system, as the auditory attention network to self-generated sound is relatively well preserved in these subjects.