Effects of expectations and sensory unreliability on voice detection – a preregistered study

“Hearing voices” is a phenomenon prevalent not only in psychotic disorders, but also in the general population. However, it is unclear under what conditions people detect voices that are not really there. In our preregistered study, we investigated a possible mechanism of such experiences, grounded in a predictive processing account of agency detection. The model, in line with previous research on predictive processing and hearing, predicts that detecting auditory agent stimuli, such as voices, depends both on one’s prior expectations and unreliability of the sensory data. We tested this model using a signal detection task, manipulating sensory noise (within subjects) and expectations regarding the number of voices embedded in auditory stimuli (between subjects). We hypothesized that expecting less or more voices than actually present would drive the response bias towards a conservative and liberal response strategy, respectively. Moreover, we expected that including sensory noise would enhance these expectancy effects. We corroborated the first hypothesis and found our data to align with the general predictions of the model. We also report the results of exploratory analyses concerning absorption, supernatural beliefs, and participants’ subjective response confidence. Our findings provide support for a predictive processing model, showing that one detects more or less voices than actually present depending on their expectations and that this effect is especially pronounced in the case of unreliable data. As such, the study contributes to our understanding of the predictive processing approach to hearing and the mechanism behind voice hearing experiences.