The Active Bayesian Brain and the Rorschach Task

Abstract. The Rorschach offers a unique and interesting paradigm from the perspective of the (Bayesian) brain. This contribution to the cross-disciplinary special issue considers the Rorschach from the perspective of perceptual inference in the brain and how it might inform subject-specific differences in perceptual synthesis. Before doing so, we provide a broad overview of active inference in its various manifestations. In brief, active inference supposes that our perceptions are the best hypothesis to explain sensory impressions. On a Bayesian account, the requisite belief updating rests sensitively upon the precision or confidence ascribed to sensory input, relative to prior beliefs about the causes of sensations. This focus – on the balance between sensory and prior precision – has been a useful construct in both cognitive science (e.g., as a formal explanation for attention) and neuropsychology (e.g., as a formal explanation for aberrant or false inference in hallucinations). In this setting, false inference is generally understood as abnormally high precision afforded to high-level hypotheses or explanations for visual input, which may compensate for a failure to attenuate sensory precision. On this view, the Rorschach offers an interesting paradigm because the amount of precise information about the causes of visual input is deliberately minimized – and rendered ambiguous – thereby placing greater emphasis on prior beliefs entertained by the respondent. We close by exploring this issue and several other areas of intersection between Rorschach responding and active inference.