The role of active inference in conscious awareness.

Active inference, a first-principles framework for modelling the behaviour of sentient agents, is beginning to be applied in consciousness research. One hypothesis arising from the framework is that active inference is necessary for changes in conscious content. As one component of an extensive adversarial collaboration among competing theories of consciousness, active inference will be contrasted with two other theories of consciousness, neither of which posit that active inference is necessary for consciousness. Here, we thus present a Study Protocol designed to test the active inference hypothesis using a carefully controlled adaptation of the motion-induced blindness paradigm, where an 'active' condition with richer active inference is contrasted with a 'passive' condition. In the active condition, participants direct their gaze towards a target stimulus following its disappearance from consciousness, and report on its subsequent reappearance. In the passive condition, participants maintain central fixation, while the stimulus array is moved across the visual field (in a replay of the active condition based on eye-tracking data acquired during active trials). In two experiments, we plan to investigate target reappearance across active and passive conditions to evaluate the contribution of active inference to conscious awareness. Results will eventually be considered in the context of all the experiments conducted as part of the overall adversarial collaboration.