<p>Awareness of the visual environment is a primary aspect of consciousness for most human beings. Predictive coding theories suggest that the contents of our visual awareness are determined through the combination of our prior expectations with the momentary evidence provided by our senses. An important, and untested, tenet of this framework is that the extent to which awareness is shaped by prior expectation relative to sensory stimulation is determined by the precision of each source of information. This is known as precision-weighting. Here, we aimed to investigate whether precision affects the impact of top-down expectation on awareness of an inherently ambiguous stimulus: a structure-from-motion sphere. Across three experiments, we induced expectations for rotation direction by presenting a sphere that had an initial phase of unambiguous rotation before a subsequent phase of ambiguous rotation. The extent to which awareness of the ambiguous phase was biased by the unambiguous phase was measured as an observer’s probability of reporting a switch in rotation between the two phases (lower probabilities indicate a stronger bias). To manipulate precision, we leveraged natural differences in perceptual acuity across the visual field by presenting stimuli either at a fixated location (where precision is high) or in the periphery (where precision is low). In Experiment 1, the two sphere phases on each trial were presented to the same visual field location. This revealed an effect of visual field eccentricity such that switches were more likely to be reported (i.e., awareness was less biased towards the prime) on peripheral compared to fixated trials. While the effect was consistent with precision-weighting, it could not rule out a systematic influence of other factors that might also be specific to each retinal location. In Experiment 2, we aimed to rule out these other factors by having the sphere move from one location to the other so that the ambiguous phase was at a different visual field location to the unambiguous phase. Interestingly, results from this task demonstrated no difference in switch rates, regardless of where in the visual field the sphere started. This finding ruled out our precision-weighting hypothesis since it suggested that the effect of expectation on awareness does not depend on the precision with which it was encoded. Next, because visual adaptation is known to play an important role in perception of bistable stimuli, we hypothesised that differences in adaptation to motion signals between the two visual field locations might explain the eccentricity effect observed in Experiment 1. To test this hypothesis, in Experiment 3, we successfully replicated the effect of eccentricity on switch rates and measured the extent to which the viewing the unambiguous sphere causes motion aftereffects using a perceptual nulling task. Confirming our expectation, stronger motion aftereffects were observed for peripheral compared to fixated vision. However, contrary to our hypothesis, this effect was not predictive of the effect of eccentricity on switch rates, and thus is unlikely to explain it. Finally, possible alternative explanations for the eccentricity effect are discussed, as well as the implications of our results for the predictive coding framework. This includes a consideration of more radical versions of predictive coding theory, under which precision can be understood differently.</p>