This project was supported by the German Research Foundation, International Research Training Group, IRTG 1901, “The Brain in Action”.
Our (-)-JQ1 must reconstruct the outside visual world from a sensory evidence that is always incomplete and is always intrinsically ambiguous (Gregory, 2009; Metzger, 2009; Yuille & Kersten, 2006). To make things worse retinal input constantly changes due to self-motion, changes in illumination, object-motion, etc. This poses an enormous challenge, as very different physical changes can produce identical changes in sensory evidence. An object changing its size (an inflated frog) and an object getting closer (you are walking toward the frog) could produce the same change in sensory evidence (a change in the size of a retinal image) (Combe & Wexler, 2010; Koenderink, 1986). A change of a retinal projection’s shape may imply that object moved (a leaf was moved by a wind), that you moved (you walked past the tree), or some combination of both (a leaf was moved by a wind as you were walking past the tree) (Wexler, Panerai, Lamouret, & Droulez, 2001). An activation-pattern of cone cells on the retina that corresponds to somebody’s face may change because the person blushed (surface has changed) or because the person stepped from a direct sunlight into an ambient illumination in the shadow or because a cloud obstructed the sun (illumination has changed) (Jameson & Hurvich, 1989). Ambiguity of change in sensory evidence makes it hard for the perceptual system to identify a unique physical cause and to determine whether constancy of a particular visual feature must be maintained. Yet, this unique physical cause is all that matters and is what our visual system is trying to correctly represent in perception.
None of the examples above correspond to a rare and exceptional event. On the contrary, they are the norm for the dynamic environment that we actively explore and which is full of objects, animals, clouds, wind, etc. It is the ubiquity of these events that raises the question of how the visual system resolves their dynamic ambiguity. The general answer to the problem is to gather and exploit prior knowledge (Friston, Breakspear, & Deco, 2012; Gregory, 2009; Metzger, 2009; Yuille & Kersten, 2006), and this process has been well studied from both behavioral (Kristjánsson & Campana, 2010; Pastukhov & Braun, 2011, 2013b) and theoretical (Friston et al., 2012; Pastukhov, García-Rodríguez, et al., 2013) perspectives, even though the neural implementations are still poorly understood (Daw, O’Doherty, Dayan, Seymour, & Dolan, 2006). The main focus of prior research was the knowledge about physical states (Hansen, Olkkonen, Walter, & Gegenfurtner, 2006; Weiss, Simoncelli, & Adelson, 2002; Yang & Purves, 2003; Yuille & Kersten, 2006), however this type of knowledge serves only as a weak constraint because the number of transformations by far outstrips the number of states.
Accordingly, our visual system also relies on the knowledge about physical transformations (in addition to, and independent of, the similar information on physical states) to determine the most likely cause of a change in sensory evidence. Because of that in examples above certain transformations are more likely to be perceived than other. Previous work of on transformation priors (Barbur & Spang, 2008; Combe & Wexler, 2010; Pastukhov, Vonau, & Braun, 2012; Tse, 2006; Tse & Logothetis, 2002; Wexler & van Boxtel, 2005) demonstrated their importance to the dynamic perception and their link to ecological constraints of the outside world. Present work extends this by asking the question whether this prior knowledge is gathered from the recent perceptual experience or can be considered to be static.
Materials and methods
Our results provide further evidence for the importance of prior knowledge about physical transformations for the visual system by showing that it gathers the information from recent perceived changes to update prior knowledge about physical ones. Critically, this repetition priming was observed for three disparate displays, whose perception relies on distinct neural representations. In addition, the specificity of priming speaks against a notion of a central mediator, but in favor of a locally implemented canonical mechanism. Taken together this suggests that both knowledge about physical transformations and its dynamic update reflect an overall adaptive strategy of the visual system and are an integral part of a general perception. Accordingly, the influence of dynamic prior knowledge must be taken into account when studying dynamic visual scenes.