Manipulable object processing reveals distinct neural and behavioral signatures for visual, functional, and manipulation properties
摘要
Our interactions with objects involve processing a range of object-associated features to assess whether an object can fulfill our intentions. For human-made manipulable objects, these features mainly encompass three interconnected types of knowledge: visual appearance, manner of manipulation, and functional purpose. We investigated these components by breaking them down into their features and exploring how each type of object-related information is processed. Using a release-from-adaptation paradigm - we tested vision, manipulation, and function in three behavioral experiments (21, 20, and 22 participants) and three fMRI experiments (20 participants each) - to explore whether the similarity between objects within each knowledge type impacts behavioral and neural responses while controlling for the other knowledge types. Our findings suggest that an object’s visual, functional, and manipulation properties are processed independently in distinct brain areas, including the fusiform gyrus and collateral sulcus, the lateral occipitotemporal cortex, and regions within the dorsal stream. Moreover, object similarity shapes how information is organized within each knowledge type and affects the ability to detect changes between objects. Importantly, the brain may follow a strategy of breaking down the incoming sensory stimulus into different knowledge types and properties in the process of building bottom-up representations that can then serve object recognition. Nevertheless, interacting with objects requires integrating these knowledge types, with our data suggesting that the medial fusiform gyrus and collateral sulcus might be important candidates for this integration. Once integrated, information may be transmitted to parietal and frontal areas to achieve a successful interaction with the object.