Abstract
Contextual violations have long been known to cause deficits in object detection and recognition (Biederman et al., 1982). Incongruent objects attract earlier and longer eye fixations (Underwood & Foulsham 2006) and evoke stronger ERPs (Mudrik, Lamy, and Deuoell 2010), suggesting that the brain rapidly marshals additional resources to aid in processing unexpected objects. Despite a wealth of psychophysical results on context, cortical models of contextual facilitation are still speculative (Bar, 2004). In this study, we use MVPA methods with fMRI data to explore the effects of context on neural representations. MVPA allows us to assess the quality of the representation as opposed to simple changes in overall activity afforded by more traditional fMRI analysis methods.
First, we showed that a classifier trained using lateral occipital complex (LOC) responses to isolated boats and cars generalized (achieved above-chance decoding accuracy) to the same objects placed in scenes. We then presented these objects in scenes that violated a semantic relationship (e.g. a boat sitting on a city street) and/or a geometric relationship (e.g. a car flying over a city street). Although these violations had similar slowing effects on reaction time in a detection task, they had very different impacts on object representations in LOC. Decoding performance decreased to chance when a geometric relationship was violated, but actually increased when a semantic relationship was violated. These results suggest that context not only impacts object detection but influences object representations in complex ways, presumably via interactions with other cortical areas. We explore decoding rates from other ROIs to further explore these contextual interactions.
NIH R01 EY019429 (D.B. & L.F-F) NSF GRFP (C.B.) Stanford SGF (M.C.I.).