The observation that events compete during learning (a.k.a. cue competition phenomena) has been widely reported and these phenomena have become a gold standard that theories of learning need to account for. However, in some domains (i.e., spatial cognition) these effects have not been reliably observed, and recent discrepancies in the literature suggest that the observation of competition may occur under a restricted set of parameters. In this talk, I will present recent work from our laboratory in which we investigated the role of temporal and spatial distance (i.e., contiguity) in predictive, action-outcome and spatial learning domains. In all these domains, we reliably observed competition when events were close in space and time, but no competition (and sometimes the opposite, that is facilitation) when events were discontiguous. Overall, these results suggest that contiguity is necessary for competition between events, and hence that we need to better account for contiguity effects in computational models.

Biography: