During encoding, participants navigated a city consisting of eight stores randomly distributed around a circle in a virtual environment. They made a defined series of deliveries of a passenger to a store, with the spatial layout of the city uncorrelated with the temporal order of deliveries to ensure dissociation between the contextual elements of an episode (Fig. 1A). During retrieval, participants performed a spatiotemporal contextual retrieval task while being scanned. For each trial, participants were first asked whether they recognized a previously encountered store (item) compared to lure stores. To assay context memory, participants were presented with an additional question after each item recognition question regarding the contextual details associated with that item. Each trial was followed by a baseline task of variable duration based upon the jittered inter-stimulus interval (Fig. 1B). The experimental design is further described in Ekstrom et al. (2011), where this dataset was previously analyzed using a univariate approach.
