Behavioral and Neural Effects of Familiarization on Object-Background Associations

Abstract

Associative memory is the ability to link together components of stimuli. Previous evidence suggests that prior familiarization with study items affects the nature of the association between stimuli. More specifically, novel stimuli are learned in a more context-dependent fashion than stimuli that have been encountered previously without the current context. In the current study, we first acquired behavioral data from 62 human participants to conceptually replicate this effect. Participants were instructed to memorize multiple object-scene pairs (study phase) and were then tested on their recognition memory for the objects (test phase). Importantly, 1 day prior, participants had been familiarized with half of the object stimuli. During the test phase, the objects were either matched to the same scene as during study (intact pair) or swapped with a different object's scene (rearranged pair). Our results conceptually replicated the context-dependency effect by showing that breaking up a studied object-context pairing is more detrimental to object recognition performance for non-familiarized objects than for familiarized objects. Second, we used functional magnetic resonance imaging (fMRI) to determine whether medial temporal lobe encoding-related activity patterns are reflective of this familiarity-related context effect. Data acquired from 25 human participants indicated a larger effect of familiarization on encoding-related hippocampal activity for objects presented within a scene context compared to objects presented alone. Our results showed that both retrieval-related accuracy patterns and hippocampal activation patterns were in line with a familiarization-mediated context-dependency effect.

Keywords: associative memory; context-dependent; functional magnetic resonance imaging; hippocampus; medial temporal lobe.

Figure 1

Experimental paradigm for Familiarization on Day 1 and Encoding and Recognition on Day 2, for both experiments. On Day 1, participants were familiarized with exemplars from 10 object categories. On Day 2, participants memorized object-scene pairs (Experiments 1 and 2) and objects alone (Experiment 2). Afterwards, participants were tested on how well they could recognize old objects (i.e., objects shown during Encoding) from new objects. The background scenes during recognition were either matched to the same object as in Encoding (“intact”), swapped with a different object’s background (“rearranged”), or were with a new object entirely (“new object with old scene”). Half of the objects in Encoding and Recognition were familiarized (i.e., shown in the Familiarization stage) or non-familiarized. Note that, in Experiment 2 (highlighted by the gray box), there was an additional distraction task in between Encoding and Recognition, where participants subjectively matched images that shared either global or local geometric properties.

Figure 2

Experiment 1 object recognition accuracy for different pairings with background scenes. (A) The hit rates and false alarms (FAs) are shown for participants responding to whether objects presented in the test phase had been presented in the previous study phase. Objects could be familiarized (blue) and non-familiarized (orange), and could be presented in intact, rearranged, and new-object/old-scene pairings. (B) Signal detection performance is indicated by D prime (d’) scores, using the corresponding new object condition, to determine the false alarm rate. Error bars indicate standard error of the mean.

Figure 3

Results for the medial-temporal volume of interest analysis. (A) Coronal, (B) Axial, and (C) Sagittal (left) MRI brain slices depicting a bilateral posterior hippocampal regions that showed a significant interaction effect for the factors of Familiarization (non-familiarized vs. familiarized) and Arrangement (single object vs. pair).

Figure 4

Parameter estimates (±1 SE) from the interaction analysis shown separately for the four conditions and for left and right hippocampus. *p < 0.05.