Visual Integration and the Role of Structural and Functional Brain Changes in the Age-Related Associative Memory Deficit

Abstract

In the present set of experiments, we investigated the effects of visual integration, and age-related changes in brain structure and function on associative memory for objects and scenes. The results of experiment 1 demonstrated improved associative memory when objects and scenes were visually integrated, through an increase in hits without a corresponding increase in false alarms. This profile was interpreted to reflect an increase in recollection-based responding when associative pairs were visually integrated. Although processing of visually integrated stimuli resulted in greater activation across medial temporal lobe (MTL) structures (e.g. hippocampus (HC), perirhinal cortex (PRC), parahippocampal cortex (PHC)) compared to non-integrated pairs and their respective controls, no difference in MTL or prefrontal (PFC) activation was observed after accounting for control activation. In young adults, activation across all three MTL structures predicted discrimination for non-integrated pairs, but only PRC activation predicted discrimination for visually integrated pairs. In older adults, MTL activation was not related to performance for non-integrated pairs, but HC and VLPFC activation predicted performance for visually integrated pairs. In experiment 2, we investigated differences in reaction time based on visual integration, and the neural correlates of associative recognition. Both young and older adults responded faster to visually integrated than non-integrated pairs, reflecting increased recollection-based responding. However, the difference between conditions was greater in older adults. While both groups exhibited a reduction in right HC and left PRC activation during recognition of visually integrated compared to non-integrated pairs, an interaction occurred in left medial HC. Young adults exhibited reduced activation in this region for visually integrated pairs, whereas high functioning older adults activated this region more for visually integrated than non-integrated pairs. Similar to results from encoding (Experiment 1), HC activation was associated with recognition of non-integrated pairs in young adults, but no MTL region predicted performance for visually integrated pairs. In contrast, HC and PHC activation was only related to memory for visually integrated pairs in older adults. Contrary to findings from the verbal unitization literature that demonstrate improved memory in older adults through increased familiarity-based responding (Zheng et al., 2015; Ahmad et al., 2015), our findings suggest that visual integration improves performance across age-groups through an increased reliance on recollection. Notably, the neural correlates of this shift vary based on age. Finally, in experiment 3, we examined the role of white matter integrity in the tracts connecting frontal and temporal brain regions in predicting associative memory. White matter integrity in the fornix, uncinate fasciulucs, and PHC cingulum predicted associative memory in older adults, even after controlling for global white matter changes. The present findings demonstrate the benefit of visual integration as a strategy to improve associative memory across age groups. Further, we identified age-related changes in brain function and structure that are related to memory for visual pairs of objects and scenes.