Our laboratory is involved in the application of rodent models of neurocognitive aging to healthy older adults and those with amnestic mild cognitive impairment (aMCI), a subset of cognitively impaired individuals that are likely to progress to Alzheimer’s disease. Rodent work has strongly suggested that one of the hallmark features of neurocognitive aging is a partial knockdown of the dentate gyrus that results in a failure to represent similar environments as distinct from old (reduced pattern separation). In the absence of the dentate’s separation signal, the CA3 could be driven by its recurrent collaterals, thus biasing it towards completion. We have tested this hypothesis in aged humans and in cognitively impaired individuals, finding strong support for the model’s predictions. the figure below is a working model of the key changes in hippocampal circuitry with age based on a combination of findings from animals and humans. This work is intimately related to some of the work conducted in the Gallagher laboratory with aged rodents.
Behavioral pattern separation (discrimination) deficits in older adults
In collaboration with the Stark Lab, we have developed a suite of behavioral tasks that assess hippocampal pattern separation across multiple domains. In the objects version (BPS-O), we find a consistent decline in discrimination ability with age (shown in the plot below). Older adults are shown pictures of every day items during study. During test, they are shown some pictures they have seen before, some new pictures and some pictures that were similar but not identical to ones they had previously seen (i.e. lures). These lure items allow us to evaluate the individual’s ability to discriminate among similar experiences in memory, a capacity we believe is dependent on hippocampal pattern separation. The pattern of behavioral decline is consistent with our hypothesized model and suggest that older adults require stimuli to be more dissimilar in order to correctly identify them as items they have not seen before. These results were published earlier this year (Stark, Yassa, et al. Neuropsychologia 2013).
Age-associated alterations in the DG/CA3 network
Using high-resolution fMRI we were able to determine the neural basis of discrimination deficits in older adults. Running a similar task inside the scanner, we found that the dentate and CA3 subregions (DG/CA3) exhibited “representational rigidity”. In other words, the DG/CA3 required input stimuli to be much more dissimilar before it can properly engage in pattern separation (see figure below). In the same investigation, we found that this was related to the integrity of the perforant path and also related to the functional connectivity between the entorhinal cortex and the DG/CA3 (Yassa et al. PNAS 2011). The altered dynamics are hypothesized to stem from reduced inhibition in the CA3 region which additionally leads to hyperactivity in this region (observed previously in Yassa et al. Hippocampus 2010). This work continues in collaboration with the Stark Laboratory at UC Irvine.
Perceptual vs. conceptual interference in older adults
Similar pictorial stimuli invoke interference both in the perceptual and conceptual domains, and do not allow one to be disentangled from another. For example, it is very difficult to design a set of pictorial stimuli that are perceptually similar yet conceptually unrelated. Verbal stimuli, on the other hand, allow us to independently manipulate conceptual and perceptual interference. This is an important question, especially, since perceptual and conceptual processing likely involve divergent neural mechanisms. We tested discrimination on conceptually similar (semantically related) and perceptually similar (phonologically related) verbal stimuli in young and older adults, and found that older adults were selectively impaired in perceptual discrimination and intact in conceptual discrimination. This deficit was not secondary to failure in working memory, attention, or visual processing. Since conceptual discrimination was intact in older adults despite known DG/CA3 dysfunction in this demographic, we take this to suggest that DG/CA3 pattern separation is only required for perceptual and not conceptual pattern separation (Ly, Murray and Yassa, Hippocampus 2013). We have since extended these results to include multiple levels of interference (by varying the number of similar items shown during study) and began conducting neuroimaging studies to investigate hippocampal contributions to verbal discrimination in young and older adults. For more information on this topic, contact Maria Ly.