Relationship between hippocampal structure and memory function in elderly humans.

With progressing age, the ability to recollect personal events declines, whereas familiarity-based memory remains relatively intact. It has been hypothesized that age-related hippocampal atrophy may contribute to this pattern because of its critical role for recollection in younger humans and after acute injury. Here, we show that hippocampal volume loss in healthy older persons correlates with gray matter loss (estimated with voxel-based morphometry) of the entire limbic system and shows no correlation with an electrophysiological (event-related potential [ERP]) index of recollection. Instead, it covaries with more substantial and less specific electrophysiological changes of stimulus processing. Age-related changes in another complementary structural measure, hippocampal diffusion, on the other hand, seemed to be more regionally selective and showed the expected correlation with the ERP index of recollection. Thus, hippocampal atrophy in older persons accompanies limbic atrophy, and its functional impact on memory is more fundamental than merely affecting recollection.

Four types of novelty-familiarity responses in associative recognition memory of humans.

Animal studies show that, like inferior temporal neurons, dorsolateral prefrontal and parietal neurons often respond more strongly to individual novel than to individual familiar stimuli. It is currently unclear whether the novelty preference of prefrontal and parietal neurons extends to associative memory. We used electromagnetic recordings (MEG/EEG) and functional magnetic resonance imaging in two groups of healthy young adults to identify neural populations outside the inferior temporal cortex that exhibit associative novelty (stronger responses for new than for old configurations of two familiar items), and to distinguish them from associative familiarity (stronger responses for old than for new configurations of two familiar items). Subjects were required to learn and were later tested for associations based on the spatial configurations of two stimuli (a face and a tool). At test, learned (old) and rearranged (new) spatial stimulus configurations had to be discriminated. This recognition memory test could only be solved through the associative relationship between individual items because all component items of the stimulus configurations were equally familiar. In both imaging modalities, right dorsolateral prefrontal cortex and right parietal cortex showed an associative novelty response, whereas the right superior temporal cortex showed an associative familiarity response. With EEG/MEG only, the right extrastriate cortex showed an early associative familiarity and a late associative novelty response, whereas the opposite pattern emerged in bilateral frontopolar cortex. Thus, through a multimodal approach, it was possible to identify four types of associative novelty/familiarity responses outside the inferior temporal cortex.

Measures of hippocampal volumes, diffusion and 1H MRS metabolic abnormalities in temporal lobe epilepsy provide partially complementary information.

We assessed whether interictal measures of hippocampal volume, hippocampal diffusion and metabolic abnormalities yield correlated or complementary information about hippocampal pathology in patients with temporal lobe epilepsy (TLE). Volumes, apparent diffusion coefficients (ADC) and ratios of N-acetyl-aspartate (NAA) to Creatine/Phosphocreatine (Cr) and Choline (Cho) were measured from each hippocampus during one magnetic resonance imaging (MRI) session in patients with TLE. Structural MRI showed unilateral hippocampal sclerosis (HS) in 13 patients and was normal in the remaining nine patients. Pearson's correlation (two-tailed) between ADC values and NAA/(Cr + Cho) ratios was significant (P = 0.04, r = -0.45) for the hippocampus ipsilateral to the epileptogenic zone as determined on the basis of interictal and ictal scalp EEG recordings. This finding was driven by a very high correlation between the two measures in the presence of HS (P < 0.001, r = -0.96). Furthermore, ipsilateral ADC values but not NAA/(Cr + Cho) ratios were correlated with disease duration (P = 0.001, r = 0.67). Hippocampal volumes did not correlate with either ADC values, NAA/(Cr + Cho) ratios or disease duration. These data suggest that hippocampal volumes, NAA/(Cr + Cho) ratios and ADC values capture partially complementary aspects of hippocampal pathology.