Back

Evaluating the position of the uncal apex as a predictor of episodic memory across the adult lifespan

Nordin, K.; Bahrd, P.; Andersson, M.; Salami, A.

2024-03-20 neuroscience
10.1101/2024.03.18.585629 bioRxiv
Show abstract

Structural decline of the hippocampus follows heterogeneous spatial patterns, and is an important determinant of episodic memory dysfunction in aging. However, evidence indicate that the anatomical landmark uncal apex, used to demarcate anterior and posterior hippocampal subregions, changes position as the hippocampus atrophies. This risks misclassifying gray matter into the incorrect subregion when using standard demarcation methods, potentially contributing to over- and underestimation of age-related anterior and posterior hippocampal volume loss, respectively. Yet, it remains unexplored whether inter-individual differences in uncal apex position itself predict episodic memory performance. Here, we manually identified the uncal apex in anatomical MRI data from a healthy adult-lifespan sample (n=180; 20-79 years), analyzed age differences in its position, and examined associations with word recall performance. Increasing age was linked to a more anteriorly located uncal apex (retracting [~]0.04 mm/year), and this anterior shift was linked to lower memory performance. Using the manually identified uncal apex for subregion segmentation in native space revealed a lower proportion of anterior hippocampal volume compared to segmentation in MNI space (y = -21), with effects of age on volumes differing between methods. Memory performance was not predicted by hippocampal subregion volumes, but the ratio of posterior to anterior volume was significantly linked to memory when accounting for tissue misclassification due to uncal apex variation. These results indicate that uncal apex position may provide an estimate of hippocampal integrity sensitive to inter-individual differences in memory, independent of limitations associated with different segmentation methods.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.