Entorhinal‐based path integration selectively predicts midlife risk of Alzheimer's disease

Coco Newton, Marianna Pope, Catarina Rua, Richard N. Henson, Zilong Ji, Neil Burgess, Christopher T. Rodgers, Matthias Stangl, Maria‐Eleni Dounavi, Andrea Castegnaro, Ivan Koychev, Paresh Malhotra, Thomas Wolbers, Karen Ritchie, Craig Ritchie, John T. O’Brien, Li Su, Dennis ChanAlzheimer 2024 · DOI

TL;DR

This study reveals that a specific type of spatial navigation, called path integration, is impaired in midlife individuals at risk for Alzheimer's disease (AD), even before other cognitive functions decline. This impairment, linked to the entorhinal cortex, could serve as an early behavioral marker for AD onset.

What Problem Does This Paper Address?

Alzheimer's disease (AD) is a devastating neurodegenerative disorder, and a major hurdle in developing effective treatments is the difficulty in identifying the disease at its earliest stages when interventions might be most impactful. Current AD progression models suggest that significant cognitive decline only occurs after substantial brain pathology has developed. However, subtle cognitive deficits may precede these overt symptoms. The entorhinal cortex (EC), a brain region crucial for memory and spatial navigation, is one of the first areas affected by AD pathology. Specifically, EC grid cells are involved in path integration (PI), a fundamental spatial behavior. This paper investigates whether PI impairment is the very first behavioral sign of AD risk, even before other cognitive domains typically associated with AD, like episodic memory, are affected.

Methods

The researchers employed a multi-faceted approach combining behavioral tasks, advanced neuroimaging, and detailed risk factor assessment.

  • Virtual Reality (VR) Path Integration Task: Participants navigated a virtual environment using a head-mounted display. The core task involved mentally reconstructing a triangular path. They walked to three numbered cones and then had to return to the starting cone without visual cues. The key manipulation was altering the visual environment during the return path (e.g., removing surface textures or distal landmarks) to isolate the reliance on internal spatial representations.
  • Cognitive Comparator Tasks: To determine the specificity of PI deficits, participants also underwent a battery of standard neuropsychological tests assessing allocentric spatial memory, egocentric spatial orientation, episodic memory (visual short-term binding, name-face association, narrative recall), and global cognition.
  • 7 Tesla (7T) MRI: A subset of participants underwent ultra-high field MRI to examine structural (volume of brain regions like the entorhinal cortex) and functional (fMRI activity related to grid cell-like representations) correlates of PI performance.
  • AD Risk Factor Stratification: Participants were carefully selected and stratified based on established AD risk factors: family history of dementia (FH), APOE ε4 allele status, and the Cardiovascular Risk Factors, Aging and Dementia Study (CAIDE) risk score (a measure of physiological risk factors).

Datasets

The study utilized data from the PREVENT Dementia prospective cohort study, which recruited cognitively healthy individuals aged 40-59. This cohort provided the participants for the behavioral and neuroimaging assessments. No new benchmark datasets were introduced; rather, existing data from this well-established cohort were leveraged.

Key Insights and Ideas

  • PI as a Sensitive Early Marker: The most striking finding is that PI impairment, specifically in the absence of external orientation cues, appears to be a highly sensitive indicator of AD risk, even in midlife individuals who are otherwise cognitively healthy.
  • Entorhinal Cortex Specificity: The deficit is linked to the entorhinal cortex, the brain region known to be affected early in AD. This suggests that PI impairment is not a general cognitive decline but a specific manifestation of early AD-related neuropathology.
  • Timing is Crucial: The study strongly suggests that PI deficits precede impairments in other cognitive domains, including episodic memory, which has historically been considered an early AD marker. This shifts the potential timeline for early AD detection.
  • Sex Differences in Risk: The study uncovered sex-specific effects, with males showing a more pronounced PI impairment in the presence of hereditary risk factors (FH and APOE ε4).
  • Functional MRI Correlates: fMRI data revealed that altered grid-like activity in the posterior-medial EC was associated with PI impairments, further solidifying the link between EC function and spatial navigation deficits in AD risk.

Key Results

  • PI Impairment in At-Risk Individuals: Individuals at risk for AD (due to FH, APOE ε4, or high CAIDE score) showed significantly worse performance on the PI task, particularly when orientation cues were removed.
  • No Impairment in Other Domains: Crucially, these same at-risk individuals did not show significant impairments on comparator tasks assessing episodic memory, allocentric spatial memory, or egocentric spatial orientation.
  • Predictive Power of PI: When compared to other cognitive tests, PI performance was the only significant predictor of "double-risk" status (combined hereditary risk factors).
  • Structural MRI Findings: Structural MRI did not reveal significant differences in the volume of key brain regions (including the entorhinal cortex) between high- and low-risk participants, suggesting that volumetric changes may not be the earliest detectable correlate of PI impairment.
  • fMRI Findings: Negative hexadirectional grid-like fMRI signals in the posterior-medial EC were associated with PI impairment. This signal was modulated by risk factors and sex, with a stronger association observed in males with hereditary risk and individuals with higher CAIDE scores.

Limitations and Future Work

  • Exclusion of "Out-of-Bounds" Trials: A significant proportion of trials were excluded due to participants moving outside the designated virtual test space. While this was controlled for in analyses, it highlights a limitation of using VR in certain settings and suggests that this behavior itself might be indicative of impaired wayfinding.
  • Sample Size for Interactions: The sample size, while adequate for many analyses, was relatively small for exploring complex interactions of multiple risk factors and sex, warranting replication in larger cohorts.
  • Time-Locking of Assessments: The study acknowledges potential variability in the timing of cognitive assessments and PI tasks, which was addressed by including visit date as a covariate.
  • Future Directions: The authors propose follow-on studies to investigate PI in presymptomatic familial AD mutation carriers and to explore the relationship between tau spread, EC neuronal activity, and PI in AD mouse models. They also emphasize the need for multimodal imaging (including PET for amyloid and tau) to clarify the relative contributions of different pathologies to PI impairments. The potential of PI as an early detection tool and a target for future therapeutic interventions is a key takeaway.

Summary updated

Abstract

Altered PI may represent the transition point from at-risk state to disease manifestation in AD, prior to impairment in other cognitive domains.
BibTeX
@inproceedings{newton2024entorhinalbased,
  author = {Coco Newton and Marianna Pope and Catarina Rua and Richard N. Henson and Zilong Ji and Neil Burgess and Christopher T. Rodgers and Matthias Stangl and Maria‐Eleni Dounavi and Andrea Castegnaro and Ivan Koychev and Paresh Malhotra and Thomas Wolbers and Karen Ritchie and Craig Ritchie and John T. O’Brien and Li Su and Dennis Chan},
  title = {{Entorhinal‐based path integration selectively predicts midlife risk of Alzheimer's disease}},
  booktitle = {Alzheimer},
  year = {2024},
  doi = {10.1002/alz.13733},
}
Figure