Cognitive Organization Breaks Down in Aging and Alzheimer’s Disease

Apochi Obed Okwoli¹², Nikolai Axmacher¹²
¹ International Graduate School of Neuroscience, Ruhr University Bochum, Bochum, Germany
² Institute of Cognitive Neuroscience, Department of Neuropsychology, Ruhr University Bochum, Bochum, Germany

Key takeaway: A topology-based index reveals entorhinal-anchored cognitive disorganization that predicts clinical progression.

---

Background & Motivation

Introduction

The entorhinal cortex is affected early in Alzheimer’s disease, but clinical cognition is usually measured as isolated test performance. We asked whether disease progression also reflects disruption in how cognitive domains are organized.

---

Standard Scores Miss Cognitive Organization

Task–task dissimilarity matrices show that cognitively normal participants have organized task relationships, whereas high-ETI participants with Alzheimer’s disease show disrupted cross-domain structure.

---

Organization ≠ Impairment — A Distinct Dimension

Conventional models quantify deviation in raw scores.
ETI quantifies deviation in cross-domain task relationships.

ETI and dcor-based deviation showed partial but non-redundant overlap:

Pearson r = 0.54; Spearman ρ = 0.23

---

Methods & Sample

Large-Scale NACC Cohort — ADRC Multi-Site

Structural MRI from FreeSurfer-processed T1-weighted scans linked within ±365 days of ETI baseline visit.

Total participants with ETI computed: 105,904

Diagnostic groups:

• Cognitively normal: 61,006

• Subjective cognitive impairment: 5,410

• Mild cognitive impairment: 19,432

• Dementia: 20,056

ETI stratified 3-year conversion risk and improved prognostic performance beyond Trail Making Test Part B alone.

---

Results

Entorhinal-Anchored Disorganization

ETI associations localized most strongly to bilateral entorhinal cortex, with additional posterior-medial and lateral temporal effects.

|β| from region-wise linear models.
n = 705.
FDR q < 0.05 across 198 regions.
CT = cortical thickness.
Models adjusted for age, sex, education, diagnosis, and intracranial volume.

---

Functional Outcomes — EC + PMN, Not Hippocampus

EC and PMN retained independent associations across outcomes. Hippocampal volume did not improve joint models.

All models adjusted for age, sex, education, diagnosis, and APOE ε4.
***p < 0.001, **p < 0.01.
Ordinal logistic regression was used for the FAQ travel item.

---

Take-Home

1. Cognitive decline reflects breakdown in cross-domain organization, not only loss of performance.

2. Disorganization is preferentially anchored to entorhinal cortex integrity, with additional posterior-medial network contributions.

3. ETI independently predicts clinical progression beyond conventional neuropsychological measures.

---

Key References

1. Besser et al. (2018). NACC UDS v3 development. Alzheimer Disease & Associated Disorders.

2. Weintraub et al. (2018). UDS v3 neuropsychological battery. Alzheimer Disease & Associated Disorders.

3. Braak & Braak (1991). Neuropathological staging of Alzheimer’s disease. Acta Neuropathologica.

4. Ranganath & Ritchey (2012). Two cortical systems for memory-guided behaviour. Nature Reviews Neuroscience.

---

Acknowledgements

Data were obtained from the National Alzheimer’s Coordinating Center, funded by the National Institute on Aging/National Institutes of Health Grant U24 AG072122. NACC data were contributed by NIA-funded Alzheimer’s Disease Research Centers across the United States.

This work was supported by DAAD and Ruhr University Bochum. The authors thank the ADRC participants and staff.