Structural brain changes identified up to fifteen years before Alzheimer’s symptoms in people with Down syndrome

A study led by the Sant Pau Research Institute (IR Sant Pau) has succeeded in describing, for the first time in detail, the structural evolution of the medial temporal lobe (MTL) regions across the clinical stages of Alzheimer’s disease in people with Down syndrome. The results, published in the journal Brain, reveal that volume and cortical thickness loss in these regions can begin 13 to 15 years before the onset of symptoms, marking a significant advance in early diagnosis and the design of preventive clinical trials.

The study is based on a large cohort of 259 adults with Down syndrome and 138 euploid controls (individuals without Down syndrome), all of whom underwent high-resolution magnetic resonance imaging (MRI), cerebrospinal fluid biomarker analysis, and detailed clinical and neuropsychological characterization. It is the first study to systematically map the trajectory of MTL subregions—including the anterior and posterior hippocampus, entorhinal cortex, parahippocampus, and Brodmann areas 35 and 36—in relation to cognitive decline and biological markers of the disease.

“We chose to focus on the medial temporal lobe because it is one of the first regions affected by the accumulation of tau neurofibrillary tangles, one of the two key pathologies of Alzheimer’s disease,” explains Dr. Alexandre Bejanin, from the Neurobiology of Dementias group at IR Sant Pau and coordinator of the study. “These regions are also involved in functions such as episodic memory and spatial orientation, which are altered early in the disease.”

An Ideal Model for Studying the Onset of Alzheimer’s Disease

People with Down syndrome have a unique genetic predisposition: overexpression of the APP gene due to trisomy of chromosome 21 leads to overproduction of beta-amyloid (Aβ) protein, making them a natural genetic model of Alzheimer’s. It is estimated that over 90% of this population will develop the disease during their lifetime.

This genetic singularity allowed the researchers to use participants’ age as an estimate of the time to symptom onset (Estimated Years to Onset, EYO), providing an objective temporal framework to study brain changes prior to dementia. According to their calculations, the inflection point in structural degeneration occurs between 15 and 9 years before the average age of symptom onset.

“This approach allowed us to observe that the entorhinal cortex and posterior hippocampus are the first regions to show structural loss, even before symptoms appear,” notes Alejandra Morcillo-Nieto, also a researcher in the Neurobiology of Dementias group and first co-author of the study. “Moreover, some regions like the parahippocampus initially show cortical thickening, possibly reflecting inflammatory or compensatory processes, before entering a phase of atrophy—although this is a hypothesis that should be confirmed in future studies.”

High-Precision Neuroimaging and Biomarkers

The work is part of the Down-Alzheimer Barcelona Neuroimaging Initiative (DABNI) and used MRI images acquired at two hospitals in Barcelona: Hospital de Sant Pau and Hospital Clínic. For automated segmentation of MTL subregions, the study used ASHS-T1 software, previously validated in Alzheimer’s populations. This system enables segmentation not only of the anterior and posterior hippocampus, but also key cortical structures such as the entorhinal cortex, Brodmann areas 35 and 36 (including the transentorhinal cortex), and the parahippocampus. In addition, a complementary algorithm (CRASHS) was used to model cortical thickness with greater surface precision.

“We didn’t use novel images per se, but advanced analytical techniques applied to standard images typically obtained in clinical practice,” emphasizes Dr. Alexandre Bejanin. “What we did was apply a specific segmentation software that allows very precise quantification of these medial temporal lobe regions—something that had not been done before in the context of Down syndrome.”

The researchers normalized volumes and thicknesses using scores adjusted for age, sex, and intracranial volume, and harmonized data from different scanners using the ComBat method. Image quality was verified through automatic analysis (CAT12) and visual evaluation by experts.

In parallel, cerebrospinal fluid samples from 243 participants with Down syndrome were analyzed to determine levels of Aβ42/40, phosphorylated tau at position 181 (pTau181), and neurofilament light chain (NfL). The researchers correlated these biomarkers with the structural measures obtained through imaging.

Topographic Progression and Predictive Value

The analysis revealed an orderly anatomical progression of atrophy, consistent with the classical pattern of tau propagation. The entorhinal cortex was the first region to show a cortical thickness inflection point (EYO = -15.7 years), followed by the posterior hippocampus (EYO = -13.5), Brodmann 35 (-13), the anterior hippocampus (-11.5), the parahippocampus (-9.8), and Brodmann 36 (-9).

Posterior hippocampal volume showed particularly strong correlations with all cerebrospinal fluid biomarkers: positive with Aβ42/40 and negative with pTau181 and NfL. Among all structures studied, this region was also the most effective in distinguishing asymptomatic individuals from those with clinical symptoms, with an accuracy of 86.3%. When included in a multivariate model with age, pTau181, and parahippocampal thickness, diagnostic accuracy rose to 96.4%.

“Our data indicate that structural MRI not only enables early detection of changes, but can effectively complement fluid biomarkers in predicting clinical status,” says Benjamin Buehner, first co-author of the study. “This is especially relevant in Down syndrome, where cognitive testing may be limited by baseline intellectual disability.”

Implications for Research and Clinical Trials

This study provides, for the first time, a detailed map of how the medial temporal lobe is affected by Alzheimer’s disease in Down syndrome and places the onset of structural changes more than a decade before clinical symptoms appear. This information is critical for designing future therapeutic trials aimed at intervening during preclinical stages to prevent neurodegeneration.

The authors emphasize that while the cross-sectional design limits the establishment of a causal temporal sequence, the genetic homogeneity of Down syndrome and the reliability of the EYO model provide exceptional robustness to the findings.

The study involved collaboration from more than twenty national and international centers and institutions, including Hospital Clínic de Barcelona, the University of Barcelona, the Center for Networked Biomedical Research on Neurodegenerative Diseases (CIBERNED), the University of Pennsylvania, and the French INSERM.