TL;DR
This study used multimodal imaging (MRI and OCTA) to investigate the intricate connections between cognitive function, retinal health, and brain structure/function across the Alzheimer's Disease (AD) spectrum. They found that while brain structure changes are linked to retinal biomarkers even in early stages (SCD), brain function alterations become more prominent in later stages (MCI/AD), suggesting a progression of AD pathology across these interconnected systems.
What Problem Does This Paper Address?
Alzheimer's Disease (AD) is a devastating neurodegenerative disorder characterized by progressive cognitive decline. While AD is well-studied, the early stages, known as the AD spectrum (ADS) disorders, including Subjective Cognitive Decline (SCD) and Mild Cognitive Impairment (MCI), are less understood. Specifically, the relationships between cognitive function, changes in the eye (retina), and alterations in the brain (structure and function) remain unclear. The retina, being an extension of the central nervous system, offers a unique window into brain health. However, previous research has often examined these connections in isolation (e.g., brain-retina or retina-cognition). This paper aims to bridge this gap by simultaneously investigating the interplay between cognition, retinal biometrics, and brain alterations across the entire AD spectrum using a multimodal imaging approach.
Methods
The researchers employed a two-step multivariable linear regression model to analyze the complex relationships.
1. Data Acquisition: * Participants: Four groups were studied: Healthy Controls (HC, n=16), Subjective Cognitive Decline (SCD, n=35), Mild Cognitive Impairment (MCI, n=18), and Alzheimer's Disease (AD, n=7). * Brain Imaging: * Structural MRI: 3D T1-weighted BRAvo sequence on a 3-T scanner to measure cortical thickness using the Desikan-Killiany-Tourville (DKT) atlas. * Functional MRI (fMRI): Resting-state fMRI on a 3-T scanner to assess brain function, specifically measuring the Fractional Amplitude of Low-Frequency Fluctuations (FALFF) and Regional Homogeneity (ReHo) using voxel-based analysis. * Eye Imaging: * Optical Coherence Tomography Angiography (OCTA): Used a deep learning model (FARGO) to segment the foveal avascular zone (FAZ) and retinal vessels. Key retinal biometrics extracted included FAZ area and perimeter, retinal blood vessel curvature (RBVC), and thicknesses of the retinal nerve fiber layer (RNFL) and ganglion cell layer-inner plexiform layer (GCL-IPL). * Cognitive Assessment: A comprehensive battery of neuropsychological tests was administered, including MMSE, MOCA, BNT, CDT, BDST, and TMT-A/B.
2. Data Analysis: * Preprocessing: Standard MRI preprocessing pipelines (FreeSurfer, DPARSF, SPM12) were used for structural and functional MRI data. OCTA images were processed for retinal biometrics. * Statistical Modeling: A two-step regression approach was used. The first step identified associations between brain metrics and cognitive function. The second step then examined the relationship between these brain metrics (or their associations with cognition) and retinal biometrics. This was done to isolate specific cognition-eye-brain connections. * Covariates: Analyses controlled for age, sex, education, hypertension, and diabetes. For structural analyses, estimated total intracranial volume (ETIV) was also controlled. * Statistical Significance: FWE correction was used for fMRI data, and Bonferroni correction for structural analyses.
Datasets
Not explicitly detailed as separate datasets. The study utilized data collected from 76 participants across four groups (HC, SCD, MCI, AD). The data comprised neuroimaging (MRI, OCTA) and neuropsychological assessment results.
Key Insights and Ideas
- Retinal Changes Precede Brain Function Deficits in Early AD: While brain function (FALFF, ReHo) showed correlations with retinal biometrics in MCI and AD, these connections were absent in the SCD group. This suggests that retinal changes might be an earlier indicator of AD pathology than detectable functional brain alterations.
- Widespread Cortical Thickening in SCD: Interestingly, individuals in the SCD group exhibited widespread cortical thickening that correlated with various retinal biomarkers. This could represent an early compensatory mechanism in the brain as it begins to be affected by AD pathology.
- FAZ as a Key Indicator: Enlarged FAZ areas and perimeters were consistently observed across the AD spectrum, aligning with previous findings. This suggests that damage to retinal blood vessels, potentially due to amyloid-beta deposition, plays a significant role.
- Interconnectedness Across the Spectrum: The study highlights that while the specific manifestations differ, there's a continuous thread of connection between the eye, brain, and cognition throughout the AD spectrum.
Key Results
- MCI/AD: Reduced FALFF in temporal regions (left STG) in MCI, and reduced ReHo in occipital regions (right MOG) in AD, were linked to increased FAZ area and decreased RNFL thickness.
- SCD: Widespread cortical thickening was observed, correlating with increased FAZ area/perimeter and RBVC, and decreased RNFL thinning. Notably, thickening in the right fusiform gyrus (FG) and right parahippocampal gyrus (PHG) correlated with reduced GCL-IPL thickness.
- Cognition-Eye-Brain Correlations:
- In MCI, cortical thickness of the right ERC was negatively correlated with FAZ area, and left STG thickness was positively correlated with FAZ perimeter. Left ERC and right SPG thickness were negatively correlated with RNFL thickness.
- In AD, cortical thickness of the left and right cingulate isthmus (CI) and right PCUN were negatively correlated with FAZ area.
- In SCD, cortical thickness of the left SPG, right PCUN, and right RMF were positively correlated with FAZ area and perimeter. Right FG and right PHG thickness were negatively correlated with GCL-IPL thickness.
Limitations and Future Work
- Cross-sectional Design & Small Sample Size: The study's cross-sectional nature limits causal inferences. The small sample size, particularly for the AD group, is a significant limitation, partly due to strict exclusion criteria and the need for ocular measurement compatibility.
- Lack of Biomarkers: The absence of amyloid-beta (Aβ) or tau biomarkers prevented direct analysis of their relationship with ocular measurements, making it difficult to confirm if observed changes are specific to AD.
- Brain Segmentation Resolution: The use of the DKT atlas for brain segmentation might have masked finer-grained structural and functional heterogeneity within subregions.
Future Directions: * Longitudinal studies with larger sample sizes are needed to establish causal relationships. * Incorporating Aβ and tau biomarkers would provide a more complete picture. * Future research could benefit from subregion-level analysis and advanced imaging techniques to explore these complex interactions in greater detail.
Summary updated