The retina is increasingly recognized as a functional extension of the central nervous system, sharing embryological origins, layered neural organization, and vulnerability to many of the same pathological processes that affect the brain. Yet while retinal imaging has become an established tool in ophthalmology, its relevance for neuropsychiatric and neurocognitive disorders is only beginning to be understood. A large new study published in Brain Communications provides one of the most detailed examinations to date of how retinal nerve fibre layer thickness corresponds to cognitive performance and to risk states that precede dementia.
The research draws from the population-based LIFE-Adult Study in Leipzig, which included high-resolution circumpapillary optical coherence tomography (OCT) scans from 1,300 adults between 60 and 79 years of age. After excluding individuals with eye disease and, in a stricter subsample, those with neurological or psychiatric conditions, the investigators analyzed 768 measurement points around the optic nerve in more than 2,000 eyes. This unusually fine spatial resolution allowed them to assess localized retinal nerve fibre layer thickness (cpRNFLT) with precision rarely achieved in previous dementia-related ocular studies.
Participants completed a comprehensive neuropsychological test battery aligned with the six cognitive domains defined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5): attention, executive function, learning and memory, perceptual-motor abilities, language, and social cognition. They were also evaluated for mild cognitive impairment (MCI) and mild neurocognitive disorder (mild NCD), two risk states that fall below the threshold for major dementia syndromes yet indicate measurable decline. Unlike many smaller or memory-focused studies, this design allowed for a broad and systematic mapping between cognitive performance and retinal structure.
The strongest and most consistent association involved the domain of attention. Across thousands of measurement points, thinner retinal nerve fibres - especially in temporal, temporal-superior, and nasal-superior regions - were linked to worse attentional performance. This relationship persisted even after adjusting for age, sex, scanning radius, and education, and it replicated in the subsample where participants with brain-relevant conditions were removed. A thicker nerve fibre layer, in turn, was associated with better attention, suggesting that cpRNFLT differences may reflect underlying neurobiological vulnerabilities affecting attentional processing circuits.
Interestingly, the pattern was driven almost entirely by male participants. In men, up to a quarter of measured retinal locations showed significant associations between cpRNFLT and attention. Women did not display similar correlations, a sex-specific divergence that the authors note aligns with known differences in the prevalence and presentation of various neurocognitive conditions. Executive function showed a weaker and less consistent relationship, detectable mainly in a subset of females and localized to nasal regions. The remaining cognitive domains - including memory, language, perceptual-motor skills, and social cognition - did not show significant linear associations with cpRNFLT.
The authors then examined whether retinal measurements predicted the presence of MCI or mild NCD. Here, too, the temporal and temporal-superior regions emerged as key. Individuals with MCI had thinner cpRNFLT in these segments, while the same pattern appeared in participants with mild NCD. When the analysis narrowed specifically to amnestic MCI and amnestic mild NCD - subtypes associated most closely with Alzheimer's disease - the associations strengthened further. Thinning in temporal regions of the retina corresponded clearly to these risk states, suggesting that the retina's temporal fibres may be especially sensitive to early neurodegenerative change.
To contextualize these retinal findings, the researchers examined structural brain imaging data from the same participants. Consistent with a predominantly Alzheimer-type etiology, individuals with MCI and mild NCD showed significantly smaller hippocampal volumes, whereas white matter lesion burden - more characteristic of small vessel disease - was relatively less divergent from cognitively healthy participants. The alignment between hippocampal atrophy and retinal thinning reinforces the possibility that cpRNFLT reflects broader neurodegenerative processes rather than isolated ocular changes.
Receiver operating characteristic analyses further explored whether retinal metrics could meaningfully classify risk states. While cpRNFLT alone did not achieve levels sufficient for diagnosis, certain temporal segments achieved area-under-the-curve values up to 0.81 in predicting amnestic MCI or amnestic mild NCD in specific subgroups. This suggests that, although retinal imaging should not be interpreted in isolation, it may contribute to multimodal early-screening strategies combining cognitive assessment, brain imaging, and biomarker evaluation.
The psychiatric frameworks used to define cognitive impairment shape how these findings are interpreted. Diagnostic categories like MCI and mild NCD describe patterns of functional deviation but do not pinpoint underlying mechanisms. The present study used DSM-5 domains and risk states as organizing labels, yet the biological signal detected in the retina - especially in attentional circuits - may reflect early stress on distributed neural networks rather than discrete psychiatric entities. The retina, in this sense, bypasses some of the ambiguities that arise when psychological constructs are used to infer pathology. It provides a direct structural measure of neuroaxonal integrity that does not rely on subjective reporting, symptom clustering, or culturally influenced cognitive expectations.
Through the lens of Seven Reflections' Dimensional Systems Architecture (DSA), the retina's role becomes even clearer. DSA treats cognition as a distributed field interaction rather than a set of isolated traits, and the sensory-cortical interface plays a central role in determining system coherence. The retina, as an exposed extension of the brain, offers a structural marker of this coherence. Thinning in temporal retinal fibres reflects weakening in the broader attentional-integration field - an early sign that the system's capacity to sustain stable information flow is diminishing. The eye becomes not merely a diagnostic surface, but an accessible entry point into the architecture of cognitive decline.
As interest in non-invasive biomarkers grows, the present findings reinforce the idea that subtle neurodegenerative changes may be detectable years before clinically recognizable dementia emerges. Deep ophthalmic imaging does not replace cognitive assessment or neuroimaging, but it offers a rapid, low-burden method for capturing early structural signatures of decline. Future longitudinal work - and integration with machine learning approaches - may clarify whether retinal nerve fibre trajectories can meaningfully predict who will progress from mild impairments to more severe neurocognitive disorders.
