Russian Scientists Train AI to Model Brain Activity Through Eye Movements
Russian researchers have learned to “see” the brain through eye movements, marking a new step in assessing human cognitive state and neurological health.
When the Eyes Reflect Neural Activity
At the St Petersburg campus of the Higher School of Economics, a team of researchers has made a breakthrough. Using artificial intelligence, they established a stable and predictable link between two processes that appear unrelated at first glance: eye movements and the brain’s electrical activity. The algorithm can reconstruct gaze trajectories from electroencephalogram data and, conversely, infer patterns of brain activity from eye-tracking signals.
The findings, published in the peer-reviewed journal Science Progress, open practical pathways to earlier and more accurate diagnosis of a wide range of neurological conditions, from Alzheimer’s disease to autism spectrum disorders.
Searching for Early Markers
The experiment involved 40 volunteers. Participants were asked to read text while specialised equipment simultaneously recorded two data streams: a standard electroencephalogram captured via a sensor-equipped cap and eye movements tracked using high-precision cameras. The core scientific question was whether machine-learning algorithms could accurately reconstruct one signal after analysing the other.
The answer was yes. The AI demonstrated not only that a relationship exists, but that it can be measured with practical precision. Based on EEG data – particularly alpha rhythms in the occipital region responsible for visual processing – the algorithm predicted reading speed with up to 95% accuracy and estimated text comprehension with 89% accuracy. The reverse task also proved feasible. Eye-tracking data made it possible to model brain activity, while gaze trajectories pinpointed specific words or phrases where readers lingered longest. These pauses directly indicate cognitive effort, surprise, or difficulty in perception.
Why This Matters
The implications for clinical medicine and neuroscience are substantial. Diagnosing many neurodegenerative diseases such as Alzheimer’s and Parkinson’s, as well as psychiatric and neurodevelopmental conditions including autism spectrum disorder and ADHD, remains a lengthy and resource-intensive process involving interviews, behavioural testing, and costly imaging techniques like MRI and PET scans.
The approach proposed by the Russian researchers points to a simpler, non-invasive screening protocol. Atypical eye-movement patterns – saccades and fixations – have long been recognised as early indicators of these conditions. By establishing a clear digital correlation with specific EEG rhythms, it becomes possible to build automated analysis systems. Following a short reading session with eye tracking or even a simplified EEG setup, such systems could provide clinicians with objective, quantitative indicators of potential abnormalities, supporting diagnosis at a preclinical stage.
Benefits for Patients and Families
Early and accurate diagnosis is critical for effective disease management. Initiating therapy, rehabilitation, or corrective interventions for autism at an earlier stage can significantly improve quality of life. The non-invasive nature and relative simplicity of the method – especially eye tracking – also reduce the stress associated with testing, which is particularly important for children and older adults.
Beyond diagnostics, the discovery offers a powerful new tool for studying human cognition in more natural, real-world settings. EEG equipment is bulky and sensitive, making it difficult to deploy in schools, workplaces, or homes. Eye-tracking glasses or panels, by contrast, are far more practical. With established correlations in place, researchers can infer engagement levels, concentration, and emotional responses by analysing eye movements during learning, conversation, or play.
Implications for Russia and Global Medicine
The work conducted at the Higher School of Economics provides a foundation for an entire class of new medical and research technologies.
In Russia, the discovery could underpin import-substituting solutions in high-tech medicine, aligning with national goals of technological sovereignty in critical sectors. Domestic hardware-software systems combining affordable eye-tracking devices with locally developed AI analytics could be deployed across outpatient clinics, neurological centres, and research laboratories, serving both major cities and regional healthcare systems.
At the global level, the study makes a meaningful contribution to international neuroscience and the evolution of digital medicine.
