Identifying EEG-Based Neuroinflammation Biomarkers in Dyslexia Using ANN Models

Background& Objectives: In the development of chronic neurological disorders, recent studies have highlighted the interplay among immune dysregulation, vitamin D deficiency, and epigenetically mediated mechanisms, specifically DNA methylation. The potential of home-based electroencephalography (EEG) as a non-invasive technique for identifying neuroinflammatory markers in immunocompromised populations is examined in this work. Our goal is to identify EEG-based biomarkers that may reflect neurophysiological patterns associated with dyslexia and potential immune-related processes, though causal relationships remain to be established. Methods: A home-based EEG system was used to gather resting-state EEG data from two groups: age-matched neurotypical controls and children with developmental dyslexia diagnoses. The EEG recordings were analysed using sophisticated machine learning algorithms, with a focus on locating topographic and spectral characteristics linked to neuroinflammatory activity. Results: Using EEG characteristics suggestive of possible neuroinflammation, machine learning models were able to differentiate dyslexic children from controls with high classification accuracy. The hypothesis of immune-related neurophysiological changes was supported by the dyslexic group's altered activity in particular frequency bands, especially in frontal and temporal regions. Conclusions: This study offers early evidence in favour of using home-based EEG as a useful and approachable neurodiagnostic method for identifying abnormalities in brain function linked to the immune system. Early detection of at-risk individuals and prompt, individualised interventions to improve brain health outcomes may be made possible by the combination of EEG and machine learning.