What lies beneath the EEG: aperiodic and complex dynamics as potential biomarkers of brain health and disease

This doctoral thesis investigates the potential of non-traditional EEG metrics, specifically the aperiodic component of the power spectrum and measures of signal complexity, as quantitative indicators of neural integrity, dysfunction, and recovery. By combining meta-analytic and experimental approaches, the work aims to clarify whether these parameters can serve as mechanistically grounded and clinically meaningful biomarkers of brain function. The first study presents a quantitative meta-analysis of 22 Electroencephalographic (EEG) and Magnetoencephalographic (MEG) datasets comparing the aperiodic exponent and offset across healthy and pathological aging. Results revealed a consistent reduction in both parameters in older adults, supporting their interpretation as markers of neural noise and reduced synaptic efficiency. However, no significant differences were observed between neurological patients and controls, suggesting that the aperiodic component captures general cortical alterations rather than disease-specific changes. The second study explores the role of the aperiodic EEG component in post-stroke patients with and without epileptic seizures. While the exponent did not distinguish patients with clinical epilepsy, it reliably differentiated those showing interictal abnormalities or background slowing from both patients without EEG alterations and healthy controls. These findings indicate that the aperiodic component is more sensitive to cortical dysfunction than to seizure propensity per se. The third study integrates aperiodic, periodic, and complexity measures (Lempel–Ziv complexity, LZC; Higuchi fractal dimension) with clinical and cognitive predictors in hierarchical regression models of functional recovery after stroke. EEG features from the lesioned hemisphere, particularly delta power and LZC, explained unique variance in two clinical outcome scales (Barthel and FIM) improvement, highlighting their relevance as indicators of residual neural capacity and adaptive potential. Taken together, the results suggest that aperiodic and complexity measures offer complementary perspectives on post-stroke brain function, reflecting stability and flexibility, respectively. While neither metric alone qualifies as a definitive biomarker, their combination provides a promising multidimensional framework for quantifying neural health and guiding future applications in precision neurorehabilitation.