Authors :
Presenting Author: Roberto Santalucia, MD – Institute of Neurosciences, UCLouvain, Brussels, Belgium
Arnaud Gérain, MD – Institute of Neurosciences, UCLouvain, Brussels, Belgium
Emiel Vereycken, M.Sc. – Clouds of Care NV, Ghent, Belgium
Pieter van Mierlo, PhD – Clouds of Care, NV, Ghent, Belgium
Riëm El Tahry, MD, PhD – Institute of Neurosciences, UCLouvain, Brussels, Belgium
Rationale:
EEG functional connectivity reflects statistical dependencies between cortical regions and offers insights into brain network organization. Graph theory applied to EEG data enables visualization of how these networks evolve. However, normative developmental patterns of both functional connectivity and power spectral features remain underexplored, limiting our ability to detect pathological deviations.
Methods:
We retrospectively analyzed de-identified long-term EEGs from 144 children, aged 0–16 years, referred to Saint-Luc University Hospital for paroxysmal events. All participants had normal EEG and a final diagnosis of non-epileptic events or other benign conditions, such as migraine or vasovagal syncope. We computed several connectivity metrics : weighted Phase Lag Index (wPLI), corrected imaginary part of Phase Locking Value (cIPLV), and orthogonalized Amplitude Envelope Correlation (ortho-AEC) and derived network measures (mean degree and efficiency).
Results:
Absolute power decreased with age, while relative power increased in high frequencies and decreased in low frequencies, indicating a developmental shift toward high-frequency activity. Aperiodic components (offset, exponent) also declined. Although connectivity results varied, they suggested age-related changes in network organization, with evolving balance between integration and segregation, consistent with a "small-world" topology.
Conclusions:
These findings highlight developmental trends in EEG power and connectivity. A larger longitudinal study is needed to establish normative references for clinical application in pediatric neurology.
Funding: None