Characterization of aperiodic and theta activity in preterm infants using EEG: Insights into cerebral maturation and inter-individual variability

Preterm birth interferes with the maturation of brain networks and functional activity. Theta oscillations are thought to play a key role in early networks formation, but their maturation appears vulnerable to prematurity. Traditional EEG spectral analyses have indicated marked development of theta power in early infancy, but these approaches mix oscillatory and non-oscillatory activity, limiting insights into the mechanisms underlying the neural changes. Using spectral parametrisation, we aimed to evaluate developmental changes in aperiodic activity and periodic theta power in infants born very preterm compared to full-terms, and to further explore whether clinical factors and brain microstructure could explain the inter-individual variability within preterms. High-density EEG was acquired during active/REM sleep at term-equivalent age (TEA) and 2 months corrected age (2mCA) in 41 very preterm infants and 13 full-term controls. Spectral parameterization was used to extract metrics of aperiodic activity (offset, exponent) and periodic theta power, globally and across spatial clusters (anterior, central, posterior). From TEA to 2mCA, offset, exponent, and theta power, increased with no differences between preterm and full-term infants. At TEA, all metrics were stronger in anterior compared with posterior areas, but this regional landscape shifted by 2mCA for aperiodic activity, driven by pronounced increases in aperiodic offset and exponent in posterior areas from TEA to 2mCA. Within preterms, inter-individual variability in aperiodic and periodic metrics at TEA was partly explained by clinical risk factors and variations in brain microstructure, as assessed with diffusion MRI at the same age. Male sex, lower gestational-age at birth, small weight at birth, and invasive ventilation were linked to alterations of aperiodic activity and theta power. Additionally, higher theta power correlated with lower cortical fractional anisotropy, consistent with more advanced cortical maturation. Collectively, these findings indicate that EEG spectral parameterization combined with spatial analysis provides a sensitive framework for characterising the postnatal maturation of brain activity, and early vulnerabilities associated with prematurity and perinatal adversity.