Resting-state EEG offers a powerful tool for studying the aging brain and can help identify neurodegenerative changes. The resting-state signal consists of periodic components (alpha-band power) and aperiodic activity (1/f slope). Decomposing these two elements provides features that can act as markers of cortical health. Prior research has shown that healthy older adults typically have a flatter aperiodic slope (a sign of increased neural noise) along with slowing of individual alpha frequencies. Parkinson's Disease patients show similar spectral alterations, although the specific patterns may differ between groups. Moving these laboratory findings into clinical practice requires computational resources capable of processing large datasets with complete reproducibility. We applied our analysis pipeline to the ds002778 OpenNeuro dataset, which contains 93 EEG sessions from healthy older adults and Parkinson's patients recorded across multiple visits. By using ARCHER2, we developed a standardised workflow that extracts physiologically meaningful biomarkers through Welch Power Spectral Density estimation and SpecParam (FOOOF) parameterisation. Deploying on ARCHER2 established a reproducible, scalable framework ready to extend to large-scale population studies, where parallel processing of thousands of subjects becomes essential.
Amirfarhang Miresmaeili (Thu,) studied this question.