Multimodal EEG-fNIRS complexity analysis can reveal the neurovascular coupling change with aging

Objective This study aimed to investigate age-related changes in neurovascular coupling (NVC) function of the brain using a multimodal electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) approach combined with nonlinear entropy analysis. Methods The study included 20 young adults and 20 older adults. We simultaneously recorded EEG and fNIRS data during four cognitive tasks (auditory attention task, verbal fluency task, upper limb motor task, and recognition memory task). We calculated sample entropy (SampEn) and conditional entropy (CondEn) for five EEG frequency sub-bands (Delta: 1–4 Hz, Theta: 4–8 Hz, Alpha: 8–12 Hz, Beta: 12–30 Hz, Gamma: 30–40 Hz) as well as for fNIRS signals (oxygenated hemoglobin and deoxygenated hemoglobin, HbO and HbR) to quantify neural and hemodynamic complexity of brain activity. Additionally, cross-sample entropy (Cross-SampEn) and cross-conditional entropy (Cross-CondEn) between EEG and fNIRS signals were computed to evaluate NVC function. Group differences were assessed using analysis of covariance (ANCOVA), with age group as the fixed factor and sex and years of education as covariates. Results Older adults showed higher EEG SampEn or CondEn in selected frequency bands during the auditory attention task and the encoding and recognition phases of the recognition memory task. For fNIRS, older adults showed lower HbO SampEn during memory recognition. Cross-modal analysis showed lower beta-HbR Cross-SampEn during the left upper limb motor task and higher theta-HbR Cross-SampEn during memory encoding in older adults. Conclusion Aging is associated with task-specific alterations in neural complexity, hemodynamic complexity, and the cross-modal temporal relationship between EEG and fNIRS signals. These findings suggest that age-related changes in NVC-related dynamics are better characterized as dynamic reorganization rather than a uniform enhancement or impairment, and support the potential of multimodal entropy analysis for characterizing age-related neurovascular coupling change.