ИСТИНА

The relationship between age and blood levels of hemoglobin and cholesterol has been extensively studied. Specifically, hematocrit may reflect certain biochemical changes occurring in various health conditions, including both aging and vascular-dependent cognitive impairment. Patterns of electroencephalography (EEG) have also shown age-dependent dynamics. While altered blood markers of hematocrit and cholesterol have been observed in different health conditions, there are few studies focusing on their association with brain changes during aging. To address this gap, we employed regression modeling of EEG features associated with blood index data. METHODS: We utilized resting-state EEG and blood index data from the elderly group in the Max Planck Institute Leipzig Mind-Brain-Body Dataset (Lemon) (N=74, 67.6±4.7 years, range 59–77, 37 female). Preprocessing was performed using the MNE-BIDS-Pipeline. For feature extraction, we employed Source Power Comodulation (SPoC) to identify neural activity patterns associated with blood indices of hematocrit and low-density lipoprotein cholesterol (LDL-C), Alpha Peak Frequency (APF), and Riemannian Decomposition of EEG covariance matrices, followed by linear regression. To evaluate model quality, we conducted 5-fold cross-validation and used the adjusted R2 metric. RESULTS: The adjusted R2 metrics for predicting hematocrit index were 0.58±0.03, 0.71±0.04, and 0.73±0.02 for APF, SPoC, and Riemannian Decomposition, respectively. For predicting blood concentrations of LDL-Cl, the adjusted R2 metrics were 0.63±0.01, 0.66±0.01 and 0.71±0.01 for APF, SPoC and Riemannian Decomposition, respectively. CONCLUSIONS: Hematocrit declined with age, while increased LDL-C is directly related to atherosclerosis. Our results show a relatively strong predictive power of EEG in detecting changes in hematocrit and LDL-C, and suggest that resting-state EEG may display age-dependent brain changes influenced by the physiological state of the body as a whole.