The intricate relationship between sleep quality and cognitive function may depend more on what happens inside the brain during sleep than on circulating inflammatory or metabolic signals in blood. This finding challenges the growing emphasis on blood biomarkers as primary predictors of cognitive decline in aging populations. Analysis of 1,898 community-dwelling men aged 65 and older revealed that detailed sleep brain wave patterns—including sleep spindles and other electroencephalographic microstructures—showed stronger correlations with cognitive test scores than nine different blood markers of inflammation and metabolism. The research team measured 41 distinct sleep EEG features alongside blood levels of markers like leptin, creating predictive models for cognitive performance on the Modified Mini-Mental State Examination. Sleep microstructures achieved a correlation of 0.45 with actual cognitive scores, marginally outperforming models using demographic factors alone. Interestingly, leptin was the only blood marker showing significant association with cognition, suggesting that metabolic signaling may matter more than inflammatory pathways for brain function in older men. This research adds nuance to the emerging field of sleep medicine and cognitive aging. While blood-based biomarkers offer convenience for clinical assessment, the superior predictive power of sleep brain waves suggests that direct neural measurements capture cognitive health more accurately. The modest improvement over demographic predictors alone indicates that neither approach provides a complete picture. For longevity-focused adults, this reinforces that sleep architecture—the detailed structure of sleep stages and brain oscillations—deserves as much attention as metabolic health markers when optimizing cognitive preservation strategies.