Abstract The aim of this study was to test the hypothesis that the association between sleep duration and brain activation as assessed by regional cerebral oxygenation using near-infrared spectroscopy (NIRS) is dependent on chronotype. Sleep was tracked across two weeks by actigraphy in 22 adults instructed to keep their normal sleep behavior. Chronotype was assessed by the midpoint of sleep on free days corrected for sleep debt on workdays (MSFsc). Prefrontal cerebral oxygenation (ΔHbDiff) during a visuospatial working memory task was measured in the morning after a night of normal sleep and after one night of extended sleep. Sleep extension was included to experimentally test the robustness of the association between sleep duration and ΔHbDiff. Habitual sleep duration (r = 0.43, p = 0.04) and MSFsc (r = − 0.66, p textless 0.001) were significantly correlated with ΔHbDiff. After adjusting for MSFsc the relationship between sleep duration and ΔHbDiff was reduced to nonsignificant levels (r = 0.34, p = 0.11), while adjusting for sleep duration did not change the significant relationship between MSFsc and ΔHbDiff (r = − 0.62, p = 0.001). One night of sleep extension increased sleep duration by 140 min, on average, but no change in ΔHbDiff was observed. Dividing participants into earlier and later chronotypes revealed greater ΔHbDiff responses in earlier chronotypes that persisted after the night of sleep extension (mean ΔHbDiff difference = 1.35 μM, t = 2.87, p = 0.006, Hedges’ g = 0.89). These results find chronotype to predict regional cerebral oxygenation responses during working memory processing under conditions of normal sleep and following a single night of sleep extension.