Research has shown that exposure to bright white light or blue-enriched light enhances alertness, but this effect is not consistently observed in tasks demanding high-level cognition (e.g., Sustained Attention to Response Task-SART, which measures inhibitory control). Individual differences in sensitivity to light effects might be mediated by variations in the basal level of arousal. We tested this hypothesis by measuring the participants' behavioural state of vigilance before light exposure, through the Psychomotor Vigilance Task. Then we compared the effects of a blue-enriched vs. dim light at nighttime on the performance of the auditory SART, by controlling for individual differences in basal arousal. The results replicated the alerting effects of blue-enriched light, as indexed by lower values of both proximal temperature and distal-proximal gradient. The main finding was that lighting effects on SART performance were highly variable across individuals and depended on their prior state of vigilance. Specifically, participants with higher levels of basal vigilance before light exposure benefited most from blue-enriched lighting, responding faster in the SART. These results highlight the importance of considering basal vigilance to define the boundary conditions of light effects on cognitive performance. Our study adds to current research delineating the complex and reciprocal interactions between lighting effects, arousal, cognitive task demands and behavioural performance.