Lighting is one of the most significant factors in the built environment that affect human health and wellbeing. In addition to stimulating visual responses, light induces a range of non-visual responses in humans including synchronizing circadian rhythms and directly alerting the brain. These effects are primarily mediated via a novel photoreceptor that contains the photopigment melanopsin. The spectral sensitivity of melanopsin is shifted towards the blue part of the spectrum compared to rod and cone photoreceptors used for vision. Therefore, making the most of the available daylight, which is naturally rich in the blue part of the spectrum, is a promising approach and might play a large role in lighting recommendations for health. Lighting simulation software tools are designed to predict and analyze the dynamically changing nature of functions that take place in architectural settings influenced by occupants’ behavior and the outside environment. Because non-visual responses adapt to changes in light intensity and spectral composition over much longer time periods than visual responses, they must be evaluated based on dynamic threshold values with regard to intensity, spectrum, duration, history and timing of light exposure. However, existing lighting design methods, used to assess visual performance and comfort, are based on static threshold values, and thus cannot be directly applied to evaluate non-visual responses. The goal of this thesis is to develop a computer-based lighting simulation framework able to predict direct non-visual responses to light and to validate novel guidelines that can inform designers about how lighting might affect human non- visual responses in the built environment. One of the main challenges to be addressed is the dynamic interaction between daylighting, occupants’ behavior and human non-visual responses to light.