This study proposes an integrated choice and latent variable model (ICLV) for commute mode choice that incorporates satisfaction of human needs and perceived functional and psychological barriers to using certain modes. The modelling framework is validated by data from a survey of commuters in the Greater Copenhagen area, which has large numbers of car users, public transport riders and bicyclists. The model results suggest that higher bicycle use is correlated to positive cycling self-concepts. Similarly, the commute choice of driving is positively correlated with car self-concepts and negatively correlated with functional difficulties in car use. Respondents with a strong focus on functional travel needs are most likely to commute using a car and least likely to use public transport. Evaluation of the effects of improving conditions for bicycles showed that the latent variables had a large influence on the potential mode shifts, highlighting that the mode choice of travellers is largely associated with mode-specific perceptions and fulfilment of travel needs rather than solely level-of-service characteristics. By analysing the mode shifts across the latent variables, further insights on the motives for travel behaviour decisions were obtained, thereby highlighting the superiority of ICLV models to simple multinomial logit models. Furthermore, the study also revealed that socio-economic variables could explain mode choice both directly and indirectly through their impact on the latent variables. This means that a given policy might have a different impact according to the present ICLV model than when estimated by traditional models.