Motivation is key for successful behaviours. Over the last few years, research in cognitive and behavioural neuroscience has begun to characterise motivation within a cost-benefit decision-making framework. In such accounts, motivation is arises through an evaluation of how much effort (cost) someone is willing to exert to obtain a reward (benefit). Such approaches have proven fruitful for identifying the systems in the brain that underlie subjective evaluations of effort and also impairments to motivation across health and disease. Yet, motivation is not static. The willingness to exert effort is dynamic and changes in different contexts. However, little is known about how motivation can differ so drastically depending on the context. Here, I present novel behavioural and computational accounts of the dynamic nature of the willingness to exert effort. Using computational modelling and brain imaging, in healthy people and in Parkinson’s Disease (PD) patients, I examine the what, when and for whom of motivation. Specifically, I will highlight (i) the domain-general nature of systems that decide whether it is worth exerting different types of effort, (ii) how selfish people are in their willingness to exert effort, and (iii) provide an account of how motivation waxes and wanes over-time due to fatigue. Finally, I will show that such an approach can provide computational signatures of motivational impairments in PD and also variability in motivation in healthy people.