Parkinson’s disease is a common neurological condition which results in progressive and incurable motor and non-motor impairments. The considerable heterogeneity in the presentation and progression of PD remains poorly understood. I will present data adopting a multimodal (empirical and theoretical) approach, to provide insights into the network mechanisms of specific PD phenotypes. I will focus on how this approach can be used to understand and to predict cortico-subcortical interactions leading to movement impairments, cognitive impairments and impairments of motivation – all of which characterise the parkinsonian state. Finally, I will discuss how these insights may be used to inform the next generation of neuromodulation therapies.