Frontotemporal dementia (FTD) is a heterogenous neurodegenerative disease and is caused by an autosomal dominant mutation in around one third of cases. This pattern of inheritance enables FTD to be studied in the presymptomatic phase, where individuals carry the genetic mutation but have yet to develop symptoms. There are currently no approved treatments for FTD, although clinical trials aiming to target interventions at the earliest disease stage, are underway. There is an urgent need for biomarkers that can reliably detect and monitor the progression of disease in the presymptomatic period, though there are a distinct lack of sensitive cognitive measures. This thesis aims to establish the validity and sensitivity of a set of digital biomarkers that can be used to measure cognitive function in FTD. I begin this thesis by describing the Ignite computerised cognitive assessment, developing normative properties for the tests through a remote data collection study in over 2,000 healthy controls. I build upon this validation by establishing the concurrent validity of Ignite with gold-standard pen and paper tasks, the test-retest reliability upon repeated administration, and demonstrate the tests are sensitive to presymptomatic impairment across several cognitive domains. I also describe a novel portable eye tracking experiment that can be completed outside of the lab, first highlighting the validity of the tests as measures of cognitive function and demonstrating their sensitivity in detecting early changes in social cognition in the presymptomatic period. Finally, I investigate a smartphone app that passively monitors human-device interactions to generate digital biomarkers of cognitive function. I establish the acceptability of the app in the general population before demonstrating the measures produced can detect differences in keyboard interactions in presymptomatic FTD mutation carriers. This work provides evidence that biomarkers generated from different digital devices are valid and sensitive measures of cognitive impairment in FTD. Therefore, digital biomarkers could replace outdated pen and paper tasks and be used as outcome measures in clinical trials.