Introduction: Dementia is becoming increasingly prevalent in the UK. Older adults from black and south Asian communities have a higher risk for dementia due to an increased prevalence of dementia specific risk factors such as hypertension, diabetes, and heart disease. Deprivation has also been linked to an increased risk of dementia. Ethnic minority and lower socioeconomic groups are underrepresented in dementia research. The aim of this study was to explore factors influencing diversity in dementia and rehabilitation research within the context of the Promoting Activity, Independence, and Stability in Early Dementia (PrAISED) randomised controlled trial (RCT). Methods: We conducted an exploratory sequential mixed methods study to explore disparities in socioeconomic and ethnic diversity between the PrAISED RCT population and recruitment pathways used in one study site (Nottinghamshire) and compared these with regional and national data. We aimed to collate and summarise data available on ethnicity and deprivation for recruitment/referral pathways (Nottinghamshire site) and the PrAISED cohort (all sites). Additionally, we interviewed healthcare professionals (n=2), researchers (n=2) and members of black and south Asian communities (n=4) to explore barriers to participating in research for people with dementia. Results: Under 2% of the overall PrAISED RCT sample (across all sites) were from a non-white ethnic minority background and a third of participants lived in areas with the least deprivation. Referrals to memory assessment services in Nottinghamshire included people from diverse socioeconomic backgrounds, with 7.3% being from non-white ethnic minority communities. Through interviews, several barriers to healthcare, research and rehabilitation were identified. Healthcare barriers included lack of awareness of dementia, mistrust, stigma, fear, and lack of culturally appropriate services. Research barriers included recruitment routes, awareness of research, language, and recruiter beliefs. Barriers to rehabilitation research included a lack of use of culturally appropriate language, more culturally specific barriers, and lack of representation. Conclusion: Participants recruited to the PrAISED RCT were mainly white and socioeconomically privileged. Data recording and access around ethnicity is still inconsistent, making it difficult to ascertain at which point services and research become inaccessible for people from underserved communities. Future research needs to work with these communities to develop innovative solutions to overcome the barriers identified in this study and to put recommendations made into practice.