Hidden trajectories and cognitive impairment risk: the use of latent change score models to assess the processing speed theory of cognitive ageing

OBJECTIVES: To test the processing speed theory of cognitive ageing by modelling latent changes in processing speed in older adults and examining whether change predicts later global cognitive status and cognitive impairment risk. METHOD: Data were drawn from 2161 community-dwelling adults aged ≥65 years (mean age = 72.57, SD = 4.68; 51% female) in The Irish Longitudinal Study on Ageing. Latent change score modelling was applied to Waves one and three. A multi-indicator latent processing speed factor was defined using Colour Trails Test 1 and 2 and Sustained Attention Reaction Time, with measurement invariance constraints across waves. Latent change in processing speed was used to predict Wave three Mini-Mental State Examination (MMSE), adjusting for baseline MMSE and covariates (age, sex, education, socio-economic status). RESULTS: Evidence supports a stable latent processing speed factor across waves. There were significant interindividual differences in processing speed change, with poorer baseline performance associated with greater subsequent decline (β = 0.338, p < .001). Greater decline in processing speed predicted lower MMSE at Wave three (β = -0.211, p < .001). CONCLUSION: Decline in processing speed throughout time is associated with later global cognitive status beyond baseline levels, supporting processing speed change as a clinically relevant marker of cognitive ageing.