Practice effects as a dynamic biomarker of early cognitive change in far-from-onset Huntington's disease

Although cognitive performance typically appears preserved in far-from-onset Huntingtin Gene Expansion Carriers (HGECs), underlying neuropsychological mechanisms may be subtly altered and remain undetected. Practice effects from repeated task exposure provide a sensitive measure of cognitive adaptability, with early disruptions signalling reduced neural efficiency before standard deficits emerge. This study aimed to examine longitudinal practice effects across annual neuropsychological testing in far-from-onset HGECs and to identify the disease burden threshold where these patterns diverge. Data from 2777 HGECs and 2777 age-matched controls were drawn from the ENROLL-HD cohort across five annual assessments. Longitudinal cognitive trajectories were modelled using linear mixed-effects models, adjusting for demographics. Segmented regression models identified disease burden thresholds marking the onset of practice effect attenuation. Quartile-based stratification assessed the influence of genetic burden. An internet-based Practice Effects Trajectories Calculator (PE-TraC) was developed to model normative longitudinal distributions across tests and visualize individual-level trajectories. HGECs showed reduced practice effects in tasks involving processing speed and executive function, with divergence from controls emerging by Year 2. Segmented regression identified the earliest disease burden breakpoint in Stroop Word Reading Task (249.6) and the latest in the Symbol Digit Modalities Test (291.9). Quartile (Q) analyses showed that participants with the highest genetic burden (Q4) exhibited minimal or absent practice effects, Q3 showed delayed decline, and Q1-Q2 followed similar trajectories to controls. Reduced practice effects constitute a sensitive marker of early cognitive dysfunction in far-to-onset individuals. These findings support the inclusion of dynamic cognitive measures in far-from-onset HGECs trials as potential endpoints.