Experimental design

Similar to recent work in individuals with PD, 15 participants were asked to perform cognitive and motor tasks simultaneously to examine dual tasking ability. Each task was performed for 20 seconds. To ensure the reliability of single and dual task assessments: (a) the same examiners were used to administer assessments, b) participants were familiarized with single and dual task assessments prior to administration, (c) assessments were administered in the same order for participants (cognitive tasks prior to motor tasks and single tasks prior to dual tasks [see Table 3 and ​and44 for testing order]), and (d) test–retest data were collected to evaluate the reliability of assessments for individuals with pre‐HD, pro‐HD and healthy controls. Importantly, all assessments, including clinical (UHDRS‐TMS), dual tasking, and neuroimaging assessments were performed within 4 weeks for all participants. All examiners had significant experience working with people with HD and therefore were not blinded to group status. Single and dual task assessments were performed three times each. Specific information on cognitive and motor tasks are detailed below.

Single and dual task performance for pre‐HD, pro‐HD, and healthy controls using mixed‐model ANOVAs. Models were adjusted for gender and age. Adjusted means (95% confidence intervals) are presented for each group. Only the notable interactions/variables (P < 0.05) are shown in the table.

Dual task cost results for pre‐HD, pro‐HD, and healthy controls using mixed‐model ANOVAs. Models were adjusted for gender and age. Adjusted means (95% confidence intervals) are presented for each group. Only the notable interactions/variables (P < 0.05) are shown in the table.

Two numeracy assessments were used to evaluate cognitive performance under single and dual task conditions. Numeracy assessments included the Serial Threes Test (STT) and a Progressive Subtraction Test (PST). 23 The STT requires participants to make multiple subtractions of three from a whole three‐digit number, e.g., 256, 253, 250, 247. The PST required participants to verbalize progressive subtractions of 1, 2, 3, 4, and 5 from a three‐digit whole number and the resulting minuends, e.g., 455, 453, 450, and 446. Different three‐digit numbers were used for each trial to reduce learning effects. Numeracy tests were purposefully selected to evaluate cognitive performance given their proven sensitivity in individuals with HD and ecological validity (use in everyday life, particularly the management of finances and time). 24

Postural stability and force steadiness tasks were used to evaluate motor performance under single and dual task conditions. The sensory organization test (SOT) on the Neurocom Smart Balance Master was used to evaluate postural stability. The SOT comprises six different sensory conditions: 1) eyes open, fixed support, and surroundings, 2) eyes closed, fixed support, and surroundings, 3) eyes open, fixed support, moving surroundings, 4) eyes open, unstable support, fixed surroundings, 5) eyes closed, unstable support, fixed surroundings, and 6) eyes open, unstable support, and moving surroundings. For the purpose of this study, only eyes open conditions (1, 3, 4, and 6) were used to ensure that outcomes were ecologically relevant. Values are expressed as a percentage of the theoretical maximum angle of sway, therefore a score of 100 indicates good stability and no movement of the centre of gravity. The Biodex System 4 was used to evaluate force steadiness in the right plantar flexors at 10% of the maximum voluntary force that participants can generate. Prior to the commencement of force steadiness trials participants performed three maximum voluntary isometric contractions (MVIC) of the right plantar flexors. The highest force (Nm) generated in a single MVIC trial was recognized as the participant’s maximum force output and was used to calculate a 10% submaximal plantar flexion contraction target force. With a television screen positioned in front of the participant, the participant was asked to maintain the real‐time force generation line on the horizontal target force line (10% MVIC). The amplitude of force fluctuations above and below the horizontal target force line (force steadiness) during trials was quantified and used for analysis. The higher the force fluctuation the worse the performance. These tests were selected based on their proven sensitivity in individuals with HD. 25 , 26 , 27 , 28 , 29 , 30 For more detailed information see Supplementary files.

Dual task cost (DTC) values were calculated and analysed to assess performance on dual tasks. The formula used to calculate dual task cost values is provided below 31

This formula enables calculation of DTC values (% change) for cognitive and motor components of each dual task. Negative values indicate a reduction in cognitive and motor performance during dual tasking, compared to single task performance, indicating an interference effect or dual task cost.

T₁‐weighted structural images of the brain were obtained from each participant using a GE Healthcare Discovery MR750W 3T MRI scanner. Images were acquired with a 24‐channel head coil using a 3DIR‐SPGR sequence (TA = 9 m 59 s, TR = 3 s, TE = 3.1 ms, TI = 400 ms, flip angle = 11˚, field of view = 256 mm³, image matrix = 256 × 256 × 256, 1 mm³ isotropic voxels). The T₁‐weighted MRI images were automatically processed with the processing pipeline available in FreeSurfer. 32 FreeSurfer was used to parcellate the T₁‐weighted MRI data into cortical and subcortical brain regions according to the Desikan–Killiany atlas. The analyses were performed on MASSIVE HPC (www.massive.org.au) using the “recon‐all” function. The neuroanatomical labels were inspected for accuracy in all HD and healthy control cases. Using the FreeSurfer processing outputs (aseg.stats), we extracted volume of the striatum (caudate and putamen), which were used in statistical analyses.

Sample size was calculated based on the results reported by Vaportzis et al 5 , 6 on dual task cost in individuals with manifest HD and controls performing two level tasks. At the time of this study, there were no previous studies describing dual task cost in individuals with pre‐HD or pro‐HD. The sample size calculation was therefore based on three group effects and interactions. From the results of these previous studies it suffices that the minimum detectable effect size (Eta‐squared) is set at 0.07. For a mixed‐model analysis of variance (ANOVA), to examine the effects of group, gender, task, and interactions and using an alpha level of 0.05, a statistical power of 0.8, it was estimated a sample size of at least 10 participants per group.

Reliability of the single and dual tasks for each group are given by the intraclass correlation (ICC), estimated with a two‐way mixed model for absolute agreement. These values were estimated using trials collected as part of the single dual tasking testing session. For ICC, values less than 0.5, between 0.5 and 0.75, between 0.75 and 0.9, and greater than 0.90 are indicative of poor, moderate, good, and excellent reliability respectively. 33

Mixed‐model analysis of variance (ANOVA) was used to examine the effects of task (single vs. dual), group (pre‐HD, pro‐HD vs. control), gender (male vs. female) and the two‐way task × group and task × gender interactions on STT, PST, force steadiness, and postural stability, whilst adjusting for participant’s age. Pairwise comparisons were conducted with Bonferroni post‐hoc test, and where appropriate, contrasts were used to compare the premanifest and prodromal HD individuals collectively to healthy controls.

Effect sizes for the ANOVAs are described by partial Eta‐squared with 0.01, 0.06, and 0.14 identified as small, medium, and large effects respectively. Post‐hoc effect sizes are described by Cohen’s d with small, medium, and large effect sizes (in absolute terms) defined by 0.2, 0.5, and 0.8 respectively. 34

General linear modeling (GLM) was used to assess the associations between dual task performance with clinical disease outcomes and striatal volume separately for pro‐HD, pre‐HD, and the healthy controls. The GLMs for the HD groups were adjusted for gender and CAP as covariates, whilst gender and age were accounted for in healthy control GLMs. Benjamini‐Hochberg correction was applied to all p‐values in order to minimize false discovery rate. Within group analysis for dual task performance and dual task cost are presented in the Supplementary Files. All analyses were carried out using IBM SPSS version 25 (IBM SPSS, Chicago, IL). Results were considered significant at P ≤ 0.05.