The Impact of a 12-Month Randomized Exercise Intervention on Cognitive Function and Brain Volume in Adults with Down Syndrome

Abstract

BACKGROUND:

Adults with Down syndrome (DS) have a high risk for Alzheimer’s disease (AD). While exercise improves cognition and brain health in the general population, few long-term studies have evaluated its effects in DS.

OBJECTIVE:

We examined the impact of a 12-month remotely delivered exercise program on cognitive function and brain volume in adults with DS.

METHODS:

81 adults with DS (mean age = 27 years) were randomized to high-frequency remote exercise (3 sessions/week, RH), low-frequency remote exercise (1 session/week, RL), or a support and education control (SE). Cognitive function was assessed using the DS-adapted Cambridge Neuropsychological Test Automated Battery (CANTAB), and brain volumes were measured via MRI at baseline and 12 months.

RESULTS:

There were no significant changes in any group in executive function or episodic memory (all p > 0.05), but the RH arm significantly improved processing speed across the 12-month intervention (EMM change: 0.17; p = 0.04). No between-group differences were observed for changes in overall cognitive scores. The RL group showed a decrease in total grey matter volume (EMM change −0.82; p=0.02) and the RH group had no change (EMM change: 0.08; p=0.72), yielding a significant group effect across time (EMM: 0.90; p=0.04). Additionally, the RH arm had an increase in right hippocampal volume over 12 months (EMM change: 0.004; p = 0.04).

CONCLUSION:

The improvements in reaction time, right hippocampal volume, and grey matter preservation suggest that structured exercise may influence cognition and brain health in adults with DS.

INTRODUCTION

Down syndrome (DS) or trisomy 21 is the most common chromosomal abnormality associated with an intellectual disability (ID)¹. Due to the overproduction of amyloid-beta (Aβ) from the additional copy of the amyloid precursor protein found on chromosome 21, all adults with DS will develop pathology related to Alzheimer’s disease (AD)². Approximately 90% of individuals with DS will develop clinical dementia in their lifetime², with a median age of onset around 53.8 years³. Notably, Aβ deposition begins decades earlier in individuals with DS compared to the general population². Despite this earlier onset, recent biomarker studies suggest that the trajectory of AD pathology in DS mirrors that of late-onset AD², indicating that prevention strategies effective in the general population may also benefit adults with DS.

In individuals without DS, an estimated 45% of AD cases are linked to 14 modifiable risk factors, including physical inactivity⁴. Moderate-to-vigorous physical activity (MVPA) in both midlife⁵ and older adulthood⁶ has been shown to maintain or improve cognitive domains such as attention, memory, and executive function. Similar findings have emerged in individuals with DS^7,8. For example, Pape et al.⁸ followed 214 participants across 12 months, and observed that higher baseline MVPA was associated with a 62% reduced risk of decline in memory and orientation.

In adults without DS, MVPA obtained through structured exercise appears to have the greatest benefit on cognition^9,10. Group-based structured exercise simultaneously addresses three of the five recommended lifestyle interventions for enhancing cognition and preventing dementia: increased physical activity, mental stimulation, and social engagement¹¹. Beyond cognitive benefits, structured exercise may also improve brain health by increasing hippocampal volume, improving cerebral perfusion, facilitating neurogenesis and synaptogenesis, and mitigating Aβ accumulation^12,13. These protective effects are thought to occur through multiple mechanisms, including enhanced cerebrovascular function and improved cardiorespiratory fitness. However, evidence on the impact of exercise on cognition and brain health in adults with DS remains sparse. Existing studies are generally small, underpowered, and of short duration (≤12 weeks), yet preliminary findings suggest that exercise may offer cognitive benefits in adults with DS as well^14–19. We are unaware of any previous long-term exercise interventions examining the effects of sustained physical activity on cognitive function or brain health in adults with DS.

We recently completed a 12-month randomized controlled trial involving 81 adults with DS without dementia to evaluate the feasibility and efficacy of a remotely delivered exercise intervention designed to increase daily MVPA and cardiorespiratory fitness²⁰. Participants were randomized in a 2:2:1 ratio to one of three groups: low-frequency remotely delivered group exercise (1 session/week; RL), high-frequency group exercise (3 sessions/week; RH), or a support and education control (SE) arm that did not receive group exercise sessions. After 12 months, the RH group demonstrated improvements in both MVPA and cardiorespiratory fitness (VO₂ Peak) that could be associated with improved health and cognitive outcomes in adults with DS²¹. The purpose of the current analysis was to examine exploratory outcomes of the 12-month exercise intervention, specifically evaluating changes in cognitive performance and brain volume by intervention arm.

METHODS

Overview

Adults with DS were randomized in a 2:2:1 ratio to the RL, RH, and SE arms. All participants received a Fitbit^® Versa to track steps, resources designed to promote physical activity, and twice-monthly, 20-minute individualized support sessions delivered remotely. Participants in the RL and RH arms also took part in group exercise sessions (~40 minutes per session, including ~30 minutes of MVPA) conducted via Zoom^™ on iPads^® provided by the study. While the session content was identical across RL and RH groups, they differed in frequency: RL participants attended once per week, while RH participants attended three times per week. All intervention sessions were led by trained health coaches with experience supporting adults with DS. The study was approved by the University’s Institutional Review Board and registered at ClinicalTrials.gov. It was conducted between February 2020 and November 2023. A comprehensive description of the study design²⁰ and the primary outcome (change in MVPA) has been published previously²¹.

Participants

Eligible participants were adults (≥18) with DS living with a caregiver willing to act as a study partner. Inclusion required basic functional ability, verbal communication, physical activity capacity, and home internet access. Exclusion criteria included dementia (per DSQIID)²², regular exercise (≥20 min/day, ≥3 days/week), serious medical risks, or inability to ambulate unassisted.

Recruitment/randomization

Cohorts (n=4) of ~ 20 participants were recruited using flyers, list-serves, and social media posts distributed by local organizations that provide services to adults with DS between January 2020 and October 2022. Written informed consent was obtained from the participant (self as guardian) or a legal guardian with participant assent. Eligible participants were stratified by age (18–30 or >30 years) and sex and computer-randomized to the RL, RH, and SE arms in a 2:2:1 allocation.

Intervention

MVPA recommendations.

All participants were encouraged to meet the Physical Activity Guidelines of 150 minutes/week of MVPA (≥3 METs)²³, progressing from 60 to 150 minutes by month 4. They wore a Fitbit^® Versa on their non-dominant wrist throughout the 12-month trial to self-monitor activity.

Individual support sessions.

All participants received an iPad^® preloaded with Zoom^™ and educational materials. Twice-monthly 20-minute remote sessions with a health coach supported their 150-minute/week MVPA goal through data review, goal setting, and tailored strategies. To minimize variability, the same health coach served all study arms within each cohort.

Remote group exercise (RH and RL arms).

All exercise sessions were led via Zoom^™ by a health educator experienced with DS and accessed on study-provided iPads^® (with optional TV display via HDMI adaptor). Each 40-minute session included a warm-up, ~30 minutes of moderate-to-vigorous aerobic and resistance training, and a cool-down. Intensity progressed from light-to-moderate (<64% HRmax) to ≥64% HRmax²⁴ over the first 6 weeks and was maintained for the remainder of the 12-month intervention.

Outcomes

All outcomes were assessed in our on-campus laboratory at baseline 6, and 12 months by trained staff masked to intervention arm. This study was conducted during the COVID-19 pandemic, and restrictions prevented some subjects from coming in for in-person testing. Thus, we were only able to collect imaging data on 52% of the sample.

Cognitive function

Cognitive function was assessed using the Cambridge Neuropsychological Test Automated Battery for DS (CANTAB^®, Cambridge Cognition, LTD, Cambridge, UK)²⁵ administered on an iPad^® in a quiet room, or in the participant’s home in the case of remote testing. The CANTAB^® DS Battery, which has been used in previous trials in individuals with DS²⁶, has demonstrated sensitivity to disease-specific cognitive deficits in DS, including those related to hippocampal dysfunction, and sensitivity to changes in cognitive function associated with AD²⁷. Outcomes from the CANTAB^® computerized cognitive battery were combined to form three cognitive composite scores: executive function, processing speed, and episodic memory, and Z-scores were created based on the sample. All composites were scored such that lower scores indicate worse cognitive function.

Executive function, the ability of participants to ignore task-irrelevant information and direct higher-order cognitive processes, was formed from each component in the multitasking task, including multitasking cost, reaction latency, incongruency cost, and the total incorrect items. For this task, an arrow was presented on either side of the screen, pointing in either direction (right or left). Participants had to pay attention to either the side of the screen where the arrow appears or the direction of the arrow (indicated by the cues SIDE or DIRECTION on the screen), by pressing a button on the left or right corner on the screen, respectively. Cues were provided either in blocks (single task condition) or randomized from trial to trial (multi-tasking condition). Items could also be congruent (arrow direction and side are the same) or incongruent (item direction and side are different).

Processing speed was formed from the reaction time test, a measure of attention and psychomotor speed. During the task, circles (one for the simple task and five for the choice mode) were shown at the top of the screen, in which a random yellow light appears. Participants had to hold a button on the bottom of the screen and release it to select the circle above in which they detected the yellow light as fast as possible and then return their finger to the hold button. Outcomes were the median time taken to a) release the button or b) select the new button in either the simple or five-choice conditions.

Episodic Memory was formed from the paired associates learning test. Participants were presented with boxes on the screen in which different visual patterns are shown one by one. After the encoding phase, the different patterns were shown in the middle of the screen and participants had to select the box in which the pattern was previously presented. Outcomes were the number of correct choices on the first attempt as well as the total number of errors across the entire task.

Brain volume

Scanning was performed on a 3-Tesla Siemens Skyra scanner (Siemens, Erlangen, Germany) fitted with a 20-channel head coil. Following automated scout image acquisition and shimming procedures to optimize field homogeneity, participants completed a 40-minute scan sequence, which included a structural scan and resting state functional MRI (rsMRI), Diffusion tensor imaging (DTI), and Arterial Spin Labeling (ASL). Participants were able to terminate the scan at any time during the procedure. Due to a high amount of movement in the rsMRI, DTI, and ASL sequences, this analysis only included the structural scan data. The structural scan consisted of a T1-weighted MRI (3D MPRAGE sequence, TR/TE = 2400/2.31 ms, field of view = 300×320 mm², 0.8 mm isotropic resolution, 224 slices). Anatomic images obtained in each session were aligned and normalized to a Montreal Neurological Institute (MNI) brain template. Volumetric analyses of overall and regional, e.g. hippocampus, brain volumes were performed using FreeSurfer segmentation. Brain volumes were normalized by taking the total gray, white, and hippocampal volumes (mm3) divided by total intracranial volume (mm3).

Analysis

Sample descriptive characteristics and outcome measures were summarized using means and standard deviations for continuous variables and frequencies and percentages for categorical variables. For all cognitive and imaging outcomes, change scores from baseline were analyzed using mixed models for repeated measures (MMRM) with an unstructured covariance matrix using the MMRM package in the R statistical computing environment. All models included fixed effects of intervention arm, time point (6 months and 12 months), the interaction between intervention arm and time point, baseline scores, age, and sex. We report means and group differences in change from baseline at the 12-month visit estimated using emmeans. For imaging outcomes, there were only 4 individuals in the SE arm with imaging data at baseline and 3 with imaging data at 6 and 12 months. Given the small sample size, we excluded the SE arm from our analysis and compared changes between the RH and RL arms to explore dose effects of exercise. For all analyses, effect sizes were calculated using Cohen’s d with contrast by group for each timepoint.. Analyses were conducted using R version 4.3.1.

RESULTS

Participants

Adults with DS (n=81, age = ~27 years, 55% female) were randomized to one of three intervention arms (RH= 34, RL=32, SE=15). Baseline participant characteristics are presented in Table 1 and our consort diagram is presented in Figure 1. Participant retention was 100% at 12 months, and attendance at remote group exercise sessions across the 12 months was 90% in the RH arm and 79% in the RL arm. Across all arms, MVPA and cardiorespiratory fitness increased by 5 ±15 minutes/day and 0.88 ± 4.4 mL/kg/min (5.5%), respectively.

Table 1.

Descriptive characteristics of adults with Down syndrome by intervention arm

	Remote High (RH)	Remote Low (RL)	Support/education (SE)
	n=34	n=32	n=15
Age (yrs.)a	25.5 ± 8.5	27.8 ±10.4	25.5 ± 6.6
BMI (kg/m2)a	30.5 ± 5.6	34.4 ± 7.1	30.5 ± 11.7
Female	19 (56%)	19 (59%)	7 (47%)
Hispanic/Latino	3 (9%)	3 (9%)	1 (6.7%)
Race
White	32 (94%)	21 (66%)	15 (100%)
Black	2 (6%)	3 (9%)	0 (0%)
Other	0 (0%)	8 (25%)	0 (0%)
Severity of IDD
Mild	25 (74%)	21 (66%)	10 (67%)
Moderate	9 (26%)	11 (34%)	5 (34%)
Co-Occurring Health Conditions
Congenital Heart Defect	20 (59%)	19 (59%)	7 (47%)
Type 2 Diabetes	2 (6%)	2 (6%)	0 (0%)
Hypertension	1 (3%)	0 (0%)	1 (7%)
Hypothyroidism	14 (41%)	11 (34%)	8 (53%)

^aData are mean ± Standard deviation or n (%).

Figure 1.

Consort Diagram

Changes in Cognition

Cognitive data were available for 72 participants, as 9 participants were unable to complete the full cognitive battery. Composite scores at each timepoint are presented in Table 2, and change over time is presented in Figure 2.

Table 2.

Executive function, processing speed, and episodic memory Z-score in adults with Down syndrome across 12 months by intervention arm. Cohen’s d values represent standardized between-group differences in longitudinal change estimated from mixed models.

	Remote High (RH)		Remote Low (RL)		Support/Education (SE)		Effect sizes (Cohen’s d)
	n	Mean ± SD	n	Mean ± SD	n	Mean ± SD
Executive function
Baseline	32	0.012±0.363	29	−0.025±0.393	12	0.066±0.398	—
6-months	32	−0.002±0.387	27	0.030±0.401	12	−0.058±0.551	SE-RH= −0.08
							SE-RL= −0.13
							RH-RL= −0.06
12-months	32	0.007±0.410	25	0.038±0.393	11	−0.099±0.473	SE-RH= −0.15
							SE-RL= −0.21
							RH-RL= −0.06
Processing speed
Baseline	32	0.132±0.690	29	−0.180±1.108	12	−0.036±1.010	—
6-months	32	0.184±0.512	27	0.114±0.600	12	0.168±0.521	SE-RH= 0.07
							SE-RL= −0.02
							RH-RL= −0.09
12-months	32	0.245±0.465	25	−0.024±0.865	11	0.172±0.497	SE-RH= 0.05
							SE-RL= 0.22
							RH-RL= 0.17
Episodic memory
Baseline	32	0.070±1.039	29	−0.126±0.956	12	0.019±0.899	—
6-months	32	0.219±1.018	27	−0.247±0.925	12	0.676±0.836	SE-RH= −0.05
							SE-RL= 0.19
							RH-RL= 0.24
12-months	32	0.175±0.976	25	0.127±1.058	11	0.108±0.827	SE-RH= 0.07
							SE-RL= 0.08
							RH-RL= 0.01

Figure 2.

Change in Z-score for executive function, processing speed, and episodic memory in adults with Down syndrome across 12 months in the RH and RL arms

In executive function, no group had a statistically significant change in mean Z-score across the 12-month intervention. The SE group had a mean change of −0.17 (95% CI: −0.41 to 0.07, p=0.16). The RH had a mean change of 0.00 (95% CI: −0.14 to 0.14, p=0.98) and the RL group had a mean change of 0.06 (95% CI: −0.10 to 0.22, p=0.43), respectively. Between-group comparisons were also not statistically significant, with a change of −0.17 (95% CI: −0.45 to 0.11, p=0.225) between the SE and RH arms, −0.23 (95% CI: −0.52 to 0.06, p=0.112) between the SE and RL arms, and −0.06 (95% CI: −0.27 to 0.15, p=0.558) between the RH and RL arms.

For processing speed, the RH group had a statistically significant improvement in Z-score (0.17; 95% CI: 0.01 to 0.34, p = 0.04). Changes in Z-score for the RL and SE arms did not reach the threshold for statistical significance (RL 0.01, 95% CI: −0.17 to 0.20, p=0.89; SE 0.22, 95% CI: −0.06 to 0.50, p=0.12). Between-group differences at 12 months were 0.05 (95% CI: −0.28 to 0.38, p=0.764) between SE and RH, 0.21 (95% CI: −0.13 to 0.54, p=0.22) between SE and RL and 0.16 (95% CI: −0.89 to 0.41, p=0.204) between RH and RL arms.

For episodic memory, none of the changes from baseline to 12 months reached statistical significance in any group. Change in the SE group was 0.21 (95% CI: −0.18 to 0.58, p=0.296), 0.10 (95% CI: −0.13 to 0.33, p=0.375) in the RH group, and 0.08 (95% CI: −0.18 to 0.34, p=0.533) in the RL group. Similarly, there were no significant between-group differences at 12 months. The SE vs. RH showed a difference of 0.10 (95% CI: −0.35 to 0.55, p=0.655), SE vs. RL also differed by 0.12 (95% CI: −0.34 to 0.58, p=0.602), and RH vs. RL differed by 0.02 (95% CI: −0.32–0.36, p=0.904).

Change in Brain Volume

Normalized total white matter, normalized total gray matter, and left and right hippocampal volumes for the RH and RL arms are presented in Table 3 and change over time is presented in Figure 3.

Table 3.

White matter, grey matter, and left and right hippocampal volume in adults with Down syndrome across 12 months in the RH and RL arms. Cohen’s d values represent standardized between-group differences in change from baseline estimated from mixed models, scaled by baseline pooled standard deviations.

	Remote High (RH)		Remote Low (RL)		Effect sizes (Cohen’s d)
	n	Mean ± SD	n	Mean ± SD
White Matter Volume (%)
Baseline	23	26.6 ± 1.2	15	27.2 ± 1.7	—
6-mos.	22	26.8 ±1.3	13	27.3 ± 2.0	0.12
12-mos.	20	26.7 ±1.2	13	27.1 ± 1.9	0.06
Grey Matter Volume (%)
Baseline	23	43.1 ± 2.9	15	43.1 ± 2.3	—
6-mos.	22	43.1 ± 2.6	13	42.7 ± 2.5	0.15
12-mos.	20	43.3 ± 2.9	13	42.3 ± 1.9	0.34
Left Hippocampal Volume (%)
Baseline	23	0.260 ± 0.015	15	0.251 ± 0.020	—
6-mos.	22	0.258 ± 0.017	14	0.246 ± 0.019	0.14
12-mos.	20	0.262 ± 0.017	14	0.247 ± 0.021	0.28
Right Hippocampal Volume (%)
Baseline	23	0.269 ± 0.022	15	0.257 ± 0.025	—
6-mos.	22	0.270 ± 0.021	14	0.256 ± 0.029	0.08
12-mos.	20	0.276 ± 0.023	14	0.261 ± 0.026	0.20

Figure 3.

Change in percent normalized total white matter, grey matter, and right and left hippocampal volume in adults with Down syndrome across 12 months in the RH and RL arms

Normalized total white matter volume did not change over 12 months in both groups (RH: −0.04, 95% CI = −0.31 to 0.23, p=0.77; RL: −0.12, 95% CI = −0.47 to 0.23, p=0.49). with no significant between-group difference (0.08, 95% CI = −0.37 to 0.53, p = 0.72).

In contrast, normalized total gray matter trajectories were significantly different. The RH group had no change in normalized total grey matter volume (0.08, 95% CI = −0.40 to 0.57, p=0.72), whereas the RL group experienced a significant decline (–0.82, 95% CI = −1.48 to −0.15, p = 0.02). This yielded a significant between-group difference at 12 months (0.90, 95% CI = 0.07 to 1.73, p = 0.04).

Neither group changed significantly in left hippocampal volume (RH: 0.001, 95% CI:–0.003–0.005, p=0.67; RL −0.003, 95% CI: −0.01–0.002, p=0.20), but the RH group showed a significant increase in right hippocampal volume (0.004, 95% CI = 0.001 to 0.009, p=0.04), while the RL group did not (0.000, 95% CI = −0.005 to 0.006, p=0.95). Between-group differences were not statistically significant (left: p = 0.212; right: p = 0.287).

DISCUSSION

Over the 12-month intervention, significant gains in the processing speed composite score were observed in the RH arm, demonstrating the potential cognitive benefits of long-term exercise in adults with DS. However, there were no significant changes in the executive function or episodic memory domains in any group. Notably, this exploratory analysis was not the primary study aim; thus, these cognitive findings should be considered exploratory. The sample sizes provided limited statistical power to detect small-to-moderate differences in the cognitive assessments, and replication in larger samples is needed to confirm these preliminary trends. Additionally, there are no thresholds currently defined for the CANTAB^® to detect clinically meaningful changes in DS, so any clinical interpretation of these findings is limited.

Previous acute and short-term interventions suggest that exercise may enhance executive function and processing speed in adults with DS^14–19. For example, Chen et al¹⁶ compared the response of 3 components of executive function (choice-response time, attention shifting and inhibition) to an acute bout of moderate intensity exercise (20 min treadmill walk, n=10) with an attentional control arm (20 min video, n=10) in adults with DS. Compared to the control group, participants in the exercise group had non-significant improvements in choice-response time (p = 0.32) and attention shifting (p = 0.13). Similarly, Shields et al¹⁸ observed improvements in executive function after a 12-week community-based program that involved two-1 hour aerobic-based exercise sessions per week.

Notably, the current analysis did not reveal any significant changes or non-significant trends in episodic memory over time or between intervention arms. This contrasts with findings from our prior 12-week pilot study, which demonstrated significant improvements in visual episodic memory (p = 0.05, d = 0.76) among participants who completed two exercise sessions per week, compared to those who participated once weekly¹⁷. Several factors may explain this discrepancy. First, differences in baseline cognitive status, which were not assessed and may have varied across groups, could have masked intervention effects. Second, improvements in episodic memory in the pilot may reflect practice effects due to the short 12-week interval between assessments. Practice effects have been well documented as a consideration when administering cognitive assessments²⁸, including in populations with DS, and should be considered in future interventions. Future trials should include baseline cognitive evaluations and use methods like alternate test forms or longer intervals to reduce test–retest effects and improve the validity of cognitive outcome measures in long-term interventions.

Normalized total grey matter volume was preserved in the RH group but declined significantly in the RL group, yielding a significant group difference across time, suggesting a potential neuroprotective effect of the higher dose intervention. Grey matter loss is a marker of poor brain health²⁹, with declines in volume associated with decreased cognitive function³⁰. Previous work in adults without DS demonstrates that higher amounts of physical activity are associated with slower decline in grey matter volume and a lower risk for cognitive impairment³¹. In adults with DS, cross-sectional studies have shown that greater cardiorespiratory fitness is linked to higher gray matter volume^32,33, and intervention studies have reported increases in gray matter volume following aerobic exercise³³. Given that earlier findings from this trial revealed a greater improvement in VO₂ Peak in the RH group compared to the RL group²¹, the observed preservation of grey matter volume aligns with patterns reported in those without DS. Together, these findings suggest that higher doses of aerobic exercise may promote brain health in adults with DS,; however, given the exploratory nature of this analysis, these findings need to be confirmed in a larger study that is adequately powered to detect these differences.

Right hippocampal volume, a key structural biomarker linked to memory, increased in the RH group, and did not change in the RL arm. Hippocampal atrophy is reliably associated with cognitive decline in both adults with and without DS^2,34. Specifically, the right side of the hippocampus has been associated with spatial memory and processing as opposed to the left side, which has been associated with verbal memory³⁵. To our knowledge, this is the first study examining changes in hippocampal volume following an exercise intervention in adults with DS. However, previous work in adults without DS has demonstrated that aerobic exercise is associated with hippocampal neurogenesis and volume preservation or growth, especially when coupled with increases in cardiorespiratory fitness³⁶. Thus, the observed hippocampal gains in the RH group are promising, especially given the known trajectory of accelerated hippocampal atrophy in adults with DS².

Improvements in reaction time, preservation of grey matter, and increases in right hippocampal volume may reflect exercise-related adaptations associated with cognitive resilience and brain health in adults with DS. Executive dysfunction, including deficits in inhibition, working memory, and cognitive flexibility, is often among the earliest cognitive changes observed in individuals with DS who later develop AD³⁷. Improvements in reaction time may therefore reflect enhanced processing efficiency or compensatory neural functioning that could support maintenance of daily functioning and cognitive performance over time³⁸. Similarly, hippocampal atrophy is well documented in adults with DS and is associated with age-related memory decline and progression toward AD-related pathology, with structural changes occurring years before clinical symptom onset². Given the relatively young age of our cohort, the current findings are more appropriately interpreted as evidence of short-term neuroplastic or brain health adaptations rather than disease modification or AD prevention. Nevertheless, the observed improvements in reaction time, right hippocampal volume, and preservation of grey matter compared with the RL group suggest that structured exercise may support mechanisms related to cognitive resilience and maintenance of brain health in this high-risk population.

This study had several notable limitations. First, cognitive and neuroimaging outcomes were exploratory, and the study was underpowered to detect small effects. In addition, the COVID-19 pandemic substantially reduced opportunities for in-person assessments, limiting the neuroimaging sample to approximately 52% of the original cohort. Participants who completed MRI assessments may also have differed systematically from those who did not complete imaging, including differences in cognitive functioning, behavioral tolerance, caregiver support, or overall health status, introducing the potential for selection bias and limiting interpretation and generalizability of the neuroimaging findings. Second, although the CANTAB^® battery has been validated for use in individuals with Down syndrome (DS), it may lack the sensitivity needed to detect subtle, domain-specific cognitive changes. Emerging assessment tools, such as the Down Syndrome Mental Status Examination (DSMSE)³⁹ and modified cued recall tasks⁴⁰, may provide greater precision and responsiveness in future trials. Third, the relatively young age of participants (mean age ~27 years) limited our ability to examine the effects of exercise on AD related cognitive decline, which more commonly emerges in the fourth and fifth decades of life in individuals with DS. In addition, the findings may have limited generalizability, as participants were volunteers enrolled in a structured exercise intervention and may not fully represent the broader population of adults with DS, including those with greater medical complexity, lower baseline functional abilities, or limited access to community-based exercise programs. Finally, due to the exploratory nature of this analysis, being the first to our knowledge to evaluate the effect of a 12-month exercise intervention on changes in cognition and brain volume in adults with DS, we did not adjust for multiple comparisons in our analyses. The priority of this analysis was to identify meaningful trends that could inform future studies for people with DS. Thus, there were no corrections for multiple comparisons, which could increase the risk of Type I error. Future studies should include broader age ranges, larger and more diverse samples, and more sensitive cognitive and biomarker measures to better evaluate the potential of exercise to prevent or delay AD-related decline in DS.

CONCLUSION

This 12-month randomized controlled trial demonstrates that a remotely delivered, structured exercise intervention can positively impact brain structure and improve processing speed in adults with DS. These findings underscore the feasibility and value of long-term physical activity interventions for improving brain health in DS. Future trials with larger samples and age-diverse cohorts are warranted to validate these results and refine strategies to promote brain health in adults with DS.

DATA AVAILABILITY STATEMENT

Deidentified individual participant data (including data dictionaries) are publicly available through Include Data Coordinating Center at the following link: https://portal.includedcc.org/studies/BrainPower. The exercise session curriculum can be downloaded from Move-DS© (touchnet.net) (https://secure.touchnet.net/C21581_ustores/web/product_detail.jsp?PRODUCTID=1778&SINGLESTORE=true)