Motor, not cognitive, performance relates to amyloid status in normal older adults

Abstract

INTRODUCTION

The National Insitutes of Health Toolbox (NIHTB) measures may be useful for Alzheimer's disease (AD) risk detection. This study investigates whether cognitively normal older adults with and without elevated brain amyloid, an early AD biomarker, differed on NIHTB cognitive and motor subtests and whether inclusion of pTau‐217 enhanced this differentiation. Motor measures were of interest based on past research reporting linkages to early AD pathology.

METHODS

Data from 112 consensus‐diagnosed cognitively normal older adults who completed amyloid‐PET imaging, blood draw for pTau217 analysis, and NIHTB Cognition and Motor batteries were analyzed using linear regression.

RESULTS

Amyloid positivity significantly predicted performance on only NIHTB Standing Balance (F(2,86) = 9.90, p < 0.001). Inclusion of pTau217 enhanced prediction of Standing Balance (F(4,84) = 5.88, p < 0.001) and Walking Endurance (F(4,88) = 9.70, p < 0.001) and Dominant‐hand Grip Strength (F(2,92) = 51.59, p < 0.001).

DISCUSSION

For AD‐related detection research, findings support inclusion of NIHTB motor measures, which may prove more sensitive early in the disease course before overt cognitive change.

Highlights

• The National Institutes of Health ToolBox motor performance may signal early Alzheimer's disease (AD) risk, even before memory change.

• Balance performance may be sensitive to early cumulative AD‐neuropathology loading.

• Motor measures should be included in early AD‐risk detection batteries.

1. BACKGROUND

Early detection of those at risk of developing cognitive impairment is crucial for timely implementation of interventions and safety initiatives. Elevated accumulation of amyloid in the brain is a potential marker of early Alzheimer's disease (AD) in healthy individuals who are not yet showing overt clinical signs of cognitive impairment ¹ , ² ; however, stigma, high cost, and invasiveness of procedures to assess amyloid burden, including positron emission tomography (PET) and lumbar puncture for cerebrospinal fluid (CSF), limit broad implementation. Work is underway to identify easy‐to‐administer, widely available, cost‐effective tools to screen cognitively normal individuals for early AD risk. ¹ The National Institute of Health Toolbox (NIHTB) is a noninvasive test battery designed to assess cognitive, motor, sensory, and behavioral functioning and is widely available, time‐efficient, cost‐effective, and easy‐to‐administer within research settings. ³ The NIHTB is conormed across dimensions and thus may be suitable for characterization of domain‐based decline in cognitively normal older adults that signals preclinical AD.

Two NIHTB domains, Cognition and Motor, are of primary interest. Cognition is important as objective cognitive decline is a defining clinical symptom of AD. However, neuropathological changes may appear 18 or more years before the emergence of overt cognitive decline ² , ⁴ and potential changes early in the disease course are likely subtle. Previously, Cheng and colleagues looked at NIHTB Cognition Battery (NIHTB‐CB) data from older adults with varying cognitive status (cognitively normal, mild cognitive impairment (MCI), and early‐stage AD) who had documented amyloid burden via PET or CSF, and found multiple cognitive features predicted amyloid positivity including performance on memory, working memory, processing speed, picture vocabulary, total cognitive composite, and fluid cognition composite. ⁵ These findings are expected as amyloid levels are typically higher among those with AD compared to cognitively normal adults, and the hallmark clinical symptom of AD is poor episodic memory. Interestingly, noncognitive features assessed by the NIHTB Motor Battery (NIHTB‐MB) were also predictive of amyloid status including performance on hearing, odor identification, and walking tasks.

RESEARCH IN CONTEXXT

1. Systematic review: We reviewed literature regarding cognitive and motor performance signs of early Alzheimer's disease (AD). The widely available, easy‐to‐administer National Institutes of Health ToolBox (NIHTB) Motor and Cognition batteries have not yet been widely studied for potential use in determining early AD‐risk among cognitively normal older adults.

2. Interpretation: Among older adults, motor performance may signal the earliest signs of AD‐risk even before overt cognitive change. This appears particularly true for Standing Balance. These results support the inclusion of NIHTB Motor measures in early AD risk detection research.

3. Future directions: Future studies should investigate sensitivity curves for pTau‐217 and amyloid beta with regard to early balance changes. Given early AD risk may be captured by several biomarkers, some still in development, there is opportunity for studies to merge multiple biomarkers of early AD into a single “hazard score” against which motor performance can be compared as a potential non‐invasive early AD marker.

Changes in motor functioning, particularly gait and balance, consistently relate to the onset of cognitive impairment but are often overlooked in diagnostic settings. For example, slower gait speed, shorter step length, and greater variability in stride length are seen in individuals with MCI and AD compared to cognitively normal adults, and are associated with future cognitive decline. ⁶ , ⁷ , ⁸ In fact, faster decline in gait speed was observed 12‐years prior to MCI onset. ⁹ The incidence of injurious falls, which may relate to worsening balance, also increases about four years before dementia diagnosis. ¹⁰ Further, in cognitively normal older adults, both slowing gait speed and worsening balance appear linked to AD pathology (e.g., CSF amyloid‐beta [Aβ] 42/40 ratios, regional brain amyloid burden). ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Thus, measurement of balance and walking metrics in cognitively normal older adults may provide valuable information regarding risk of conversion to cognitive decline.

This study aimed to determine whether cognitively normal older adults with elevated brain amyloid (e.g., “amyloid positive” individuals at heightened risk of developing cognitive impairment due to AD) differed in performance on select NIHTB‐CB and NIHTB‐MB tasks compared to those without elevated brain amyloid (e.g., “amyloid negative” individuals at lower risk of developing clinical manifestations of AD). We hypothesized that amyloid‐positive older adults would show subtle, early weakness in episodic memory performance as well as lower walking and balance performance when compared to their amyloid‐negative peers.

As blood‐based biomarkers, such as plasma phosphorylated threonine 217 tau (pTau‐217), are increasingly seen as promising early and accessible markers of AD, we also investigated whether inclusion of pTau‐217 enhanced the NIHTB performance differentiation of our cognitively normal amyloid‐positive and amyloid‐negative participants. Plasma pTau‐217 measurement detects soluble tau and is considered an early marker of AD, similar to amyloid‐PET, that increases early in the course of AD prior to deposition of tau in the neocortex, ¹⁶ , ¹⁷ predicts future cognitive decline among cognitively unimpaired individuals with elevated brain amyloid, ¹⁸ and has relatively high accuracy in detecting AD pathology among individuals with memory complaints or MCI due to AD pathology. ¹⁹ , ²⁰ As such, we hypothesized that inclusion of pTau‐217 will enhance the NIHTB performance differentiation of amyloid‐positive and amyloid‐negative groups.

2. MATERIALS AND METHODS

2.1. Study participants

All study procedures were approved by the University of Michigan Institutional Review Board. Participants signed written informed consent prior to initiation of study procedures. The study was performed in accordance with ethical research standards as laid out in the Declaration of Helsinki and later amendments.

At the time of study analyses, 112 cognitively normal adults enrolled in a longitudinal driving study (R01AG068338) had completed the NIHTB‐CB, sections of the NIHTB‐MB, and had undergone amyloid‐PET imaging. All participants were included in study analyses. Participants were age 65 and older, free of neurologic disease that might negatively impact cognitive functioning (e.g., Parkinson's disease) and major or unstable psychiatric conditions (e.g., schizophrenia). All were coenrolled in the Michigan Alzheimer's Disease Research Center (P30 AG072931) where they were determined to be cognitively normal via Clinical Consensus diagnostic conference following National Alzheimer's Coordinating Center (NACC) guidelines. Data from NIHTB, amyloid‐PET scan, and blood‐based biomarkers were not available at the time of this Clinical Consensus conference. No participants were excluded on the basis of sex, race, or ethnicity.

2.2. NIHTB

The NIHTB is a set of standardized tests that were developed to assess cognitive, motor, emotional, and sensory functioning in a cost‐effective and time‐efficient manner within the research setting, and enable comparison of data across research studies ³ . In this study, version 2.0 of the NIHTB‐CB was administered in full via an iPad tablet. The battery includes measures of crystallized cognitive functioning, made up of oral reading recognition and picture vocabulary, and a set of fluid measures, including assessment of attention, language, processing speed, working memory, memory, and executive functioning (see Reference [21] for measurement details). Measures from version 2.0 of the NIHTB‐MB were also given including bilateral grip strength, standing balance, and walking endurance (see Reference [22] for measurement details). Uncorrected standardized scores were obtained for each of the seven cognitive and four motor tests and used for study analyses as recommended by NIHTB developers.

2.3. Amyloid‐PET

Elevated accumulation of amyloid plaques in the brain, detected by PET scanning, is a hallmark of AD and in those with normal cognitive functioning, may be an early, preclinical marker of the disease ¹ . All participants in this study underwent PET scanning on the University of Michigan Siemens BioGraph TruePoint TrueV 1094 PET/CT research scanner. [C‐11]Pittsburgh Compound‐B (PiB) was used to quantify the amount of Aβ in the brain as well as a head CT for attenuation correction of the PET emission scan. PET scans were acquired as an 80‐min dynamic sequence of 17 frames, following a bolus + infusion protocol (40% bolus, 60% infusion over the 80‐min scan) designed to achieve steady‐state concentrations in the brain by about 40–50 min postinjection. PiB tissue concentration ratios were estimated using the frames from 40 to 80 min postinjection of the radiotracer. Centiloid values were obtained for each participant in accordance with Klunk and colleagues ²³ . Centiloid values of 20 or more were considered to reflect definitive amyloid positivity (n = 36) while centiloid values of 10 or less were considered to reflect amyloid negativity (n = 59) ²⁴ . Seventeen individuals who had centiloid values between 10 and 20 were considered to have less certain amyloid status (i.e., questionable negative to mild positive) and were excluded from further analyses. With the exception of four cases, participants completed PET scans and NIHTB testing within 12 months of each other (21.5 ± 52.9 days) and 95% of participants completed these procedures within 2 months of each other (12.1 ± 12.0 days). The four cases outside of 12 months occurred when a positive amyloid PET scan was obtained 36–65 months prior to NIHTB testing; since amyloid positivity generally does not decrease over time; these individuals were included in the current study without being rescanned to minimize participant burden.

2.4. Blood biomarker: plasma phosphorylated threonine 217 tau (pTau‐217)

Plasma phosphorylated threonine 217 tau (pTau‐217) has emerged as a key blood‐based biomarker for potential prediction of AD pathology among individuals with subjective cognitive complaints and MCI. ¹⁹ , ²⁰ , ²⁵ The pTau‐217 biomarker was available for all but three participants and was included in the models to determine whether its inclusion enhanced the potential relationship between Aβ‐PET and NIHTB performance. One‐hundred and nine participants completed a blood draw within 12 months of their NIHTB testing and plasma pTau‐217 was measured in all samples. Plasma samples were centrifuged at 10,000 × g for 5 min and stored in single‐use aliquots at −80°C until the time of analysis (samples underwent two freeze/thaws; one when plasma was isolated and the second when it was aliquoted). Plasma pTau‐217 levels were determined using the Quanterix ultrasensitive Simoa technology on an HD‐X analytical platform. The commercial ALZpath pTau217 v2 assay (Quanterix, 104371) was used to measure the plasma concentration of pTau‐217. Samples were thawed and run in duplicate with onboard 3X dilution according to the manufacturer's instructions. This assay has a limit of detection of 0.0052 to 0.0074pg/mL, a functional lower limit of quantification of 0.06pg/mL, and a dynamic range of 0.007 to 30pg/mL. The ALZpath Simoa pTau 217 v2 EQC Kit (Quanterix, 104372) and two additional bridging plasma samples were used to calculate intrarun and inter‐run precision. Mean intrarun and inter‐run precision was 5.3% and 11.3%, respectively. All sample duplicates had a coefficient of variance (CV) of <15% (median = 3.7%, range = 0.03%–14.7%). PTau‐217 plasma level data are reported as dilution‐adjusted concentrations.

2.5. Statistical analyses

SPSS version 29.0 was used for all study analyses. Potential demographic group differences were assessed via independent sample t‐test or chi‐squared test across the amyloid groups, as appropriate. Group differences in pTau‐217 levels were also assessed using independent sample t‐test. Posthoc comparisons were conducted using Tukey HSD test as indicated.

Histograms were inspected and Shapiro–Wilk tests were conducted for each NIHTB outcome variable. Given that some skew was anticipated and appropriate given the use of uncorrected Standard Scores derived from the NIHTB nationally representative general adult population age 18–85, Q‐Q plots were also inspected. The majority of Q‐Q plots were normal, however, Grip Strength was slightly skewed for both dominant and nondominant hands. Overall, the suggestion of normality was reasonable and homoscedasticity assumptions were reasonably met. As such, no transformations were applied and the raw data were used for study analyses.

Six individuals had missing data across on the NIHTB Standing Balance task (four amyloid negative, two amyloid positive) and two individuals had missing data on the NIHTB Endurance task (both amyloid negative); this includes one individual who had missing data on both measures. Reasons for missingness were documented during data collection and included: two technology‐related issues (both amyloid negative groups), one recent lower limb surgery (amyloid positive group), one knee pain (amyloid negative group), one participant refusal (amyloid negative group), and two unknown (one amyloid positive group, one amyloid negative group). These values were listwise deleted from the relevant analyses. Although this approach may have slightly reduced statistical power, the potential bias introduced by listwise deletion was likely minimal as the missing motor data were attributable to identifiable reasons consistent with missing at random (MAR) assumptions and represented a relatively small proportion of the total sample.

To determine the relationship between amyloid positivity and NIHTB performance, backward elimination linear regressions were used to assess whether amyloid status predicted NIHTB performance on the eleven NIHTB tests between amyloid positive and negative groups. These analyses were conducted with all variables included in the model in step 1. Given the potential influence of age, sex, and education on neuropsychological and motor performance, bivariate correlations were run for the three variables and all 11 NIHTB tasks separately within each amyloid group. Ultimately, age and sex variables were included in the regression models along with their interactions with amyloid status as detailed below.

To determine whether inclusion of pTau‐217 enhanced potential relationships between amyloid positivity and NIHTB performance, the aforementioned linear regression models were re‐run adding pTau‐217 and the interaction of pTau‐217 and amyloid positivity as variables.

For both study questions, there was an adequate ratio of participants to predictor variables and a backward hierarchical regression approach was used to allow less informative predictors to be sequentially removed, yielding a more parsimonious final model. This, combined with the theoretical considerations to use of covariables (e.g., age, sex) help ensure the number of predictors entered into the initial model and retained in the final model are justified.

3. RESULTS

Please see Table 1 for demographic characterization of the sample by amyloid group. Overall, the sample was largely White, non‐Hispanic, highly educated, and right‐handed. There were no significant demographic differences among the groups although the amyloid positive group was, on average, two years older. As expected, pTau‐217 levels were higher in the amyloid positive group compared to the amyloid negative group (t(42.81) = 6.67, p < 0.001).

TABLE 1.

Characteristics of the study sample

Parameter	Amyloid negative (n = 59)	Amyloid positive (n = 36)	p‐Value	Degrees of freedom	Effect size
Age in years Mean (SD) [range]	71.5 (5.1) [65–86]	73.4 (4.0) [66–81]	0.07	93	4.701
Education in years Mean (SD)	16.5 (2.4)	17.0 (1.9)	0.28	93	2.243
Handedness, right Number (percentage)	54 (91.5%)	31 (86.1%)	0.39	2	0.178
Sex, female Number (percentage)	42 (71.2%)	20 (55.6%)	0.13	1	0.202
Race, White Number (percentage)	51 (86.4%)	33 (91.7%)	0.71	3	0.152
Ethnicity, Hispanic/Latinx Number (Percentage)	2 (3.4%)	0 (0%)	0.52	1	0.145
pTau‐217 (pg/mL) Mean (SD) [range]	0.315 (0.16) [0.12–0.94]	0.722 (0.34) [0.23–1.95]	<0.001	91	0.241

Note: There were no significant demographic differences between the amyloid groups. The p‐tau217 levels were higher in the amyloid positive group compared to amyloid negative group.

Abbreviation: SD, standard deviation.

As expected, sex, age, and education were associated with scores on multiple NIHTB measures within the two amyloid groups. Male sex was positively associated with NIHTB performance on measures of Walk Endurance (amyloid negative: β = 0.32, p = 0.02, amyloid positive: β = 0.33, p = 0.05), Dominant‐hand Grip Strength (amyloid negative: β = 0.67, p < 0.001, amyloid positive: β = 0.77, p < 0.001), and Nondominant hand Grip Strength (amyloid negative: β = 0.63, p < 0.001, amyloid positive: β = 0.72, p < 0.001) within each amyloid group. Age was negatively correlated with Walk Endurance (amyloid negative: β =−0.39, p < 0.01, amyloid positive: β =−0.37, p = 0.03) and Standing Balance (amyloid negative: β =−0.33, p = 0.02, amyloid positive: β =−0.41, p = 0.02) within each amyloid group, and with Picture Sequence Memory within the amyloid negative group (β =−0.27, p = 0.04). Finally, education was positively correlated with two crystallized measures, Oral Reading Recognition (β = 0.28, p = 0.03) and Picture Vocabulary (β = 0.30, p = 0.02), in the amyloid negative group. Given the association of age and sex with primary variables of interest, these demographic variables were included in final regression models investigating the relationship between amyloid status and NIHTB performance; education was not included in the models to preserve effect sizes.

Mean NIHTB scores are shown in Table 2 for each amyloid group and details of the regression analyses are shown in Table S1. Consistent with our hypotheses, NIHTB performance differed between amyloid positive and negative groups on Standing Balance. Amyloid status along with age predicted 18.7% of the variance in Standing Balance scores (F(2,86) = 9.90, p <0 .001). Both amyloid positivity and increasing age were associated with lower Standing Balance scores (β =−0.21 and β =−0.34, respectively). Although age and the interaction of age and amyloid positivity predicted a small percentage of the variance in Picture Sequence Memory scores (6.0%), the overall model did not reach statistical significance ((F(2,92) = 2.96, p = 0.057) counter to our original hypothesis. Other NIHTB measures did not include amyloid or interactions with amyloid in their final models, including Walk Endurance (see Table S1 for details).

TABLE 2.

NIHTB performance by amyloid status

Parameter	Amyloid negative (n = 59)	Amyloid positive (n = 36)
NIHTB‐Cognition Battery
Oral Reading Recognition Mean (SD)	109.5 (5.3)	108.9 (5.2)
Picture Vocabulary Mean (SD)	118.3 (8.4)	116.0 (8.8)
Dimensional Change Card Sort Mean (SD)	103.8 (6.5)	103.1 (6.7)
Flanker Inhibitory Control and Attention Mean (SD)	97.8 (5.7)	96.9 (5.6)
List Sorting Mean (SD)	100.5 (8.8)	98.6 (10.0)
Pattern Comparison Mean (SD)	93.4 (12.9)	89.4 (15.3)
Picture Sequence Memory Mean (SD)	98.5 (11.6)	97.5 (12.1)
NIHTB‐Motor Battery
Grip Strength, Dominant Mean (SD)	97.0 (9.4)	98.4 (9.8)
Grip Strength, Nondominant Mean (SD)	96.3 (9.8)	96.8 (9.9)
Standing Balance Mean (SD)	99.0 (10.5)	92.7 (11.6)
Walk Endurance Mean (SD)	89.9 (10.7)	88.7 (16.8)

Note: Mean uncorrected standard scores are shown for each task along with their standard deviations. Standard Scores have a mean of 100 and standard deviation of 15. Please note, every participant completed the full NIHTB‐Cognitive Battery and bilateral grip strength measures (no missing data). Two participants did not complete the Walk Endurance task (both amyloid negative) and six did not complete Standing Balance (four amyloid negative, two amyloid positive participants).

Abbreviations: NIHTB, National Institutes of Health Toolbox; SD, standard deviation.

Details of the regression models following the addition of pTau‐217 and the interaction of pTau‐217 and amyloid status can be found in Table S2. Addition of these two variables to the models marginally enhanced the model for Standing Balance such that the interaction of amyloid status and pTau‐217 along with amyloid status, pTau‐217, and age predicted 21.9% of the variance in Standing Balance scores (F(4,84) = 5.88, p < 0.001). Amyloid status, pTau‐217, and age were negatively associated with Standing Balance (β =−0.50, β =−0.46, and β=−0.31 respectively) while the interaction of amyloid and pTau‐217was positively associated with Standing Balance performance (β = 0.67). Within the amyloid negative group, there was a negative association between pTau‐217 and Standing Balance (r =−0.32, p < 0.05) while there was a negligible relationship within the amyloid positive group (please see Table S2 for details).

Inclusion of pTau‐217 also enhanced the models for two other motor measures: Walk Endurance and Dominant‐hand Grip Strength. The interaction of amyloid status and pTau‐217 along with amyloid status, sex, and age predicted 29.2% of the variance in Walk Endurance scores (F(4,88)  =  9.07, p < 0.001). Male sex and amyloid status were positively associated with Walk Endurance scores (β  =  0.27 and β  =  0.44, respectively) while age was negatively associated (β  =  −0.33). The interaction of amyloid and pTau‐217 was also negatively associated with Walk Endurance (β  =  −0.53). Within the amyloid positive group, there was a negative association between pTau‐217 levels and Walk Endurance performance (r  =  −0.43, p < 0.01) while there was a negligible relationship within the amyloid negative group. With regard to dominant‐hand Grip Strength, sex and, to a smaller extent, pTau‐217 explained 52.9% of the variance in dominant‐hand Grip Strength (F(2,92)  =  51.59, p < 0.001). Male sex was positively associated with dominant‐hand Grip Strength scores (β = 0.72) while pTau‐217 was negatively associated (β  =  −0.15). The models for all other NIHTB tasks, including Picture Sequence Memory, remained unchanged (please see Table S2 for details).

4. DISCUSSION

As a field, we most often look at faltering memory performance as the early cardinal clinical sign of impending AD. Findings from this study are consistent with a growing body of literature suggesting that changes in motor skills may signal the earliest signs of AD risk even before the presence of overt memory change. In this study of consensus diagnosed cognitively normal older adults, motor performance on the NIHTB related to early neurobiological markers of AD, including elevated brain amyloid assessed via PET scan and plasma pTau‐217 biomarker levels that suggest increased risk for development of AD/ADRD, while NIHTB assessed cognitive performance did not.

The relationship between amyloid positivity and lower Standing Balance in our cognitive normal older adult cohort, and the marginal enhancement of this prediction with inclusion of pTau‐217, suggests that the NIHTB Standing Balance measure may be sensitive to early AD changes and supports its inclusion in early AD risk detection research. Looking more closely at the interaction between amyloid positivity and pTau‐217 in relation to Standing Balance performance, pTau‐217 and balance performance were negatively associated when brain amyloid was low as initially hypothesized; however, once brain amyloid was high, no substantial relationship was seen between pTau‐217 and balance. This may suggest balance sensitivity in early cumulative AD‐neuropathology loading, or may relate to pTau‐217′s relationship with amyloid neuropathology. Similar to Aβ, pTau‐217 increases early in the course of AD, and pTau‐217 is effective at discriminating between individuals with positive and negative amyloid‐PET, particularly in preclinical AD where cognition is not yet impaired as is the case for out study participants. ²⁵ , ²⁶ , ²⁷ Given this study's use of amyloid positivity as a binary variable (positive/negative, removing those with less certain amyloid status) and pTau‐217 as a continuous variable (in part due to lack of consensus regarding what pTau‐217 values are considered “elevated”), worsening motor performance corresponding to increasing pTau‐217 among amyloid‐negative individuals may reflect an enhancement of the relationship between motor performance and increasing neuropathology burden that plateaus when a certain neuropathology threshold is reached. Future research investigating sensitivity curves for both pTau‐217 and Aβ with regard to early balance changes in AD is warranted.

Walking endurance, which may have different cognitive demands or be a less complex motor task compared to balance, was not predicted by amyloid status alone; however, the addition of pTau‐217 was meaningful in terms of predicting walking performance. Interestingly for walking endurance, increasing pTau‐217 was correlated with worsening performance when amyloid levels were high while no substantial relationship was seen when amyloid was low. Given the contrast with balance performance, there may be differences in early motor signs among those with elevated AD biomarkers.

One consideration is whether the early differences in motor functioning presented here reflect elevated general health/frailty, defined as “a medical syndrome… characterized by diminished strength, endurance, and reduced physiologic function that increases an individual's vulnerability for developing increased dependency and/or death,” ²⁸ in amyloid‐positive individuals who are at greater risk of cognitive decline or whether these findings relate more specifically to associations involving underlying neuropathology associated with AD/ADRD. There was no overt evidence of increased frailty among our amyloid positive individuals; however, formal assessment was not obtained and may be an important consideration for future research. Prior research suggests that while physical frailty and cognitive impairment often overlap in older adults, physical frailty alone does not confer increased risk for cognitive decline or dementia onset over at least a 5‐year period. ²⁹ , ³⁰ , ³¹ Instead, it appears that a combination of physical frailty and cognitive complaints among nondemented older adults increases the risk of developing cognitive decline and dementia, although research does not agree on whether this risk is increased for all types of dementia or only specific diagnoses (e.g., AD, vascular) ³¹ , ³² .

Overall, findings from this study support the inclusion of motor measures in batteries aimed at early risk detection for AD ⁶ , ⁹ , ³² , ³³ , ³⁴ and suggest that the measures from the NIHTB‐MB may be sensitive in capturing early differences to signal risk. Our findings build on prior work looking at NIHTB‐MB among cognitively normal, MCI, and AD adults where cognitively normal adults outperformed those with MCI and AD on measures of walking endurance and fine motor speeded dexterity. ³⁵ Although Gupta and colleagues looked at AD biomarkers in available participants, only a small subset of the cognitively normal adults had available tau or amyloid data; within this small subset, amyloid and tau burdens in three selected brain regions of interest (precentral, parahippocampus, and hippocampus) were not associated with performance on NIHTB walking endurance, grip strength, or fine motor dexterity measures. ³⁵ In our larger sample of cognitively normal adults classified based on participants’ centiloid values, NIHTB motor measures did appear sensitive to amyloid PET and plasma pTau‐217 biomarker elevations prior to the onset of cognitive decline.

At this time, it is not clear whether the observed relationship between Standing Balance and amyloid positivity found in this study is a direct result of amyloid deposition in the brain or whether there may be mediating factors, such as vascular changes that may precede increases in Aβ during the progression of AD ³⁶ and potentially impact muscle strength and function. ³⁷ Even if vascular dysfunction earlier in the Alzheimer's disease pathway is the primary contributor to balance disturbance in early ADRD, the inclusion of Standing Balance in early detection research continues to be important. Future studies may wish to further investigate the relationship of motor changes with various early markers of AD and related dementias (ADRD).

This study is not without its limitations. The study cohort is English speaking, highly educated, and non‐Hispanic White, limiting its generalizability to the general U.S. and world populations. Also, while our study cohort is larger than some previous studies of cognitively normal adults with known biomarker status, it is still small and validation of study findings in larger, more diverse cohorts is necessary to confirm the observed relationship and thus the importance of including motor measures in early detection batteries. As mentioned above, pTau‐217 is one of many blood‐based biomarkers currently under investigation for possible use in early AD risk detection and pTau‐217 blood cutoffs to indicate elevated AD risk are still under development (e.g., ²⁰ ). Future studies may wish to include robust cutoffs for pTau‐217 to categorize participants as “high risk” or “low risk” for AD‐related cognitive decline. Statistically, we did not correct for multiple comparisons in this dataset; however, the degree of significance for our primary findings (e.g., p‐values < 0.001) suggest any such correction would not alter the findings.

Future investigation of our study questions in larger participant cohorts may wish to investigate inclusion of potential interaction effects among commonly occurring medical conditions, like hypertension and diabetes, and AD biomarkers on cognitive and motor functioning performance, or utilize measures of vascular burden from magnetic resonance imaging scans to understand potential interaction/mediation effects. Similarly, larger cohorts may be better powered to investigate potential relationships between cognitive‐motor performance interactions and AD biomarker status. These cognitive–motor performance interactions may be of particular interest to early AD detection work given bidirectional relationships between motor performance, such as gait and balance metrics, and cognitive performance, particularly measures of executive functioning that have been previously demonstrated in our own laboratory as well as by other research groups (e.g., ³⁸ , ³⁹ , ⁴⁰ ). Although there were no cognitive differences seen between the amyloid positive and negative cognitively normal groups in this study, subtle interactions between cognition and motor function may be of importance in early detection work among cognitively normal older adults.

In conclusion, this study sought to determine whether performance on NIHTB cognition and motor measures differentiated older adults with normal cognitive functioning performance with and without increased brain amyloid status. Our findings support previous work suggesting motor performance is negatively associated with AD biomarkers among cognitively normal individuals.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

AUTHORS FINANCIAL DISCLOSURES

All authors are employed by the University of Michigan or Michigan State University and have salary effort supported by research grants from the National Institute of Health (NIH).

CONSENT STATEMENT

All human subjects provided written informed consent.

Supporting information

Supporting Information

Supporting Information

Supporting Information

Cook Maher A, Rose S, DuBois KN, et al. Motor, not cognitive, performance relates to amyloid status in normal older adults. Alzheimer's Dement. 2026;18:e70237. 10.1002/dad2.70237