Sensory and motor deficits as contributors to early cognitive impairment

Abstract

INTRODUCTION

Age‐related sensory and motor impairment are associated with risk of dementia. No study has examined the joint associations of multiple sensory and motor measures on prevalence of early cognitive impairment (ECI).

METHODS

Six hundred fifty participants in the Baltimore Longitudinal Study of Aging completed sensory and motor function tests. The association between sensory and motor function and ECI was examined using structural equation modeling with three latent factors corresponding to multisensory, fine motor, and gross motor function.

RESULTS

The multisensory, fine, and gross motor factors were all correlated (r = 0.74 to 0.81). The odds of ECI were lower for each additional unit improvement in the multisensory (32%), fine motor (30%), and gross motor factors (12%).

DISCUSSION

The relationship between sensory and motor impairment and emerging cognitive impairment may guide future intervention studies aimed at preventing and/or treating ECI.

Highlights

• Sensorimotor function and early cognitive impairment (ECI) prevalence were assessed via structural equation modeling.

• The degree of fine and gross motor function is associated with indicators of ECI.

• The degree of multisensory impairment is also associated with indicators of ECI.

1. BACKGROUND

Age‐related sensory and motor impairment are associated with elevated risk of mild cognitive impairment (MCI) and Alzheimer's disease and related dementias (ADRD). A growing body of literature has shown decline in sensory function of vision, ¹ olfaction, ² , ³ , ⁴ taste, ⁵ hearing, ⁶ , ⁷ , ⁸ , ⁹ and balance ¹⁰ , ¹¹ as risk factors for poor cognitive outcomes. Declines in motor function manifest as slower gait speed and greater lap time variability and have been linked with increased risk of cognitive impairment. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ These sensory and motor impairments arise up to 10 years prior to onset of cognitive impairment, ¹ , ⁷ , ⁸ , ²⁰ , ²¹ , ²² , ²³ and thus may serve as key non‐invasive preclinical indicators of MCI onset and progression to ADRD, with potentially meaningful public health implications for screening and early intervention.

To date, most studies have focused on the independent associations of sensory or motor impairment on risk of cognitive impairment. Recent findings indicate both a high prevalence of concurrent sensory and motor impairments in older adults; those with multiple sensory and motor impairments have elevated risk of cognitive impairment relative to those with an isolated impairment in sensory or motor function. ²⁴ , ²⁵ , ²⁶ , ²⁷ Prior to formal diagnosis of MCI, individuals often report changes in cognition below the detectable threshold of standard cognitive testing. This preclinical phase of cognitive decline is associated with normal functional ability and can precede a formal diagnosis of dementia by up to 20 years. ²⁸ , ²⁹ To capture this preclinical phase, a classification algorithm was recently developed to identify individuals with apparently normal cognition who subsequently progressed to MCI. ³⁰ Previous work has shown the application and validation of this tool to identify community resident older adults with early cognitive impairment (ECI). ³¹ Further, we investigated the association of multiple sensory impairments including hearing, vision, vestibular, proprioception, and olfaction functioning with risk of ECI and found that older adults with multiple sensory impairments had triple the risk of developing ECI over up to 4 years of follow‐up. ³²

The current study aims to extend this work by systematically assessing the joint associations of multiple sensory and motor impairments with prevalence of ECI in older adults participating in the Baltimore Longitudinal Study of Aging (BLSA). We employed structural equation modeling (SEM) to measure these constructs using several available indicators in a cross‐sectional study. We hypothesized that the prevalence of ECI is associated with sensory and motor impairment as assessed by multisensory, fine motor, and gross motor function.

2. METHODS

2.1. Study participants

The BLSA was established in 1958 and is conducted by the National Institute on Aging (NIA) Intramural Research Program. This longitudinal study continuously enrolls community‐dwelling volunteers, aiming to explore the interdependence of aging and disease processes on physical and cognitive performance. ³³ Participants are free of major chronic conditions and cognitive and functional impairment at time of enrollment. Once enrolled, participants are followed every 1 to 4 years for life (every 4 years if aged <60, every 2 years if 60 to 79, and annually if ≥80). BLSA participants aged 50 years and older who completed cognitive, sensory function, and motor function assessments from 2012 to 2018 were included in the analysis (Figure S1). In the main analysis, we excluded participants with stroke, Parkinson's disease, MCI/ADRD, and cognitive impairment (CI) without dementia. ³⁴ , ³⁵ , ³⁶ Participants with MCI/ADRD were included in the sensitivity analysis. The study protocol was approved by the Institutional Review Board of the Intramural Research Program of the National Institutes of Health. Informed consent was obtained from all participants at each study visit.

2.2. Measurements

2.2.1. Algorithmic classification of ECI

We have previously used the BLSA cohort to create and validate a measure of ECI that incorporates demographically adjusted cutpoints of the Card Rotations Test and the California Verbal Learning Test (CVLT) immediate recall to operationalize impairment. ³⁰ , ³² The Card Rotations test assesses visuospatial ability ³⁷ ; the score is defined as the difference between correctly and incorrectly classified items. Verbal memory was measured using the sum of immediate free recall trials from the CVLT. ³⁸ Poor cognitive performance on these two tests was operationalized as one standard deviation (SD) below age‐, sex‐, race‐, and education‐specific means in the BLSA data.

2.2.2. Multisensory function

Assessment of hearing, vision, vestibular, proprioception, and olfaction function were considered independently and jointly.

2.2.2.1. Hearing

Hearing data were collected by pure tone audiometry, a measure of peripheral auditory system sensitivity. Pure tone air conduction thresholds in each ear were obtained in decibels hearing level (db HL) at standard octaves from 0.5 to 8 kHz using insert earphones (EARTone 3a; 3 M, St. Paul, Minnesota) and an Interacoustics AD629 audiometer (Interacoustics A/S, Assens, Denmark) by a trained technician in a sound‐attenuating booth meeting American National Standard Institute standards. ³⁹ A speech frequency pure‐tone average (PTA) at thresholds 0.5, 1, 2, and 4 kHz was calculated for each ear, and hearing impairment was defined by PTA > 25 dB HL in the better‐hearing ear consistent with American Speech‐Language‐Hearing Association guidelines. ⁴⁰

2.2.2.2. Vision

A composite visual impairment variable was defined as having impairment on any of the following tests: visual acuity, visual fields, contrast sensitivity, and stereo acuity.

Visual acuity was measured unilaterally while participants wore corrective lenses (if applicable) using Early Treatment of Diabetic Retinopathy Study (ETDRS) eye charts (logMAR units). ⁴¹ Presenting better‐eye visual acuity was based on the ETDRS chart line where the participant could correctly read at least three out of five letters. Impaired visual acuity was defined as presenting visual acuity worse than 20/40 based on criteria commonly used in research studies and clinical settings. ⁴² , ⁴³

Contrast sensitivity was measured using a Pelli‐Robson chart (logContrast units) bilaterally with participants wearing corrective lenses (if applicable). ⁴⁴ Contrast sensitivity impairment was defined as 1 SD below the average (logContrast units < 1.55) of previous population‐based studies of adults ≥60 years. ⁴⁵

Visual fields were measured using a Humphrey 81‐point single intensity (24 dB) full field (60°) analyzer (Humphrey Field Analyzer, Carl Zeiss Meditec, Dublin, CA). Monocular visual fields were measured, and binocular visual fields were estimated using a previously described method. ⁴⁶ The composite binocular visual fields were scored as number of points missed (out of 96 points) (central field [56 points], upper peripheral field [18 points], lower peripheral field [22 points]). Impairment was defined as 1 SD below the population mean of total points missed in this cohort.

Stereo acuity was measured using the Randot Stereo vision test, which assesses depth perception. Impairment was defined as an inability to discern a depth differential of <80 seconds of arc. ⁴⁷

2.2.2.3. Vestibular function

Vestibular function impairment was defined as impairment in saccular or semicircular canal function. Saccular function was measured using cervical vestibular‐evoked myogenic potentials (cVEMP). ⁴⁸ In the air‐conducted cVEMP test, individuals with no response above 100 dB acoustic clicks were considered to have absent function in the tested ear. Impaired saccular function was defined by absent cVEMP bilaterally. ⁴⁹ Semicircular canal function was measured using video head‐impulse testing. Vestibulo‐ocular reflex gain was calculated ⁵⁰ as the ratio of mean eye velocity over mean head velocity during a 40‐millisecond window centered at peak head acceleration. Impaired semicircular canal function was defined as a mean vestibulo‐ocular reflex gain of <0.7. ⁵¹

RESEARCH IN CONTEXT

1. Systematic review: Age‐related sensory and motor impairment are associated with elevated risk of mild cognitive impairment and Alzheimer's disease and related dementias. A growing body of literature has shown decline in sensory and motor function as risk factors for poor cognitive outcomes, but no study has examined the joint effects of multiple sensory and motor measures on prevalence of early cognitive impairment (ECI).

2. Interpretation: In this study, the association between sensory and motor function and ECI was examined using structural equation modeling with three latent factors corresponding to multisensory, fine motor, and gross motor function. The multisensory, fine, and gross motor factors were correlated, and the odds of ECI were lower for each additional unit in the multisensory, fine motor, and gross motor factors.

3. Future directions: The relationship between more severe sensory and motor impairment and emerging cognitive impairment may guide future intervention studies aimed at preventing and/or treating ECI.

2.2.2.4. Proprioception

Proprioception was measured using the threshold for perception of passive movement (TPPM). ⁵² , ⁵³ Four trials were performed in a set sequence of plantarflexion, dorsiflexion, dorsiflexion, and plantarflexion. TPPM was the average of the best plantarflexion and dorsiflexion. Impairment was defined as TPPM > 2.2°, based on previously established thresholds in older adults. ⁵²

2.2.2.5. Olfaction

Olfaction was tested from the 16‐item Sniffin’ Sticks Odor Identification Test. ⁵⁴ Two versions of this test were administered at random at the initial assessment, and then alternated during the follow‐up visits. ³⁶ Given the lack of established thresholds for impairment in the literature, we defined impaired olfaction as scores eight or seven on test A or B, respectively, which were at the 11th percentile on each test version for the current sample data. ²⁴

2.2.3. Motor performance

Motor and physical function measures were considered based on prior work identifying these measures as reliable tools for evaluating functional mobility in older adults. ⁵⁵

2.2.3.1. Fine motor performance

Visuomotor integration and manual dexterity were assessed with the Purdue Pegboard test. ⁵⁶ Participants picked up pegs and placed them sequentially into small holes on the board as quickly as possible in 30 seconds. The number of pegs placed was recorded over two trials per hand and averaged for the dominant and non‐dominant hand separately.

2.2.3.2. Gross motor performance

Physical functioning was measured using the expanded short physical performance battery (ExSPPB) ⁵⁷ which consists of full‐tandem and single‐leg stance standing balance tests, each held for 30 seconds, 10 chair stands done as quickly as possible, and usual gait speed (m/s) over 6 m. Ability and time to hold each standing balance pose was recorded. The rate for repeated chair stands (chair stands/second) was calculated. Usual gait speed was measured using the faster of two trials.

The long‐distance corridor walk (LDCW) test measured endurance walking performance and variability. ⁵⁷ Participants completed a 2.5‐minute warm‐up walk at their usual comfortable pace followed by a 400‐m fast‐paced walk (walking back and forth along a 20‐m course for ten 40‐m laps). Time to complete each lap was recorded. Lap time variation was defined as the detrended SD of residuals of lap time. ⁵⁸ Specifically, lap times were regressed on the number of laps using a linear mixed effects model, and the SD of the model residuals was calculated.

Hand grip strength was measured using a Jamar hydraulic hand dynamometer (kg). ⁵⁹ Three trials were taken per hand; the best measurement from the left or right hand was used in the analysis.

2.3. Covariates

Sociodemographic characteristics including age, sex, race, and education in years were reported by participants in the health interview. The body mass index (BMI) was calculated as kilograms per meter squared (kg/m²). Self‐reported chronic conditions included cardiovascular disease, diabetes, hypertension, hyperlipidemia, osteoarthritis, lung disease, and cancer. Diabetes can lead to visual impairment and mobility limitations and thus was examined separately. ⁵⁵ The number of comorbidities except diabetes were classified as less than two versus two or more conditions.

2.4. Statistical analysis

Data distributions were checked for outliers for all variables. The means and SDs or frequencies and percentages of demographics, health characteristics, and sensorimotor function tests were examined. Independent t‐tests or chi‐square tests were used to examine differences in characteristics or functional tests by ECI status.

SEM methods facilitate the representation of constructs as latent variables based upon the shared covariation among indicators that form a measurement model for a given construct; such methods have less measurement error than summative scores. ⁵⁵ The association between sensory and motor functioning and ECI was tested using SEM methods. Exploratory factor analysis was conducted; three latent factors corresponding to multisensory, fine motor, and gross motor function were identified. Measurement models were estimated for these factors using confirmatory factor analysis (CFA). The multisensory latent variable was measured by five binary indicators: normal versus impaired vision, vestibular function, hearing, proprioception, and olfaction. A higher value indicated greater sensory function. Fine motor function was measured by two indicators from the Purdue Pegboard test corresponding to the dominant and non‐dominant hands. Gross motor function was measured by six indicators including binary indicators for holding each pose of a full tandem and single leg stance for 30 seconds, and continuous indicators for grip strength, rate of repeated chair stands, usual gait speed, and lap time variation with sign changed to be consistent with the polarity of other motor function variables. For both the fine and gross motor latent variables, a higher value indicates better motor performance. CFAs were estimated using weighted least squares estimation with theta parameterization. The model fit was evaluated by global model fit indices including comparative fit index (CFI), Tucker‐Lewis fit index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). Criteria for excellent model fit included CFI of ≥0.95, TLI of ≥0.95, RMSEA of ≤0.05, and SRMR of ≤0.08. ⁶⁰

Once the fit of the measurement models for each latent variable was acceptable, the three latent factors for multisensory functioning, fine motor, and gross motor were entered as exogenous variables with ECI as an endogenous outcome. This full model was adjusted by age, sex, race, education, and BMI. Additional covariates for adjustment included diabetes and other comorbidities. Regressions of ECI on latent factors were converted to odds ratios (ORs) from probit regression coefficients for a binary outcome by multiplying the coefficient by a scaling factor (1.7) and exponentiating it. ⁶¹

The significance level α was set as 0.05. Descriptive statistical analyses were conducted using R version 3.6.2 (R Foundation). SEM analysis was performed using Mplus 8.4 (Muthén & Muthén, Los Angeles, CA).

3. RESULTS

The final analytic sample consisted of 650 participants (Figure S1). The mean age was 73.8 ± 10.3 years. The participants were 57.1% female and 25.5% Black. On average, participants had 17.6 ± 2.5 years of education. Sociodemographic characteristics, health conditions, physical functioning tests and impairment status on multiple sensory functions in the total sample and by ECI status are presented in Table 1. In this BLSA cohort, participants who met the criteria for ECI were more likely to be younger and trended towards having higher rates of diabetes. Without adjusting for demographics and comorbidities, individual sensory and motor function indicators did not significantly differ between those with and without ECI (Table 1).

TABLE 1.

Demographic, health, and sensorimotor factors in middle‐aged and older adults in the Baltimore Longitudinal Study of Aging.

			Mean ± SD or N (%)
Characteristics	N	Total	ECI (Yes) (N = 153)	ECI (No) (N = 497)	p‐value
Age (years)	650	73.8 ± 10.3	71.7 ± 10.4	74.4 ± 10.1	0.006
Female	650	371 (57.1)	93 (60.8)	278 (55.9)	0.33
Race	650				0.74
White		438 (67.4)	98 (64.1)	340 (68.4)
Black		166 (25.5)	45 (29.4)	121 (24.3)
Other		46 (7.1)	10 (6.5)	36 (7.2)
Education (years)	648	17.6 ± 2.5	17.7 ± 2.6	17.5 ± 2.5	0.37
BMI (kg/m2)	649	27.3 ± 4.9	27.5 ± 5.1	27.3 ± 4.8	0.75
Comorbidities
Cardiovascular disease	650	54 (8.3)	10 (6.5)	44 (8.6)	0.46
Diabetes	650	93 (14.3)	29 (19.0)	64 (12.9)	0.08
Hypertension	650	301 (46.3)	64 (41.8)	237 (47.7)	0.24
Hyperlipidemia	650	357 (54.9)	87 (56.9)	270 (54.3)	0.65
Osteoarthritis	650	330 (50.0)	75 (49.0)	255 (51.3)	0.69
Lung disease	650	71 (10.9)	20 (13.1)	51 (10.3)	0.41
Cancer	650	207 (31.8)	44 (28.8)	163 (32.8)	0.40
Pegboard test (# of pegs)
Dominant hand	639	12.5 ± 2.1	12.3 ± 2.2	12.5 ± 2.1	0.40
Non‐dominant hand	639	12.0 ± 2.1	12.0 ± 2.1	12.0 ± 2.0	0.94
Grip strength (kg)	637	30.3 ± 10.7	30.0 ± 11.2	30.4 ± 10.6	0.71
Full‐tandem stand time	650	27.6 ± 7.3	27.8 ± 6.9	27.5 ± 7.4	0.69
Did not hold for 30 s		75 (11.5)	16 (10.5)	59 (11.9)	0.74
Single leg stand time	650	20.2 ± 12.1	21.1 ± 12.0	19.9 ± 12.1	0.28
Did not hold for 30 s		296 (45.5)	64 (41.8)	232 (46.7)	0.34
Chair stand pace (stands/s)	648	0.55 ± 0.19	0.57 ± 0.19	0.55 ± 0.18	0.22
400‐m endurance walk
Lap time variation	616	0.90 ± 0.43	0.92 ± 0.44	0.90 ± 0.43	0.56
Usual gait speed (m/s)	649	1.15 ± 0.22	1.16 ± 0.25	1.15 ± 0.22	0.50
Hearing impairment	593	288 (48.6)	62 (44.6)	226 (49.8)	0.33
Visual impairment	552	207 (37.5)	51 (38.1)	156 (37.3)	0.96
Vestibular impairment	412	87 (21.1)	19 (20.7)	68 (21.3)	1.0
Proprioception impairment	648	90 (13.9)	35 (19.4)	65 (13.1)	0.36
Olfaction impairment	490	75 (15.3)	21 (18.8)	54 (14.3)	0.32

Abbreviations: BMI, body mass index; ECI, early cognitive impairment.

In the CFA model (Model 1), the measurement model consisting of three latent factors of multisensory, fine, and gross motor function exhibited a good model fit (RMSEA = 0.04, CFI = 0.97, TLI = 0.963, SRMR = 0.055; Figure 1). Factor loadings for indicators across various domains ranged from 0.37 to 0.89. The multisensory factor was correlated with the fine motor and gross motor factors at r = 0.77 and r = 0.81, respectively. The fine motor and gross motor factors were correlated at r = 0.74.

FIGURE 1.

CFA model (Model 1) with three latent factors of fine motor, gross motor, and multisensory function: results from BLSA (N = 650) participants. Standardized factor loadings are shown on paths between latent variables and their respective indicators. Curved, double‐headed arrows indicate latent variable correlations. Model fit indices are as follows: CFI = 0.97, TLI = 0.963, RMSEA = 0.04, SRMR = 0.055. BLSA, Baltimore Longitudinal Study of Aging; CFA, confirmatory factor analysis; CFI, comparative fit index; SRMR, standardized root mean square residual; TLI, Tucker‐Lewis fit index.

In the full SEM model (Model 2, Figure 2), after adjusting for age, sex, race, education, and BMI, for each 1 SD higher in the multisensory, fine motor, and gross motor latent factors, the odds of ECI were 32% (95% CI: 12% to 47%), 30% (95% CI: 15% to 43%) and 12% (95% CI: −2% to 24%) lower, respectively. Standardized factor loadings in Figure 2 are relatively homogeneous across indicators for each factor (eg, 0.76 to 0.77 for fine motor, 0.55 to 0.89 for gross motor, and 0.41 to 0.60 for multisensory impairment), suggesting no indicator or smaller set of indicators are driving the main results. In addition to including covariates in Model 2, after further adjustment for diabetes and number of comorbidities, for each 1 SD higher in the multisensory, fine motor, and gross motor latent factors, the odds of ECI were 32% (95% CI: 13% to 46%), 29% (95% CI: 14% to 42%) and 10% (95% CI: −4% to 22%) lower, respectively (Model 3, Figure S2).

FIGURE 2.

SEM model (Model 2) of associations between early cognitive impairment (ECI) and fine motor function, gross motor function, and multisensory function: results from BLSA (N = 650) participants. Regressions of ECI on variables of interest are adjusted for age, sex, race, education in years, and BMI. Standardized factor loadings are shown on paths between latent variables and their respective indicators. Odds ratios (95% confidence intervals) are shown on the paths between latent variables and ECI. BLSA, Baltimore Longitudinal Study of Aging; BMI, body mass index; ECI, early cognitive impairment; SEM, structural equation model.

4. DISCUSSION

This study found that while individual sensorimotor indicators did not significantly differ between those with and without ECI, multisensory function and fine motor function latent factors are significantly associated with ECI. Gross motor function, on the other hand, did not reach statistical significance. Whereas some impairments in sensory and motor function may themselves represent indicators of neurodegeneration associated with ECI, the impact of systemic and institutional issues that lead to gaps in environmental accessibility and discrimination against those with disabilities should also be considered. These associations may also be related to the onset of cognitive impairment through decreased environmental perception and interaction and thus serve as modifiable targets for the prevention of ADRD. This is particularly important since many known ADRD risk factors—age, family history, and genetic susceptibility genes—are non‐modifiable. ⁶² In addition, most therapies are currently administered later in the ADRD neuropathologic process, lowering the potential to substantially alter clinical outcomes. ²⁰ , ⁶³ To this end, treatment of early adverse changes in sensory and motor function may provide promising alternatives to alter the course of ADRD progression. Moreover, early identification of cognitive changes via detection of sensory and motor impairment may allow for more effective disease modification and/or stabilization of cognitive function. Additionally, systemic and institutional changes must be implemented to support the inclusion and integration of those with impairments into society as a whole.

We used SEM to examine the influence of multiple sensory indicators for multisensory impairment and various performance tests for fine and gross motor function on the prevalence of ECI. This study confirms and extends prior work showing that multiple sensory impairments are associated with cognitive dysfunction. ³² Vision, hearing, olfactory, vestibular, and proprioceptive impairments have each been independently associated with cognitive impairment and, more recently, in studies evaluating the link between multisensory impairments and cognitive impairment. ¹ , ² , ³ , ⁴ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Brenowitz et al. found that the number of sensory impairments was incrementally associated with dementia risk; specifically, participants with 1, 2, or ≥3 impairments had greater risk of developing incident dementia in the subsequent 10 years. ²⁶ Prior work has also shown that each additional sensory impairment was associated with a greater risk of having ECI, suggesting a synergistic contribution to cognitive impairment. ⁶⁴ Collectively, this evidence suggests that sensory impairments have at least additive associations with cognitive impairment, although the mechanisms by which each sensory system is related to cognition may differ, with some sensory systems functioning as indicators of neurodegeneration while others represent risk factors for cognitive impairment. ⁶⁵ , ⁶⁶

By using a data‐driven factor analytic approach, we examined domains of fine and gross motor function and found that fine motor function was significantly associated with ECI whereas gross motor function did not reach statistical significance. Our findings are consistent with fine motor function being tightly intertwined with cognitive and neurological function. ⁶⁷ , ⁶⁸ , ⁶⁹ The Purdue Pegboard test used to operationalize manual dexterity requires visual‐motor perceptual integration as well as musculoskeletal function of the upper extremities to accurately grasp and manipulate small objects. This test also serves as a measure of neurological function and has been related to cognitive planning ability. ⁵⁶ , ⁷⁰ , ⁷¹ Recent work from our group has shown that poor manual dexterity on the Pegboard test was the strongest clinical predictor of slow gait speed. ³² Taken together, these findings suggest that impairment in fine motor performance and sensorimotor integration is a strong correlate of both ECI and mobility limitation and may play an important role in predicting progression to clinically detectable cognitive as well as physical impairment.

The gross motor function domain, comprised of measures of standing balance, strength, gait speed, and gait variability, was not significantly associated with ECI. Although this difference may simply be due to insufficient sample size, the possibility of a temporal relationship between gross motor function and more severe cognitive impairment cannot be excluded. Previous findings suggest that AD pathology affects common pathways involved in both cognition and gross motor function (ie, motor changes may represent markers of AD pathology), ¹⁶ , ⁷² and amyloid‐β (Aβ) burden has been linked with gait speed and variability in cognitively normal adults, suggesting that slowed and altered gait patterns may be sensitive markers of AD pathology. ⁷³ , ⁷⁴

The current findings of independent associations of multisensory impairment and motor function with ECI suggest that each domain has an independent link with cognitive impairment. Whether this is due to distinct causal mechanisms or from shared neuropathology of discrete brain regions remains to be determined. Additional studies are needed to investigate associations among impairments in sensory and motor function and AD pathology (ie, Aβ and tau deposition) in the preclinical phase, which may shed light on potential pathogenic mechanisms. Further investigation into the role of modifiable environmental and societal factors on cognitive function is also paramount.

This study has limitations. First, these findings may not be generalizable to the broader United States population as the BLSA recruits a healthy cohort and has high educational attainment. However, the healthy nature of the cohort helps minimize confounding by disease burden. Second, the cross‐sectional nature of the data prevents relationships between sensory and motor functions and cognitive impairment to be causally linked. While the classification of ECI was determined by poor cognitive performance on the Cards Rotation Test and CVLT, performance on these tests may be affected by overlapping sensorimotor domains such as hearing and vision impairment. However, prior work has shown that while performance on cognitive testing may be affected by sensorimotor impairments, the overall measurement of cognitive function is not significantly affected by these administration characteristics. ⁷⁵ Further, the potential bidirectional nature between sensory and motor impairment and cognitive dysfunction should be further elucidated (eg, whether sensorimotor impairments contribute to loss of cognitive function or loss of cognitive function contributes to the development of sensorimotor impairments). Lastly, we attempted to minimize the effects of confounding factors by adjusting for possible predictors of cognitive impairment including age, sex, race, level of education, BMI, diabetes, and comorbidities, but residual confounders may persist. For example, a common pathologic etiology affecting brain structure and function may underlie both sensorimotor and cognitive function. Future longitudinal studies can help to distinguish whether the associations found in this study can serve as early markers or indicators of an upstream cause.

In summary, our study shows that sensory and motor function as measured by multiple sensory and fine motor function are significantly associated with ECI in a cohort of older adults who do not meet standard criteria for cognitive impairment. Gross motor function had a similar protective association with ECI, although this did not reach statistical significance. The primary goal of our work is to guide future research studies that are aimed to prevent or intervene on cognitive decline at an early stage through modifiable mechanisms, allowing older adults to maintain independence and overall quality of life. Future refinement of these tests is needed to guide clinical decision‐making.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare. Author disclosures are available in the supporting information.

CONSENT STATEMENT

The study protocol was approved by the Institutional Review Board of the Intramural Research Program of the National Institutes of Health. Informed consent was obtained from all participants at each study visit.

Supporting information

Supporting information

Supporting information

Sayyid ZN, Wang H, Cai Y, et al. Sensory and motor deficits as contributors to early cognitive impairment. Alzheimer's Dement. 2024;20:2653–2661. 10.1002/alz.13715