Association of Age with Dual-Task Objective Cognitive Indicators and Gait Parameters in Older Adults

Abstract

Background:

Early recognition of dementia like Alzheimer’s disease is crucial for disease diagnosis and treatment, and existing objective tools for early screening of cognitive impairment are limited.

Objective:

To investigate age-related behavioral indicators of dual-task cognitive performance and gait parameters and to explore potential objective markers of early cognitive decline.

Methods:

The community-based cognitive screening data was analyzed. Hierarchical cluster analysis and Pearson correlation analysis were performed on the 9-item subjective cognitive decline (SCD-9) scores, walking-cognitive dual-task performance, walking speed, and gait parameters of 152 participants. The significant differences of indicators that may related to cognitive decline were statistically analyzed across six age groups. A mathematical model with age as the independent variable and motor cognition composite score as the dependent variable was established to observe the trend of motor cognition dual-task performance with age.

Results:

Strong correlation was found between motor cognitive scores and SCD and age. Gait parameters like the mean value of ankle angle, the left-right difference rate of ankle angle and knee angle and the coefficient of variation of gait cycle showed an excellent correlation with age. Motor cognition scores showed a decreasing trend with age. The slope of motor cognition scores with age after 50 years (k = –1.06) was six times higher than that before 50 years (k = –0.18).

Conclusions:

Cognitive performance and gait parameters in the walking-cognitive dual-task state are promising objective markers that could characterize age-related cognitive decline.

Keywords

Age Alzheimer’s disease cognitive decline motor cognitive assessment objective markers walking-cognitive dual task

INTRODUCTION

Cognitive decline in older adults is a critical issue. There were more than 11 million caregivers for people with Alzheimer’s disease (AD) and other dementias in the United States in 2021, who contributed about 16 billion hours of care [1]. Mild cognitive impairment (MCI), as a precursor to AD, is defined as abnormal cognitive decline exceeding normal aging expectations [2]. A primary care study found that the undiagnosed rate of cognitive impairment in the United States was 20.8% in the elderly population over 55 [3]. A majority of patients do not receive adequate guidance and treatment before the onset of clinical symptoms, and more validated early screening tools need to be developed and investigated.

Clinical screening tools that efficiently and objectively assess cognitive function are currently lacking. Existing technologies for detecting dementia biomarkers [2, 4], such as imaging and cerebrospinal fluid testing, require relatively significant symptoms to be presented to be detected accurately, and are costly. The methodology of scaled cognitive assessment [5] is subjective and has limited accuracy. Neuropsychological testing methods [5, 6] take hours to complete, with complex and inefficient questions, making them unsuitable as community cognitive screening instruments. Large-scale screening for cognitive impairment is limited. More instruments and objective markers that enable efficient and rapid detection are needed.

Assessment combining motor and cognitive function is considered promising for cognitive screening [7, 8]. Evidence has shown that low motor function is associated with declines in specific cognitive domains including global cognitive function, situational memory, working memory, and so on [8]. Study showed that sensory and motor measures improve risk prediction models of cognitive decline as well as cognitive impairment [9]. Gait and/or balance disturbances are proved to be associated with a higher risk of developing AD [10]. Dual-task (DT) gait test is a classic cognitive-motor interference test in which cognitive task and gait are executed concurrently and measured separately to appraise the interaction between gait and cognition [11]. Gait test, as an easy to use and cost-effective tool that provides valuable insights into an individual’s neurological well-being before more apparent cognitive decline symptoms manifest, is a potential screening tool for MCI in a large-scale population. It has been found that some of the evaluated DT gait parameters significantly correlate with cognitive impairment over time. A study suggested that gait velocity in animal picture naming dual task could act as a behavioral marker to detect MCI [12]. According to a multisite research of 500 older persons across neurodegenerative conditions, increased gait variability was proved to be associated with dementia, suggesting that high gait variability is a marker of cognitive cortical dysfunction [13]. Numbers of words during DT has shown to be one of the promising DT parameters that associated with cognitive impairment prediction [14]. This evidence underscores the value of gait assessments as a potential biomarker for predicting transitions from MCI to dementia. Basic gait spatiotemporal parameters, such as speed, step length, swing and stance time, can be assessed visually or quantitatively. However, more complex gait parameters such as gait variability and gait symmetry are rarely investigated due to insufficient application of technical equipment [15, 16]. Gait has an intricate correspondence with different aspects of cognitive function, as it characterizes multiple motor performances [17]. It is hard to comprehensively assess cognitive function using a single characteristic of walking speed [18, 19]. More diverse gait indicators that characterize cognitive decline need to be evaluated.

Cognitive problems emerge with age. Cognitive abilities such as memory, language function and thinking deteriorate with age [1]. Research studied cognitive estimation abilities across the lifespan using a digital cognitive assessment tool and found a tendency to decrease visual and auditory estimation abilities with age [20]. One’s biological age marks on one’s brain during a healthy person’s lifetime. A study predicted brain age using MRI imaging data by comparing the severity of brain atrophy and functional connectivity, with a predicted error of 7.08 years from chronological age [21]. Researchers can compute the predicted age of new samples by using age prediction models. When predicted age is greater than the chronological age and the difference is too significant, it indicates that the sample is aging too fast [22]. The approach of establishing a connection between potential indicators and age to explore cognitive-related objective markers is gradually recognized.

Subjective cognitive decline (SCD) refers to an individual’s complaint of memory or other aspects of cognitive decline and is considered an earlier or preclinical stage of MCI [5, 23]. Gifford et al. [24] developed the 9-item subjective cognitive decline (SCD-9) rapid screening scale based on a series of psychometric models reliably extracted from a large pool of questions about SCD. The SCD-9 scale will be used as a subjective counterpart for cognitive assessment in this study.

Based on the theory that walking-cognitive dual task performance is related to cognitive function, coupled with the fact that cognitive function declines naturally with age, the present study aims to conduct dual task tests in people of different ages to observe the correlation between objective indicators and age. Indicators highly correlated with age are considered valid predictors of cognitive decline, providing a theoretical basis for developing cognitive screening tools. The specific work of this study is as follows. First, the correlation of walking-cognitive dual-task performance indicators, cognitive performance, and gait parameters with age will be analyzed considering SCD, chronic disease conditions, and education level, to identify objective indicators underlying age-related cognitive decline. Second, a mathematical model with age as the independent variable and dual-task performance indicators as the dependent variable will be developed to observe the trend of objective cognitive indicators with increasing age. This study proposes two hypotheses. First, cognitive task performance, motor performance, and gait parameters during walking-cognitive dual tasks would show correlations with age; the higher of the age, the worse the performance. Objective cognitive indicators in the dual-task condition may also be correlated with SCD, chronic disease, education level, and other cognitive-related factors. Second, walking-cognitive dual-task performance will decline with age, and a turning point will exist, after which the rate of cognitive decline will accelerate.

MATERIALS AND METHODS

Subjects

Demographic information, chronic diseases, cognitive function, motor performance, and a combined motor-cognitive evaluation of 152 participants of different ages were analyzed using data obtained from the 2023 community-based cognitive function screening of older adults at Institute of Biomedical Engineering, Chinese Academy of Medical Sciences. The dataset was divided into six groups according to age (22–34, 35–44, 45–54, 55–64, 65–74, 75–84). All subjects are able to walk normally without assistance for at least 60 meters; corrected eyesight at 4.0 and above; no color blindness or color weakness; no less than 8 years of education; capable of living independently; able to complete the dual-task test; exclusion of diagnosed psychiatric illnesses. All subjects provided their written, informed consent to participate in the study, which was approved by the Biomedical Ethics Committee of Hebei University of Technology (HEBUThMEC2023021).

Demographic information is shown in Table 1.

Table 1

Basic information of subjects

Age Group	22–34	35–44	45–54	55–64	65–74	75–84	p (> 0.05)
Total Number	152
Number	15	26	23	34	42	12	–
Female	9	9	10	22	28	7	–
Female proportion (%)	60	35	43	65	67	58	–
Height (Mean±SD)	170.2±7.29	169.58±7.97	170.22±8.33	164.53±6.92	164.79±7.16	163±7.01	0.0015
Weight (Mean±SD)	65.6±18.94	70.73±12.23	71.35±11.33	69.03±11.67	62.95±10.77	63.5±11.95	0.0463
BMI (Mean±SD)	23.79±3.54	23.10±3.20	25.48±3.99	24.52±2.68	24.48±3.15	22.66±6.51	0.0736
Educational Level (number)
Scholar and above	9	14	8	1	4	4	–
three-year college	5	8	9	10	4	1
Senior high school	0	1	4	5	13	4
Junior and below	1	1	2	18	21	3

Equipment

Motor cognitive dual-task evaluation

The equipment used for motor cognitive assessment is the digital motor cognitive function assessment and training system based on human-computer interaction (developed by Institute of Biomedical Engineering, Chinese Academy of Medical Sciences, hereinafter refers as motor cognitive assessment system). The system (Fig. 1) consists of a treadmill, three high-definition (HD) cameras placed on the left, right, and front sides of the treadmill, a control host computer, an HD display placed in front of the walking direction of the treadmill, and a voice processing module assembled on the treadmill. The system recognizes human joint point coordinates using the unlabeled human key point recognition technique based on machine vision, which shapes the gait information containing time information by connecting the key point data of each frame in the extracted video.

Fig. 1

Components of motor cognitive assessment system.

The system displays cognitive tasks on the HD display (multiple tasks are supported including Stroop color-word test, digit-symbol conversion, etc.). Users perform the cognitive tasks through voice interaction to finish the single cognitive task, walk on the treadmill to finish the single walking task, and complete a walking-cognitive dual-task test by answering cognitive questions while walking on the treadmill. The gait parameters, cognitive function, motor performance, and motor cognition composite scores obtained from the testing of this system will be used to analyze this study.

Evaluation indicators

The indicators that will be analyzed in this study are demographic indicators including age and education, physiological indicators, including BMI, chronic disease status and walking speed, SCD-9 score, cognitive task score, dual-task cognitive performance, and gait parameters. The indicators and corresponding acquisition methods are shown in Table 2.

SCD-9 scores and chronic disease status were collected via questionnaires. Walking speed was obtained using 6-meter walking test. Dual-task test outcomes were captured using the motor cognitive assessment system, and Stroop color-word test was selected for the cognitive task. Participants’ cognitive task scores during the walking-cognitive dual-task test were recorded as dynamic cognitive scores. The motor cognitive assessment system obtained the motor cognition composite score which ranges from 0 to 100. When walking on the treadmill, the tester can hold on to the treadmill handrail if he/she is maladaptive to the autonomic walking condition. The recorder will record the indicator of the need to hold on to the handrail. The gait parameters were obtained by the motion evaluation portion of the system, derived from the gait information of a 10-s walking. In this study, the fundamental gait parameter indicators, such as gait cycle and lower limb joint activity angle, will be further analyzed, and the selected gait parameter indicators are shown in Table 3.

Table 2

Indicators and their acquisition methods

Indicators	Short forms	Acquisition methods
Age	Age	Basic information collection
BMI	BMI
Educational Level	Edu
SCD-9 scores	SCD_9	Questionnaires
Num of Chronic Disease	Num_chronicDisease
Walking Speed	WalkingSpeed	6-Meter Walking Test
Dual-task Cognitive Scores while walking	Dynamic_cogScore	Motor cognitive assessment system
Need to hold the handrail or not when walking	NeedAssist
Motor cognitive composite scores	Score_overall
Gait parameters	(Listed in Table 3)

Table 3

Gait parameters and their short forms

Projects	Indicators	Short forms
Gait cycle	Mean Value	gaitCycle_avg
	Left-right difference rate	gaitCycle_diff
	Coefficient of variation (CV)	gaitCycle_CV
Hip joint swing angle	Mean Value	Hip_avg
	Left-right difference rate	Hip_diff
Knee joint swing angle	Mean Value	Knee_avg
	Left-right difference rate	Knee_diff
Swing phase proportion	Mean Value	Swing
Stance phase proportion	Mean Value	Stance

Statistical analysis

Hierarchical cluster analysis was used to analyze the clustering of each indicator after standardization. The correlation between the indicators was analyzed using Pearson correlation analysis. The significance of differences between the indicators in different age groups was analyzed using t-test analysis. A quadratic curve fitting method established an age prediction model based on the motor cognitive composite scores.

RESULTS

Correlation analysis and significance analysis

The correlation heat map after clustering at each indicator level is shown in Fig. 2. The results showed that the correlation between age and each of the other indicators (except gait parameters) was significant, except for BMI. The bar charts (mean±variance) and the results of the significant difference analysis for each indicator with a strong correlation with age in different age groups are shown in Fig. 3. The motor cognitive composite scores, walking speed, and dynamic cognitive scores showed a decreasing trend with age, while the SCD-9 scores, the number of chronic diseases, and the need to hold the treadmill handrail showed an increasing trend with age. There is a correlation between education level and age. The older the age, the lower the average education level (Fig. 3a). With SCD-9 scores reflecting the degree of cognitive decline, dual-task performance (motor cognitive composite score) was negatively correlated with SCD-9 scores and negatively correlated with age (Fig. 2, Fig. 3b). Older adults had decreased walking speed relative to younger adults (Fig. 3c). Walking speed correlated with cognitive performance, dual-task performance, and SCD (Fig. 2). The number of chronic diseases was positively correlated with age (Fig. 2, Fig. 3f) and negatively correlated with cognitive performance (Fig. 2). There was no significant correlation between BMI and age and other indicators except the number of chronic diseases. BMI was positively correlated with the number of chronic diseases (r = 0.23), with higher BMI being associated with a more significant number of diseases (Fig. 2).

Fig. 2

Correlations among the demographics, cognitive features, dual-task performance and gait parameters. The indicators covered in the horizontal and vertical coordinates of the chart and their corresponding short forms are listed in Tables 2 and 3. Correlations were given as correlation coefficients from negative (pink) to positive (green). The values of correlation coefficients were shown in the corresponding cells with its significant stars (^*p < 0.05, ^**p < 0.001). Features are clustered using Hierarchical clustering on the basis of Euclidean distances.

Comparison of bar charts (mean±variance) for each indicator at different ages for (a) education level; (b) motor cognitive composite score; (c) walking speed; (d) dynamic cognitive score; (e) SCD-9 score; (f) number of chronic diseases; and (g) need for a handrail (^*p < 0.05, ^**p < 0.001).

Among the gait parameters, the indicators of the mean ankle angle, the stance phase, and the left-right difference rate indicators of the angle of hip and knee showed correlations with indicators of cognitive performance and demographic indicators (Fig. 2). There was a significant negative correlation between the mean ankle angle and age, and a significant positive correlation between the left-right difference rate of the ankle angle and the left-right difference rate of the knee angle and age. Stance-phase and swing-phase proportions were weakly correlated with age. The need to hold the handrail on the walker was significantly positively correlated with the gait cycle and positively correlated with the knee left-right difference rate and the gait cycle left-right difference rate. The CV of the gait cycle was negatively correlated with the motor cognition composite score, walking speed, dynamic cognition score, and education level, and positively correlated with age, SCD, and number of chronic diseases.

Fig. 4

Scatterplot and fitting curves of age and motor cognitive composite score

Age and motor cognitive composite scores

Scatter plots and prediction curves for age versus motor cognitive composite scores (green dashed line) are shown in Fig. 4. The results showed that motor cognitive composite scores showed a decreasing trend with age, with cognitive decline accelerating in the older age group over 50 years. Linear fittings of age as the independent variable and motor cognition composite scores as the dependent variable were separately adapted for under and over 50 years of age. The results showed that the rate of cognitive decline in people over 50 years of age (yellow solid line, slope k = –1.06) was roughly six times the rate of cognitive decline before age 50 (purple solid line, k = –0.18).

DISCUSSION

The results of this study showed that motor cognitive composite score and dynamic cognitive score in walking-cognitive dual-task condition were significantly correlated with age SCD and other related indicators, respectively. Dual-task performance (motor-cognitive composite score, dynamic cognitive score) deteriorated with age. Walking speed, number of chronic diseases, educational level, and the need for treadmill handrails correlate with dual-task cognitive performance. Under the automatic walker scenario, the mean ankle angle, ankle angle left-right difference rate, knee angle left-right difference rate, and gait cycle CV had significant correlations with other cognitive indicators and age. It is suggested that the motor cognition composite score, dynamic cognition score, and gait parameter indexes in the dual-task state are potential objective markers characterizing cognitive decline. Mathematical modeling of the motor cognitive composite score with age depicted a trend of cognitive score decline with age, and the rate of decline was 6 times more rapid over the age of 50 than in the pre-50 population, which formed the basis for the prediction of cognitive age. Overall, our findings indicate that the method of dual task gait test provides possibilities for complementary clinical diagnosis, rapid screening, and community health care, facilitating the implementation of the grading diagnosis and treatment system for large-scale rapid cognitive screening.

Age-related cognitive decline is inevitable. According to the 2022 Chinese Expert Consensus on Cognitive Impairment Assessment, aging is one of the independent risk factors for cognitive impairment in older adults [5]. With age, gait performance deteriorates, as does cognitive response, the ability to allocate resources when multitasking occupies cognitive resources, and perceptual ability. This may be related to reduced social activity and loss of muscle mass in older adults. A hypothetical model of neural pathways based on compensatory effects suggests that older adults mobilize additional neural resources to achieve or maintain the same level of task performance. The hypothesis is based on the theory that brain activation is load-dependent. Evidence shows that the mechanism of this neural compensation will reach the upper limit in older people over 70 [25]. The model of reduced hemispheric asymmetry in older adults assumes that older adults mobilize compensatory resources localized in the opposite hemisphere, thus presenting stored bilateral brain activation. In contrast, younger adults would have a unilateral activation [26]. Research has been shown that fNIRS activation in the brain’s dorsolateral prefrontal cortex is more activated with age during single-walking task performance, suggesting that brain load becomes greater in older adults when processing tasks [26].

One possible hypothesis holds that concurrently performed two tasks interfere with each other and compete for limited cortical resources, increasing the cost of cortical attention processing, and thus walking-cognitive dual-task tests have been used to assess cognitive decline [7, 19], loss of mobility, and fall risk [27]. A study conducted a cognitive-motor dual-task assessment of community-dwelling cognitively normal post-stroke older adults and found that the dual-task walking assessment had good reproducible reliability and validity [28]. Another study showed that older adults, compared to younger adults, had worse verbal fluency, complexity of answer statements, and content richness when answering questions while walking [29]. The results of this study showed that walking-cognitive dual-task performance (including dynamic cognitive score and motor cognitive composite score) was significantly correlated with age and SCD, and dual-task performance deteriorated with age, and the higher the SCD-9 score, the worse the performance, which is consistent with the conclusion of the previous study about the decline in dual-task performance caused by aging, suggesting that the dual-task performance indicators can reflect the cognitive decline and has the potential to evaluate cognitive ability objectively.

Walking speed is significantly correlated with physical function, neurocognitive function, accelerated aging, and brain volume [30]. Decreased gait speed is a precursor to cognitive decline [30 –32]. Older adults in this study had lower walking speeds than younger adults (Fig. 4(c)). Walking speed was correlated with cognitive performance, dual-task performance, and SCD, showing an overall correlation between cognitive decline and lower walking speeds.

There was no significant correlation between BMI and age and other indicators except the number of chronic diseases, suggesting that cognitive function is uncorrelated with BMI. BMI was positively correlated with the number of chronic diseases (r = 0.23), which is consistent with previous findings that obesity is a risk factor for cardiovascular disease [33]. The results of the present study showed that the BMI of 75 to 84 years old individuals was significantly lower than that of every other age group (mean 22.66, Table 1), suggesting that older elders should be vigilant for weight loss [34] and sarcopenia to mitigate the escalating cognitive decline with aging.

Studies have shown that the incidence of age-related chronic diseases increases with the severity of the cognitive state [35]. The number of chronic diseases is also related to the ability to plan, number of instruments for activities of daily living, and attentional inhibition [36]. The results of this study showed a positive correlation between the number of chronic diseases and age, and a negative correlation with cognitive performance (positive correlation with dynamic cognitive scores and motor cognitive composite scores; negative correlation with SCD), which suggests that cognitive decline may be related to disease disturbance. It may also be possible that there is an interaction between worse cognitive performance and an increased number of chronic diseases that the causality needs to be further verified, as well as the interaction effect between age and the number of chronic diseases on cognitive decline.

The results of this study showed a strong correlation between educational level and age with other indicators (except BMI), respectively. Higher levels of education are less prone to cognitive decline, based on the theory of higher cognitive reserve and brain maintenance associated with education [37]. Studies have found that improved educational conditions in the early years have great potential to improve cognitive performance in early adulthood and reduce the public health burden of cognitive aging and dementia [37]. Conducting the interaction effect between educational level and age on cognitive decline [38–39] is the next step of this study.

With the increasing use of quantitative gait parameter acquisition tools, several studies have been devoted to finding correlations between gait parameters and cognitive decline. Increased gait variability, poor symmetry, and reduced regularity were found to be predictive of cognitive decline in elderly patients with cerebral small vessel disease (CSVD), proposing that gait variability and symmetry indicators could be used as biomarkers of cognitive impairment associated with CSVD [40]. Early cognitive processing dysfunction (attention, executive function, working memory, etc.) is associated with slower gait and gait instability [27]. The results of this study showed that gait parameters correlated with cognitive performance and demographic indicators. The mean value of ankle joint angle was significantly negatively correlated with age, and the left-right difference rate of ankle and knee joint angles was significantly positively correlated with age, suggesting that the range of ankle swing became smaller and the gait symmetry became worse with age, which was consistent with the fact that mobility deteriorated with aging. The temporal phase of gait was also weakly correlated with age, with a lower proportion of the stance phase and a more significant proportion of the swing phase at older ages, which is at odds with common sense. The possible reason for this is that older adults with weak perceptual abilities walk on a treadmill with a different gait from natural walking, and their fear of automatic walking platforms is relieved by increasing the frequency of foot changes, showing a tendency for an increased proportion of the swing phase. The need for treadmill handrails was also associated with some of the gait parameters. Individuals who needed assistance had slower gait speeds and worse gait symmetry. The gait cycle CV was negatively correlated with the motor cognition composite score, walking speed, dynamic cognition score, and education level; and positively correlated with age, SCD, and number of chronic diseases, indicating that the worse the cognitive function, the greater the gait variability. It suggests that individuals with weak cognitive abilities must adjust their gait more frequently at a fixed walking speed. In contrast, those with good cognitive abilities have a more stable gait and higher adaptability to a fixed-speed walking platform. These results demonstrate the predictive role of gait parameters for cognitive performance, suggesting a strong link between motor and cognitive functions. Gait parameters can be used as supplementary indicators for cognitive evaluation.

Studies have been dedicated to investigating the predictive effects of reliable biomarkers to characterize the state of brain health. The study identified a mismatch between chronological and physiological age and pointed to the potential value of dual-task performance as an early factor in accelerated aging [41]. Cole et al. [21–22] trained the brain MRI of healthy individuals by machine learning and obtained a brain age prediction model. The correlation between brain age and some measurements as well as mortality rate was also studied. It was found that for every 1-year increase in the difference between the brain’s predicted age and the chronological age, the risk of mortality before age 80 increased by 6%. This study developed a mathematical model with age as the independent variable and motor cognitive composite score as the dependent variable to predict cognitive age. In addition, the slope of cognitive decline was found to be inconsistent before and after age 50, suggesting that people should be fully aware of the problem of cognitive decline before midlife and proactive measures should be taken to prevent and intervene.

One of the limitations of this study is that it is difficult to find perfect subjects who are “supposed” to be healthy for their age in the older age group since it is not clear whether age-related diseases or age itself is the causal factor of cognitive decline. A larger data volume combined with a more detailed inclusion of basic information about subjects could avoid such problems. Therefore, the next step in the study is to conduct digital cognitive screening of community-based older adults in a broader area and to study the efficacy of interventions and training tools for delaying cognitive decline in old age. In addition, research on brain mechanisms of cognitive decline related to cognitive behavior will be performed to find further easily accessible objective indicators that specifically reflect cognitive decline.

Conclusion

The present study suggests that walking-cognitive dual-task performance and dual task gait parameters are potentially valuable assessing cognitive function. In addition, a mathematical model of the motor cognitive composite score and age lays the foundation for predicting cognitive age. The results of this study further demonstrate the role of the interactive coordination of motor and cognitive functions on cognitive maintenance, suggesting that cognitive aging can be delayed in middle-aged and elderly populations through cognitive and physical exercises.

AUTHOR CONTRIBUTIONS

Linlin Wang (Methodology; Visualization; Writing – Original Draft; Writing - Review & Editing; Formal Analysis; Investigation); Xuezhen Zhang (Conceptualization; Data curation; Visualization; Methodology; Writing - Original draft preparation; Writing – Review & Editing); Lei Wang (Conceptualization; Investigation; Methodology; Project administration; Supervision; Validation); Miaomiao Guo (Conceptualization; Investigation; Methodology; Validation); Qihang Yang (Validation; Writing – Review & Editing); Xiaogang Chen (Funding acquisition; Validation); Hong Sha (Project administration; Validation).

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

This work was supported by National Key R&D Program of China (No. 2022YFC3602803).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

References

Alzheimer’s Association (2022) 2022 Alzheimer’s disease facts and figures. Alzheimers Dement 18, 700–789.

Sperling

, Aisen

, Beckett

, Bennett

, Craft

, Fagan

, Iwatsubo

, Jack

, Kaye

, Montine

, Park

, Reiman

, Rowe

, Siemers

, Stern

, Yaffe

, Carrillo

, Thies

, Morrison Bogorad

, Wagster

, Phelps

(2011) Toward defining the preclinical stages of Alzheimer’s disease: Recommendations fromthe National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’sdisease. Alzheimers Dement 7, 280–292.

Federman

, Becker

, Mindt

, Cho

, Curtis

, Wisnivesky

(2023) Rates of undiagnosed cognitiveimpairment and performance on the Montreal Cognitive Assessment among older adults in primary care. J Gen Intern Med 38, 2511–2518.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

(2011) The diagnosis of mild cognitive impairment due toAlzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Associationworkgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

, Wu

, Song

, An

, Jiang

, Bai

, Wang

, Yu

, Zhang

, Wu

(2022) Chinese expert consensus onassessment of cognitive impairment in the elderly. Aging Med (Milton) 5, 154–166.

Yang

, Li

, Zhang

, Chen

, Li

, Wei

, Lu

, Liang

, Liu

, Shu

, Wang

, Guan

, Tao

, Wang

, Jia

, Ai

, Cui

, Wang

, Peng

, Zhang

, Chen

, Wang

, Zhao

, Wang

, Dong

, Wang

, Zhang

(2021) Early prevention of cognitive impairment in the community population: The Beijing Aging Brain RejuvenationInitiative. Alzheimers Dement 17, 1610–1618.

Ghoraani

, Boettcher

, Hssayeni

, Rosenfeld

, Tolea

, Galvin

(2021) Detection of mild cognitiveimpairment and Alzheimer’s disease using dual-task gait assessmentsand machine learning. Biomed Signal Process Control 64, 102249.

Wang

, Wang

, Guo

, Dove

, Arfanakis

, Qi

, Bennett

, Xu

(2023) Association of motor functionwith cognitive trajectories and structural brain differences. Neurology 101, e1718–e1728.

Schubert

, Pinto

, Paulsen

, Chappell

, Chen

, Engelman

, Ferrucci

, Hancock

, Johnson

, Merten

(2024) Midlife sensory and motor functions improve long-term predictions of cognitive decline andincidence of cognitive impairment. Alzheimers Dement 16, e12543.

10.

Ahn

, Chung

, Crouter

, Lee

, Anderson

(2023) Gait and/or balance disturbances associatedwith Alzheimer’s dementia among older adults with amnestic mild cognitive impairment: A longitudinalobservational study. J Adv Nurs 79, 4815–4827.

11.

Al-Yahya

, Dawes

, Smith

, Dennis

, Howells

, Cockburn

(2011) Cognitive motor interference whilewalking: A systematic review and meta-analysis. Neurosci Biobehav Rev 35, 715–728.

12.

Wang

, Yang

, Tian

, Zeng

, Yang

, Li

, Mao

(2023) A dual-task gait test detects mild cognitiveimpairment with a specificity of 91.2% . Front Neurosci 16, 1100642.

13.

Pieruccini Faria

, Black

, Masellis

, Smith

, Almeida

, Li

KZH

, Bherer

, Camicioli

, MonteroOdasso

(2021) Gait variability across neurodegenerative and cognitive disorders: Results from theCanadian Consortium of Neurodegeneration in Aging (CCNA) and the Gait and Brain Study. AlzheimersDement 17, 1317–1328.

14.

Ramírez

, Gutiérrez

(2021) Dual-task gait as a predictive tool for cognitive impairment inolder adults: A systematic review. Front Aging Neurosci 13, 769462.

15.

Tuena

, Pupillo

, Stramba-Badiale

, Riva

(2024) Predictive power of gait andgait-related cognitive measures in amnestic mild cognitive impairment: A machine learning analysis. Front Hum Neurosci 17, 1328713.

16.

Verghese

, Robbins

, Holtzer

, Zimmerman

, Wang

, Xue

, Lipton

(2008) Gait dysfunction in mildcognitive impairment syndromes. J Am Geriatr Soc 56, 1244–1251.

17.

Verghese

, Wang

, Lipton

, Holtzer

, Xue

(2007) Quantitative gait dysfunction and risk of cognitivedecline and dementia. J Neurol Neurosurg Psychiatry 78, 929–935.

18.

Hao

, Zhao

, Kimura

, Ukawa

, Kadoya

, Kondo

, Tamakoshi

(2021) Association of gait with globalcognitive function and cognitive domains detected by MoCA-J among community-dwelling older adults: Across-sectional study. BMC Geriatr 21, 523.

19.

Montero-Odasso

, Sarquis-Adamson

, Speechley

, Borrie

, Hachinski

, Wells

, Riccio

, Schapira

, Sejdic

, Camicioli

, Bartha

, McIlroy

, Muir-Hunter

(2017) Association of dual-task gait withincident dementia in mild cognitive impairment. JAMA Neurol 74, 857.

20.

Tapia

, Rocabado

, Duñabeitia

(2022) Cognitive estimation of speed, movement and time across the lifespan. J Integr Neurosci 21, 1–6.

21.

Cole

, Ritchie

, Bastin

, Hernández

, Maniega

, Royle

, Corley

, Pattie

, Harris

, Zhang

, Wray

, Redmond

, Marioni

, Starr

, Cox

, Wardlaw

, Sharp

, Deary

(2018) Brain age predictsmortality. Mol Psychiatry 23, 1385–1392.

22.

Cole

, Franke

(2017) Predicting age using neuroimaging: Innovative brain ageing biomarkers. TrendsNeurosci 40, 681–690.

23.

Jessen

, Wolfsgruber

, Wiese

, Bickel

, Mösch

, Kaduszkiewicz

, Pentzek

, Riedel Heller

, Luck

, Fuchs

, Weyerer

, Werle

, van den Bussche

, Scherer

, Maier

, Wagner

(2014) AD dementia risk inlate MCI, in early MCI, and in subjective memory impairment. Alzheimers Dement 10, 76–83.

24.

Gifford

, Liu

, Romano

, Jones

, Jefferson

(2015) Development of a subjective cognitive declinequestionnaire using item response theory: A pilot study. Alzheimers Dement (Amst) 1, 429–439.

25.

Nóbrega-Sousa

, Gobbi

LTB

, Orcioli-Silva

, Conceição

NRD

, Beretta

, Vitório

(2020) Prefrontal cortex activity during walking: Effects of aging and associations with gait and executivefunction. Neurorehabil Neural Repair 34, 915–924.

26.

Hoang

, Paire-Ficout

, Derollepot

, Perrey

, Devos

, Ranchet

(2022) Increased prefrontal activityduring usual walking in aging. Int J Psychophysiol 174, 9–16.

27.

Montero Odasso

, Verghese

, Beauchet

, Hausdorff

(2012) Gait and cognition: A complementary approachto understanding brain function and the risk of falling. J Am Geriatr Soc 60, 2127–2136.

28.

Tsang

CSL

, Chong

DYK

, Pang

MYC

(2019) Cognitive-motor interference in walking after stroke:Test–retest reliability and validity of dual-task walking assessments. Clin Rehabil 33, 1066–1078.

29.

Kemper

, Herman

, Lian

CHT

(2003) The costs of doing two things at once for young and older adults: Talking while walking, finger tapping, and ignoring speech of noise. Psychol Aging18181–192.

30.

Rasmussen

LJH

, Caspi

, Ambler

, Broadbent

, Cohen

, D Arbeloff

, Elliott

, Hancox

, Harrington

, Hogan

, Houts

, Ireland

, Knodt

, Meredith-Jones

, Morey

, Morrison

, Poulton

, Ramrakha

, Richmond-Rakerd

, Sison

, Sneddon

, Thomson

, Hariri

, Moffitt

(2019) Association ofneurocognitive and physical function with gait speed in midlife. JAMA Netw Open 2, e1913123.

31.

Cohen

, Verghese

, Zwerling

(2016) Cognition and gait in older people. Maturitas 93, 73–77.

32.

Jayakody

, Breslin

, Ayers

, Verghese

, Barzilai

, Milman

, Weiss

, Blumen

(2022) Relativetrajectories of gait and cognitive decline in aging. J Gerontol A Biol Sci Med Sci 77, 1230–1238.

33.

Battineni

, Sagaro

, Chintalapudi

, Amenta

, Tomassoni

, Tayebati

(2021) Impact of obesity-inducedinflammation on cardiovascular diseases (CVD). Int J Mol Sci 22, 4798.

34.

Ukraintseva

, Duan

, Holmes

, Bagley

, Wu

, Yashkin

, Kulminski

, Akushevich

, Whitson

, Stallard

, Yashin

, Arbeev

(2024) Patterns of aging changes in bodyweight may predict Alzheimer’s disease. J Alzheimers Dis 97, 163–170.

35.

Nam

, Park

, Bae

, Hong

(2023) Cognitive function level and chronic disease incidence rate by age ofolder adults living in the community. Alzheimers Dement 19, e062860.

36.

Thompson

, Demiris

, Rue

, Shatil

, Wilamowska

, Zaslavsky

, Reeder

(2011) a holistic approach toassess older adults’ wellness using e-health technologies. Telemed J E Health 17, 794–800.

37.

Lövdén

, Fratiglioni

, Glymour

, Lindenberger

, Tucker-Drob

(2020) Education and cognitivefunctioning across the life span. Psychol Sci Public Interest 21, 6–41.

38.

Ardila

, Ostrosky-Solis

, Rosselli

, Gomez

(2000) Age-related cognitive decline during normal aging:The complex effect of education. Arch Clin Neuropsychol 15, 495–513.

39.

Wilson

, Hebert

, Scherr

, Barnes

, Mendes de Leon

, Evans

(2009) Educational attainment andcognitive decline in old age. Neurology 72, 460–465.

40.

Mukli

, Detwiler

, Owens

, Csipo

, Lipecz

, Pinto

, Tarantini

, Nyul-Toth

, Balasubramanian

, Hoffmeister

, Csiszar

, Ungvari

, Kirkpatrick

, Prodan

, Yabluchanskiy

(2022) Gait variabilitypredicts cognitive impairment in older adults with subclinical cerebral small vessel disease. FrontAging Neurosci 14, 1052451.

41.

Zhou

, Cattaneo

, Yu

, Lo

, Gouskova

, Delgado-Gallen

, Redondo-Camos

, Espana-Irla

, Solana-Sanchez

, Tormos

, Lipsitz

, Bartres-Faz

, Pascual-Leone

, Manor

(2023)The age-related contribution of cognitive function to dual-task gait in middle-aged adults in Spain: Observations from a population-based study. Lancet Healthy Longev 4, e98–e106.