Epidemiology of visuospatial impairment in older adults: the Harmonized Cognitive Assessment Protocol Authors: Rogan G Magee, Yang Li, Michael J Rovine, Victoria S Pelak, Ali G Hamedani, Jennifer R Turner Published: 2026 ABSTRACT AbstractBackground and ObjectivesThis study aimed to determine the prevalence of visuospatial impairment in a nationally representative sample of older adults and identify self- and informant-reported functional measures associated with visuospatial dysfunction.MethodsWe analyzed data from 3496 participants aged ≥65 years in the 2016 Harmonized Cognitive Assessment Protocol (HCAP), a sub-study of the Health and Retirement Study. Visuospatial function was assessed using the Consortium to Establish a Registry for Alzheimer’s Disease constructional praxis and Mini Mental State Examination pentagon copy tasks. Impairment was defined as a factor score >1.5 SDs below expected relative to a robust normative sample. Cognitive status was classified by HCAP as normal, mild cognitive impairment (MCI), or dementia. Logistic regression models examined associations between visuospatial impairment and functional outcomes, adjusting for demographic, health, and general cognitive factors.ResultsVisuospatial impairment was present in 435 participants (11.33%). Among these, 49.85% had MCI, and 39.97% had dementia. Over half of those with MCI and visuospatial impairment had no other cognitive domain affected. Visuospatial impairment was associated with increased odds of self-reported visual difficulty (OR 1.60, 95% CI: 1.20–2.15), difficulty using a map (OR 3.35, 95% CI: 2.35–4.80), and informant-reported navigation difficulties, including getting lost at home (OR 17.30, 95% CI: 8.02–37.31). Associations with driving, map use, and getting lost at home remained significant after adjusting for general cognition.Discussion and ImplicationsVisuospatial impairment is common among older adults, frequently occurs in isolation in MCI, and is strongly associated with functional limitations, particularly in navigation and driving.Keywords: Health and retirement study, Constructional praxis, Mild cognitive impairment, Dementia This study shows that visuospatial impairment is a highly prevalent and often isolated cognitive deficit in older adults with mild cognitive impairment. By using nationally representative data and validated neuropsychological tools and linking them to real-world functional limitations such as spatial navigation and driving, it highlights a critical but underrecognized contributor to aging-related disability. These findings support the integration of visuospatial testing into cognitive aging research and inform policy and clinical strategies to enhance safety, independence, and quality of life in older adults. Background and objectives Visuospatial impairment occurs when the brain loses its normal ability to analyze visual information to support navigation and promote an understanding of the surrounding environment and its contents.1 It occurs in typical Alzheimer’s disease (AD)2 and can be a prominent feature of non-AD causes of dementia such as dementia with Lewy bodies,3 in which it is enriched in comparison to AD.4,5 Dementia can also present primarily as visuospatial impairment with relative preservation of other cognitive domains, a phenomenon known as the posterior cortical atrophy (PCA) syndrome.6 A variety of measures have been developed to quantify visuospatial impairments. Broadly, these assays capture deficits in orientation (eg, judgment of line orientation7 or object size4), navigation,8 construction,9 or perception (eg, of motion10). Several studies report that landmark-based navigation worsens with advancing age,11,12 highlighting an age-related decline in navigation dependent on medial temporal rather than parietal structures. Visuospatial function may also depend on and interact with other variables. For example, studies have demonstrated sex-based differences in navigation strategies, whereby women may rely more on landmarks and men more on orientation cues.13 Prior to the onset of dementia, visuospatial impairment also occurs in mild cognitive impairment (MCI), a stage defined by measurably reduced function in at least one cognitive domain14 that does not impact activities of daily living.15 MCI is divided into amnestic and non-amnestic subtypes, but most population-based research has focused on memory impairment in MCI.16,17 Isolated visuospatial impairment is relatively understudied in MCI. In 2003, Mapstone et al. reported that the perception of motion was impaired in one-third of older unimpaired adults and in those with MCI, and half of those with dementia diagnoses.10 Similar gradients of visuospatial impairment have been reported in studies exploring constructional praxis9 and simulated navigation tasks.8 Visuospatial dysfunction is a key contributor to functional impairment in individuals living with dementia.18,19 For example, fall risk20,21 and driving safety22,23 have both been linked to visuospatial impairment. Because people typically perceive visuospatial problems as related to vision rather than cognition, visuospatial symptoms are often directed to optometrists and ophthalmologists rather than primary care doctors or neurologists. For example, the average time between symptom onset and diagnosis for people with the PCA syndrome is 3.6 years, and one in six receives treatment for eye disease (eg, cataract surgery, capsulotomy, epiretinal membrane peel) before visuospatial impairment is recognized as the cause of their symptoms.24 Because these symptoms are often not recognized as being due to neurodegenerative disease until other cognitive domains are affected, early visuospatial predominant MCI is likely underrepresented in clinic-based cohorts such as the Alzheimer’s Disease Neuroimaging Initiative. Additional investigation into visuospatial impairment at the early stages of cognitive decline is needed. Until recently, large nationally representative studies such as the Health and Retirement Study (HRS) relied on telephone interviews to assess cognition, which allowed for effective testing of some domains such as verbal memory, but not others such as visuospatial function. In this study, we aimed to estimate the prevalence of visuospatial impairment in a large, nationally representative sample of older adults and identify self- and informant-reported daily functioning measures that are most strongly associated with visuospatial impairment. Research design and methods Standard protocol approvals, registrations, and patient consents This study was determined to be exempt from approval by the University of Pennsylvania Institutional Review Board (IRB) due to the fully deidentified nature of the datasets. The protocols for HRS were previously approved by the University of Michigan IRB. Informed consent was previously obtained at the time of study enrollment and was not separately required for this secondary analysis. Study population The HRS is a large, nationally representative longitudinal survey of US adults older than 50.25 Participants have been surveyed biennially since 1992, with the sample replenished in 1998, 2004, 2010, 2016, and 2022, such that over 42 000 people have participated in HRS to date. In 2016, a random sample of HRS participants aged 65 or older at the time of their 2016 core interview (n = 3496) was enrolled in a substudy called Harmonized Cognitive Assessment Protocol (HCAP), which consisted of a series of questionnaires and neuropsychological tests lasting approximately 1 hour.26,27 As part of this study, participants were asked to nominate up to 3 informants, one of whom was also interviewed. HCAP cognitive measures The HCAP cognitive testing protocol included several measures of visuospatial function.26 Using the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) constructional praxis copy task, subjects were asked to copy 4 line drawings in order of increasing complexity, with total scores ranging from 0 to 11.28 After approximately 2 minutes, they were asked to draw the shapes from earlier in the interview to the best of their memory, and results were scored using the same criteria. Subjects were also asked to draw intersecting pentagons as part of the Mini Mental State Examination (MMSE). Using these data, standardized factor scores for visuospatial function and other cognitive domains were previously calculated, and normative cutoffs were established. Briefly, this process consisted of selecting a robust normative sample from within HCAP and adjusting test scores within this population for age, sex, education, race, and ethnicity. Next, neuropsychological test scores were calibrated using latent variable models, and the resulting general and domain-specific factor scores were normalized and then standardized with respect to demographic variables. Finally, t scores (which represent the difference between an individual’s actual factor score and their expected score based on demographic variables) were calculated with a mean of 50 and an SD of 10.29 Subjects with t scores of 35 or lower (indicating that they performed >1.5 SDs below their expected score) were considered to have impairment in the respective cognitive domain. Based on diagnostic criteria from the National Institute on Aging and Alzheimer’s Association workgroups, dementia was defined as having impairment in 2 or more cognitive domains plus an informant report of functional impairment.30 People who did not meet the criteria for cognitive impairment in any domain were classified as having no cognitive impairment. People with impairment in only 1 cognitive domain were classified as having MCI unless both the respondent and informant reported no concerns about cognition or function, in which case they were classified as cognitively unimpaired. HRS-HCAP respondent and informant interview Informants were asked whether the subject was able to drive or use public transportation alone. As part of the Community Screening Instrument for Dementia, they were also asked if the subject gets lost in the community (such as when finding the post office or friends’ houses) or at home (such as when finding the toilet). The Blessed Dementia Rating Scale assesses a person’s ability to perform activities of daily living and relies on an informant interview with questions regarding a person’s ability to find their way around familiar streets or around the home or familiar indoor environments.31 Finally, the core HRS survey included a question asking subjects about the quality of their eyesight (excellent, very good, good, fair, or poor). We defined self-reported visual difficulty as a rating of fair or poor for overall, distance, or near eyesight. Statistical analysis We used descriptive statistics to summarize the characteristics of HCAP participants with and without visuospatial impairment. Because neuropsychological test scores were calibrated and demographically standardized relative to reference norms, raw test scores were not reported. Within each group, we determined the prevalence of normal cognition, MCI, and dementia using existing definitions in HCAP. Logistic regression was used to examine the association between visuospatial impairment and subject- and informant-reported functional abilities, adjusting for age, race/ethnicity, gender, marital status, education, income, smoking, diabetes, hypertension, heart disease, and stroke. To determine whether these associations were specific to visuospatial function rather than markers of general cognitive impairment, we also adjusted for a generalized cognition factor score, which weighted and harmonized tests of orientation, executive function, language, and memory but did not include visuospatial function.30 We used survey design variables provided by HRS (sampling weight, stratum, and primary sampling unit) to account for the complex survey design of HCAP, and variance estimation was performed using the Taylor series linearization. Subjects with missing data (n = 483) were excluded. Two-sided hypothesis testing was conducted with an alpha of 0.05. Results There were 3496 participants in the HRS-HCAP cohort, of whom 435 (11.33%) had visuospatial impairment. The characteristics of individuals with and those without visuospatial impairment are summarized in Table 1. Those with visuospatial impairment were more likely to be older, have lower income, and have a history of hypertension, heart disease, and stroke, and they were less likely to be married than those without visuospatial impairment. Of the 435 subjects with visuospatial impairment, 215 (49.85%) were classified as having MCI, and 178 (39.97%) had dementia (Figure 1); the remainder, 42 (10.19%), were classified as normal. Among those with MCI who had visuospatial impairment, over half (n = 106, 52.71%) had visuospatial function as their only affected cognitive domain. The remainder (n = 109, 47.29%) had impairment in at least one other HCAP-defined cognitive domain (orientation, executive function, memory, or language/fluency). Table 1Characteristics of older adults with vs. without visuospatial impairment in the Health and Retirement Study.VariableNo visuospatial impairment (n = 3061)Visuospatial impairment (n = 435) Age 74.3 ± 0.276.8 ± 0.8 Race/ethnicity White/Caucasian2423 (86.32%)332 (86.52%) Black/African American490 (9.38%)73 (9.21%) Hispanic/Other147 (4.31%)30 (4.27%) Gender Female1832 (55.73%)263 (57.61%) Married 1755 (60.79%)216 (50.41%) Education Less than upper secondary548 (14.57%)105 (19.62%) Upper secondary and vocational training1787 (57.66%)232 (53.69%) Tertiary725 (27.77%)98 (26.69%) Income 1st quartile (lowest)735 (20.27%)139 (27.69%) 2nd quartile768 (22.08%)106 (21.80%) 3rd quartile771 (25.42%)102 (25.72%) 4th quartile (highest)787 (32.23%)88 (24.79%) Smoking 1655 (55.77%)258 (55.99%) Diabetes 898 (27.39%)149 (27.84%) Hypertension 2150 (66.58%)331 (73.46%) Heart disease 1017 (30.86%)176 (40.14%) Stroke 346 (10.00%)102 (19.35%) Characteristics of older adults with vs. without visuospatial impairment in the Health and Retirement Study. Prevalence of visuospatial impairment, HRS-HCAP 2016. HRS-HCAP, Health and Retirement Study-Harmonized Cognitive Assessment Protocol; MCI, mild cognitive impairment. *Other cognitive domains included orientation, executive function, memory, and language/fluency. In multivariable-adjusted regression models, visuospatial impairment was associated with an increased odds of self-reported visual difficulty (OR 1.60, 95% CI: 1.20-2.15) and difficulty using a map (OR 3.35, 95% CI: 2.35-4.80). Visuospatial impairment was also associated with informant reports of getting lost in the community (OR 3.05, 95% CI: 2.16-4.31), getting lost at home (OR 17.30, 95% CI: 8.02-37.31), difficulty navigating familiar streets (OR 3.10, 95% CI: 2.16-4.45), and difficulty navigating at home (OR 4.14, 95% CI: 2.51-6.82). Visuospatial impairment was also associated with lower odds of driving (OR 0.28, 95% CI: 0.19-0.39) or using public transportation (OR 0.36, 95% CI: 0.28-0.47). After additionally adjusting for general cognition, associations with driving, using a map, and getting lost at home remained statistically significant (Figure 2). Point estimates for all covariates are presented in Supplementary Tables 1-8. Associations between visuospatial impairment and perceived difficulty with vision-related tasks. Discussion and implications In this study, we examined the prevalence of visuospatial impairment in MCI and dementia in a nationally representative sample of older adults. We found that approximately 1 in 10 older US adults has visuospatial impairment, about half of whom have MCI, and visuospatial impairment is often the only affected cognitive domain in MCI. We also found strong associations between visuospatial impairment and functional impairment across several activities, including navigation by car and public transport. These findings highlight the prevalence and significance of visuospatial impairment in the older adult population. Previous population-based studies have relied on tests such as the Telephone Interview for Cognitive Status or MMSE to screen for cognitive impairment.32 These tools provide little to no assessment of visuospatial function, and as a result, data on the epidemiology of visuospatial impairment in the general older adult population are limited. Using a combination of the constructional praxis portion of the CERAD battery and the pentagon copy included in MMSE, we found that the prevalence of visuospatial impairment in adults aged 65 and older in the United States was 11.33%, indicating that visuospatial impairment is common and easily overlooked in the general population when using certain screening tools. Of note, estimates of the prevalence of cognitive impairment depend not only on the overall distribution of cognitive function within a population but also on the sensitivity of cognitive tests and the method by which impairment is defined. For example, using only the Block Design task (a test of visuospatial function that correlates moderately with the CERAD constructional praxis33,34), the Sydney Memory and Aging Study found that the prevalence of visuospatial impairment was 4.1%.35 In contrast, a smaller study that utilized more extended screens of visuospatial function identified deficits in approximately one-fifth of subjects with a mean age of 76.10 The estimated prevalence of visuospatial impairment in HRS-HCAP lies in the middle of this range, but while using two tasks in HRS-HCAP is an advantage, this may still underestimate true prevalence given that discrete aspects of visuospatial function may require multiple tests in order to fully capture impairment.36 Standardized batteries like the Visual Object and Space Perception battery37 and self-reporting assessments such as the Colorado Posterior Cortical Questionnaire (CPC-Q)38 could be employed in future work to better capture the spectrum of visuospatial dysfunction by capturing other aspects such as illusory movement, visual tracking, depth perception, and mental rotation. We also found that half of those with visuospatial impairment met criteria for MCI, of whom half had impairment only in visuospatial function. Estimates of the prevalence of MCI vary widely and are complicated by several factors, including the choice of screening tools and diagnostic criteria.39 Meta-analyses of the global burden of MCI accounting for these differences yield largely concordant estimates of the total MCI burden around 15%-20%,40,41 similar to the published estimate of 22% in this cohort.30 Harmonized analyses of pooled data suggest that non-amnestic MCI may be more common than amnestic MCI.42 In this study, we found a significant burden of non-amnestic MCI attributable solely to visuospatial impairment. As exemplified by the lived experience of people with the PCA syndrome, symptoms of visuospatial impairment are often first reported to optometrists and ophthalmologists rather than primary care doctors or neurologists, potentially lengthening the time to detection of progressive cognitive impairment.24 We identified robust associations between visuospatial impairment and functional difficulties with vision-related tasks. Associations with driving and other forms of self-navigation persisted even after adjusting for general cognitive impairment, indicating that our analysis captured functionally significant visuospatial impairment. We adjusted for several possible confounders, including marital status, cardiovascular comorbidity, and stroke. These factors have previously been reported in relation both with visual impairment43,44 and driving-related injury42 and driving performance,45,46 which motivated their inclusion as model covariates. From a public health and safety perspective, our findings related to driving are especially pertinent. To date, clinically reliable neuropsychological screening tools for the prediction of driving safety during on-road assessment, including those that measure visuospatial function, remain scarce. Our finding that visuospatial impairment associates with informant report of driving ability, even after adjusting for global cognitive decline, suggests that appropriate in-office screening for visuospatial impairment may be useful as a screening tool for driving safety.47 Our study has several limitations. First, while qualitative assessments of vision were made by survey, objective data on visual acuity were lacking, limiting our ability to account for the effect of eye disease on visuospatial test performance. While this is a major limitation, previous studies have shown that reduced visual acuity is not a significant source of differential item functioning on visuospatial tests.48 Assessments of vision-related functioning were also limited, which represents another major limitation. Missing data could also have served as a source of information bias, though the frequency of this was relatively low. Our study did not capture the full spectrum of visuospatial impairment, nor did it capture other visual brain dysfunctions such as simultanagnosia or visual object perception deficits, which can be screened for using other tools such as the CPC-Q.38 Although the study recruited older adults from a nationally representative sample, it was conducted only in the United States, limiting the generalizability of our findings. This is especially important when considering the functional correlates of visuospatial impairment, as activities such as driving may have less importance in other cultural contexts. Finally, given the inherent limitations of a cross-sectional study design, future studies should examine longitudinal healthcare utilization and outcomes associated with visuospatial impairment in the general population. Supplementary Material Acknowledgments The HRS (Health and Retirement Study) is sponsored by the National Institute on Aging (NIA U01AG009740) and is conducted by the University of Michigan. The RAND HRS Longitudinal File is an easy-to-use dataset based on the HRS core data. This file was developed at RAND with funding from the National Institute on Aging and the Social Security Administration. This analysis uses data or information from the Harmonized HRS dataset and Codebook, Version C as of January 2022 developed by the Gateway to Global Aging Data. The development of the Harmonized HRS was funded by the National Institute on Aging (R01 AG030153, RC2 AG036619, 1R03AG043052). For more information, please refer to www.g2aging.org. Contributor Information Rogan G Magee, Department of Neurology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States. Yang Li, Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia, Pennsylvania, United States. Michael J Rovine, Graduate School of Education, University of Pennsylvania, Philadelphia, Pennsylvania, United States. Victoria S Pelak, Departments of Neurology and Ophthalmology, University of Colorado School of Medicine, Denver, Colorado, United States. Ali G Hamedani, Department of Neurology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States; Department of Ophthalmology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, United States. Supplementary material Supplementary material is available at Innovation in Aging online. Funding This work was supported by the National Institutes of Health (K23 EY033438 to A.G.H. and UE5 2UE5NS065745 to R.G.M.) and the John and Joanne Hare Fund for Accelerating PCA Research at the University of Colorado School of Medicine to V.S.P. The National Institute on Aging provided funding for the Health and Retirement Study (U01 AG009740) and the Harmonized Cognitive Assessment Protocol (U01 AG058499). Conflicts of interest None declared. Data availability The Health and Retirement Study and the Harmonized Cognitive Assessment Protocol data are publicly available for download through the Institute for Social Research at the University of Michigan and the Gateway to Global Aging. The analyses presented in this manuscript were not preregistered. References Associated Data Supplementary Materials Data Availability Statement The Health and Retirement Study and the Harmonized Cognitive Assessment Protocol data are publicly available for download through the Institute for Social Research at the University of Michigan and the Gateway to Global Aging. The analyses presented in this manuscript were not preregistered.