Determinants of Activity Levels in African Americans with Mild Cognitive Impairment

Abstract

Engaging in cognitive, social, and physical activities may prevent cognitive decline. In a sample of older African Americans with Mild Cognitive Impairment (N=221), we investigated the cross-sectional relationships between activity levels and participants’ demographic, clinical, and neuropsychological characteristics. The average age of participants was 75.4 years (standard deviation [SD] 7.0); 177 (80.1 %) were women. Participation in cognitive/social activities was positively associated with education, depression, literacy, mobility, instrumental activities of daily living (IADL), verbal learning, and subcomponents of executive function. A linear regression identified IADLs, education, depression, and verbal learning as independent predictors. Participation in physical activities was positively associated with gender, depression, IADLs, and subcomponents of executive function. An ordinal regression identified executive function and depression as independent correlates. These data suggest that unique characteristics are associated with cognitive/social and physical activities in older African Americans with MCI. These characteristics, coupled with low activity levels, may increase the risk of progression from Mild Cognitive Impairment to dementia. Culturally relevant behavioral interventions to reduce cognitive decline in this high risk population are needed.

Introduction

Mild Cognitive Impairment (MCI) is a transition state between normal cognitive aging and dementia that affects 10% to 20% of older persons and predicts progression to dementia, particularly Alzheimer’s disease (AD).^1–3 Older African Americans are more likely to meet criteria for MCI than whites, possibly because comorbid medical problems (e.g., diabetes, hypertension), obstacles to health care, low health literacy, fewer years of education of uncertain quality, and limited economic resources increase their risk.^4–7 These factors also explain why African Americans are diagnosed with AD at more advanced stages than whites, when care is costlier and outcomes are less promising.^4,7 Culturally relevant preventive interventions are therefore needed to preserve cognition in older African Americans.

An extensive epidemiologic literature suggests that cognitive, social, and physical activities may prevent cognitive decline in cognitively normal persons.^8–13 Whether engaging in these activities reduces cognitive decline in persons with MCI is uncertain. Some studies suggest that cognitive rehabilitative interventions for persons with MCI improve targeted cognitive functions but few other cognitive domains or activities.^14–16 Physical and psychosocial interventions also have modest cognitive benefits.^17–19 Possible mechanisms for these effects may include reduced amyloid deposition, improved cardiovascular and cerebrovascular integrity, better glycemic control, lower blood pressure, and better general physical and mental health. ^20–22

We are currently conducting a randomized controlled trial to test the efficacy of a behavioral intervention to reduce cognitive decline by increasing activity participation in older African Americans with MCI.²³ As part of that research, we are interested in identifying the baseline characteristics of participants who respond best, or less well, to the intervention. One predictor of response may be their baseline level of activity. As the clinical trial is ongoing, no outcome results are yet available. In this cross-sectional study, we investigated the relationship between participation in cognitive, social, and physical activities and participants’ demographic, clinical, and neuropsychological characteristics. We tested the hypothesis that participation in these activities would be associated with education, depression, mobility, various aspects of cognition, and everyday functioning. The results may identify important treatment moderators, and thereby increase the theoretical and translational impact of the clinical trial.

Methods

The methods of this clinical trial have previously been reported.²³ Briefly, 221 African Americans were enrolled in a clinical trial testing the efficacy of a behavioral intervention to reduce cognitive decline over 2 years. The inclusion criteria are: 1) self-identified African American race; 2) age over 65 years; and 3) amnestic MCI (single or multiple domain). To meet the latter criterion, participants had self-reported memory decline; preserved general cognition (i.e., Mini Mental Status Examination [MMSE] score over 22); self-reported independence in basic activities of daily living; and scores more than 1.5 standard deviations (SD) below age-, education-, and race-adjusted norms in delayed recall on the Hopkins Verbal Learning Test-Revised (HVLT-R).^24–26 The exclusion criteria were: 1) Axis I Diagnostic and Statistical Manual-IV psychiatric diagnosis; 2) sensory deficits that precluded neuropsychological testing; 3) reduced life expectancy (given the 2 year follow-up period); and 4) institutional residence. Race-concordant Community Health Workers (CHWs) recruited participants from senior centers, senior high rise apartment buildings, churches, and primary care clinics. To screen for cognitive impairment, the CHWs administered Trial 1 of the HVLT-R (immediate recall) to potentially eligible participants. Recalling fewer than 5 of 12 words was considered screen positive. A previous study reported that the average Trial 1 HVLT-R score for African Americans aged 60–71 years is 4.4 (SD 1.3).²⁶ A total of 1,034 persons were screened and 524 (50.7%) screened positive. For the latter, the CHWs scheduled an in-home visit to obtain informed consent and complete the baseline assessment. Jefferson’s Institutional Review Board approved this approach to contacting, consenting, and evaluating persons for possible research participation. The CHWs obtained the following data at baseline:

Demographic characteristics

These included age, sex, education, marital status, and literacy as measured on the Reading Recognition Subtest of the Wide Range Achievement Test-Version 3 [WRAT-3]).²⁷

Physical health status

Participants reported the presence of chronic medical conditions, mobility (i.e., ability to walk two blocks without help), and medication use. The latter was used to obtain the Chronic Disease Score, which is a weighted score based on current medication use that predicts healthcare utilization and costs.²⁸ Higher scores (transformed into z scores) reflect greater severity of medical comorbidity.

Instrumental activities of daily living (IADL)

This was assessed with the Alzheimer’s Disease Cooperative Study Activities of Daily Living–Prevention Instrument (ADL-PI), which self-rates level of difficulty performing instrumental activities over the past 3 months (e.g., managing medications, handling money, shopping, preparing food).²⁹ Scores range from 0 to 45 with higher scores representing better function.

Depressive symptoms

This was assessed with Geriatric Depression Scale (GDS-15 items), which self-rates depressive symptoms (e.g., depressed mood, loss of interests, hopelessness) as “present’ or “absent” over the past week.³⁰ Scores range from 0 to 15 with higher scores indicating worse depression.

Participation in activities

This was assessed with the Florida Cognitive Activities Scale (FCAS) and the U.S. Health Interview Survey (US-HIS).^31–33 The FCAS is a 25-item measure of predominantly cognitive and social activities (“cognitive/social”) with known reliability and validity in older African Americans. It rates the frequency of activity participation on a 5 point scale as: “used to do activity but not in past year” (0); “a few times a year” (1); “every couple of months” (2); “a couple times a month” (3); “a few times a week” (4); and “every day” (5). Items include board games, crossword puzzles, reading the newspaper, craft work, home repairs, cooking, and going to church or social clubs. Scores range from 0 to 125, with higher scores indicating more frequent activity participation. The 9-item US-HIS characterizes participation in physical activities (e.g., walking for exercise, dancing, swimming, and biking) as the number of times, and amount of time in minutes, that a person engages in each activity in the previous 2 weeks. Each item is scored as the product of the number of times a person engages in an activity by the number of minutes spent in that activity.

Cognition

This was assessed with the Hopkins Verbal Learning Test-Revised (HVLT-R) and the National Alzheimer’s Coordinating Center’s (NACC) Uniform Data Set (UDS) Neuropsychological Battery.^{25, 34} The HVLT-R is a word-list learning and memory test that consists of a 12 item-word list presented in 3 consecutive learning trials. The Total Recall score is the sum of the 3 learning trials; scores range from 0 – 36. The Delayed Recall score is the number of correctly recalled words 20 minutes later; scores range from 0 – 12. The UDS Neuropsychological Battery includes the MMSE and tests of immediate and delayed verbal episodic memory (Logical Memory of the Wechsler Memory Scale [WMS-R] – Revised) IA and IIA, respectively); attention (Digit Span, Forwards and Reverse); semantic memory/language (Boston Naming and Category Fluency Test); processing speed/visuospatial ability (WAIS-R Digit Symbol and Trailmaking Test Part A); and various subcomponents of executive function (Trailmaking Test Part B, Digit Span, and Digit Symbol). Participants who met criteria for amnestic MCI and scored below the 7^th percentile in a second cognitive domain met criteria for amnestic MCI-Multiple Domain.

Statistical methods

Bivariate comparisons across FCAS-defined activity groups were made using ANOVA for continuous variables and chi square or Fisher’s Exact Test (for expected values less than 5) for categorical variables. A multivariable regression with FCAS score as the dependent variable was conducted, with all baseline variables considered as covariates. For the US-HIS, proportional odds logistic regression was used to estimate bivariate associations between the ordinal dependent variable of US-HIS score grouping and all baseline variables. A multivariable proportional odds model was fit with US-HIS score grouping as the dependent variable and all baseline variables considered as covariates. Both unadjusted and adjusted odds ratios for having higher activity were calculated.

Results

The average age of participants was 75.4 years (standard deviation [SD] 7.0); 177participants (80.1%) were women; and mean years of education was 12.5 (SD 2.6). One hundred seventy six participants (79.6 %) met criteria for amnestic MCI-Multiple Domain and 45 participants (20.4 %) met criteria for amnestic MCI-single domain. Table 1 shows the total sample mean scores and percentile ranks on the various neuropsychological tests. Percentile scores were based on normative scores from the appropriate test manuals, adjusted for age, race, and education.^35–37 As expected, the lowest scores and percentile ranks were on tests of memory. The higher scores and ranks on tests of semantic memory, language, and executive function suggest that the overall sample was not globally cognitive impaired.

Table 1.

Average Neuropsychological Test Scores of the Study Sample (n = 221)

	Mean	SD	Percentile
MMSE	25.81	2.40	13%
HVLT Total Recall1	14.76	3.46	15%
HVLT Delayed Recall1	1.75	1.83	1–2%
HVLT Learning Curve1	-0.02	1.39	2–3%
Boston Naming Test (15-item)	12.21	2.19	37%
WMS-R Logical Memory A-Immediate2	9.00	3.52	---
WMS-R Logical Memory A-Delayed2	6.17	3.89	---
Category Fluency (animals)	12.37	3.55	45%
Trail Making Test, Part A	78.41	34.49	4%
Trail Making Test, Part B	238.49	75.77	23%
WAIS-III Digit Symbol3	31.59	13.29	9%
Digit Span Forwards	7.07	1.88	16%
Digit Span Backwards	4.35	1.50	5%

¹Hopkins Verbal Learning Test

²Wechsler Memory Scale (mean scores are similar to those reported in persons with MCI).⁴⁸

³Wechsler Adult Intelligence Scale

FCAS scores were normally distributed (range 0 to125; median 53; and mean [SD] 54.0 [14.2]. Participants were categorized as having low (i.e., one standard deviation [SD] below the mean), intermediate (i.e., within one SD below or above the mean), or high (i.e., one SD above the mean) levels of activity. Thirty seven participants (16.7%) had low activity levels; 147 (66.5%) had intermediate activity levels; and 37 (16.7%) had high activity levels. Table 2 compares the demographic, clinical, and neuropsychological characteristics of participants by FCAS activity level. The latter was positively associated (i.e., p ≤ 0.05) with education, literacy (WRAT-3 score), mobility, IADL function, GDS score, and HVLT Total Recall, WMS-R Logical Memory IA, Digit Symbol Substitution, and Trailmaking Test Part A and B scores. All variables in Table 2 were included in a linear regression analysis with FCAS scores as the dependent variable. Table 3 shows the regression results, and indicates that IADL function, education, GDS score, and HVLT Total Recall score each contributed independently to cognitive/social activity participation. The model accounted for 25% of the variance.

Table 2.

Demographic and Clinical Characteristics (N = 221) by Cognitive/Social Activity Level (Florida Cognitive Activity Scale)

Variables	Low (n = 37)	Medium (n = 147)	High (n = 37)	Statistic	df	p
Demographic/Psychosocial
Age (mean, SD)	74.3 (7.2)	76.0 (7.2)	74.2 (5.6)	F = 1.60	2,218	.204
Female (n, %)	32 (86.5)	116 (78.9)	29 (78.4)	X2 = 1.15	2	.564
Education, yrs. (mean, SD)1	11.8 (2.7)	12.5 (2.5)	13.4 (2.6)	F = 4.04	2,218	.019
WRAT (mean, SD)2,3	23.3 (7.9)	27.6 (6.8)	28.1 (5.5)	F = 6.44	2, 218	.002
GDS score (mean, SD)4,5	5.2 (3.2)	3.2 (3.1)	2.1 (2.1)	F =11.2	2, 218	<.001
Physical
Chronic Disease Score (mean, SD)	4335.6 (4007.5)	4552.5 (4571.3)	3553.2 (2389.7)	F = .83	2, 218	.433
Able to walk 2 blocks without assistance (n, %)	18 (50)	114 (79.2)	29 (78.4)	X2 = 13.20	2	.001
ADL-PI score (mean, SD)6,7	30.8 (6.6)	36.3 (5.3)	38.8 (4.3)	F = 22.00	2, 218	<.001
Neurospsycholgocal
Mini Mental Status Score	25.6 (2.9)	25.8 (2.3)	26.2 (2.3)	F = .51	2, 218	.602
HVLT Total Recall78	13.1 (3.6)	14.8 (3.4)	16.3 (2.9)	F =8.23	2, 218	<.001
HVLT Delayed Recall8	2.1 (1.8)	1.6 (1.8)	1.9 (1.9)	F = 1.41	2, 218	.248
Boston Naming Test	11.6 (2.4)	12.3 (2.2)	12.6 (1.7)	F = 2.10	2, 218	.124
Category Fluency	11.8 (3.7)	12.3 (3.6)	13.2 (3.1)	F = 1.49	2, 218	.228
Trail Making Test, Part A9	91.6 (38.5)	77.4 (33.1)	69.1 (32.7)	F = 4.26	2, 218	.015
Trail Making Test, Part B10	249.7 (73.0	244.6 (73.6)	202.9 (78.9)	F = 5.17	2, 218	.006
Digit Symbol Substitution10	27.8 (13.2)	31.1 (12.7)	37.2 (14.2)	F = 5.11	2, 218	.007
Digit Span - Forward	6.7 (1.9)	7.1 (1.8)	7.2 (2.0)	F = .87	2, 218	.420
Digit-Span - Backward	4.3 (1.5)	4.4 (1.4)	4.4 (1.8)	F = .06	2, 218	.940
Wechsler Memory Scale-R Logical Memory IA9	7.6 (2.7)	9.1(3.7)	10.0 (3.2)	F = 4.39	2, 218	.014
Wechsler Memory Scale-R Logical Memory IIA	5.2 (3.5)	6.2 (4.1)	7.0 (3.3)	F = 1.983	2, 218	.140
Amnestic MCI-multiple domain (n, %)	6 (16.2)	25 (17)	14 (37.8)	X2 = 8.38	2	.015

¹The low activity group has fewer years of education than the high activity group.

²Wide Range Achievement Test; higher scores signify better literacy.

³The low activity group has significantly lower scores than the middle and high activity groups.

⁴Geriatric Depression Scale; higher scores signify worse depression.

⁵The low activity group has significantly higher scores than the middle and high activity groups.

⁶Alzheimer’s Disease Cooperative Study Activities of Daily Living–Prevention Instrument; higher scores signify better function.

⁷All 3 activity groups differ significantly from each other.

⁸Hopkins Verbal Learning Test; higher scores signify better memory.

⁹Scores for the low activity group are significantly worse than those of the high activity group.

¹⁰Scores for the high activity group are significantly better than scores of the middle and low activity groups.

Table 3.

Regression results: Baseline Variables associated with Florida Cognitive Activity Scale Scores.

Variable	Beta	p	R2
ADL Total Score1	.24	<.001
Years of education	.18	.005	25%
GDS2	−.20	.003
HVLT Total Recall3	.16	.01

¹Alzheimer’s Disease Cooperative Study Activities of Daily Living–Prevention Instrument

²Geriatric Depression Scale score

³Hopkins Verbal Learning Score

The distribution of US-HIS scores was positively skewed [skewness = 4.7 (.16); kurtosis = 32.3 (.33)], with nearly 20% of the sample (n = 45) reporting no physical activity. The most common activities in which participants engaged were walking for exercise (68.2%), calisthenics or general exercise (44.7%), dancing (18.5%), and gardening or yard work (14.6%). Participants engaged in an average of 5.4 hours (SD 8.4) of physical activity/2 weeks. We divided the sample into three groups: participants who engaged in fewer than 60 minutes of exercise/2 weeks; 61 - 149 minutes/2 weeks; or 150 minutes or more/2 weeks.

Table 4 shows the unadjusted and adjusted odds ratios and confidence intervals for the relationship of participants’ demographic, clinical, and neuropsychological characteristics and US-HIS physical activity. The odds ratios indicate the likelihood of being in a higher physical activity group. The unadjusted analyses show that physical activity was positively associated (i.e., p ≤ 0.05) with male gender, scores on the WRAT, GDS, and chronic disease (CDS), ability to walk two blocks, and IADL performance. Notably, none of the neuropsychological test scores were related to physical activity in the unadjusted analysis. The adjusted analyses, however, indicate that scores on the GDS, Digit Span Backwards, and Trailmaking Test Part B contributed independently to physical activity participation. The 0.90 odds ratio for GDS indicates that every one point increase in GDS score (i.e., worse depression) reduces the odds of being in the higher physical activity group by 10%. The adjusted odds ratios for Trailmaking Test Part B [i.e., 1.05; (95% CI 1.00, 1.09)] and Digit Span Backwards [i.e., 0.80; (95% CI 0.66, 0.99)] suggest that better performance on these tests was associated with lower physical activity, although these relationships may represent a statistical artifact given the absence of significant relationships in the unadjusted analysis.

Table 4.

Ordinal Regression Results Predicting Physical Activity Level (n = 221)¹

Variables	Unadjusted Proportional OR (CI)	p	Adjusted Proportional OR (CI)	p
Demographic/Psychological
Age	.97 (.94; 1.00)	.078	.96 (.92; 1.01)	.129
Female	.41 (.22; .77)	.006	.50 (.25; 1.02)	.057
Education, yrs	1.07 (.97; 1.18)	.190	1.02 (.90; 1.15)	.741
WRAT2	1.04 (1.00; 1.08)	.043	1.04 (.99; 1.10)	.095
GDS score3	.87 (.80; .95)	.001	.90 (.81; 1.00)	.040
Physical
Chronic Disease Score4	.93 (.86; 99)	.027	.97 (.90; 1.04)	.349
Not able to walk 2 blocks	.57 (.32; 1.00)	.050	.89 (.46; 1.74)	.734
ADL-PI score5	1.07 (1.03; 1.12)	.001	1.03 (.98; 1.09)	.223
Neuropsychological
Mini Mental Status Score	1.04 (.94; 1.15)	.482	1.01 (.89; 1.16)	.861
HVLT Total Recall6	1.05 (.97, (1.12)	.221	1.00 (.92;1.09)	.974
HVLT Delayed Recall6	1.07 (.94; 1.22)	.298	1.04 (.89; 1.22)	.585
Boston Naming Test7	1.05 (.94; 1.18)	.375	0.96 (.83; 1.10)	.535
Category Fluency8	1.04 (.97; 1.12)	.226	1.00 (.93; 1.09)	.910
Trail Making Test, Part A9	.93 (.87; 1.00)	.058	.96 (.86; 1.07)	.421
Trail Making Test, Part B9	1.00 (.97; 1.04)	.813	1.05 (1.00; 1.09)	.049
Digit Symbol Substitution10	1.15 (.96; 1.39)	.135	1.02 (.78; 1.34)	.866
Digit Span – Forward11	.99 (.87; 1.12)	.849	1.02 (.87; 1.19)	.832
Digit-Span – Backward11	.86 (.73; 1.01)	.066	.80 (.66; .99)	.037
Wechsler Memory Scale-R Logical Memory IA12	1.02 (.95; 1.09)	.640	.99 (.88; 1.11)	.895
Wechsler Memory Scale-R Logical Memory IIA12	1.05 (.99; 1.12)	.129	1.01 (.91; 1.13)	.838

¹The low group engaged in 0 to 59 minutes of activity in past 2 weeks (coded 1), the middle group engaged in 60 to 149 minutes (coded 2), and the high group engaged in more than 150 minutes (coded 3). Odds ratios are for odds of having higher activity. Deviance = 434.43, p = .176; Pseudo R² = .175.

²Wide Range Achievement Test

³Geriatric Depression Scale

⁴Expressed in increments of $1,000.

⁵Alzheimer’s Disease Cooperative Study Activities of Daily Living–Prevention Instrument; higher scores signify better function.

⁶Hopkins Verbal Learning Test

⁷Scored as the number of correctly identified pictures.

⁸Scored as the number of animals named in 60 seconds.

⁹Scored as the number of seconds for task completion; expressed in 10 second increments.

¹⁰Scored as the number of correct symbols; expressed in 10 symbol increments.

¹¹Scored as the number of correctly recalled sequences.

¹²Scored as the number of correctly recalled story components.

Discussion

We found that unique demographic, clinical, and neuropsychological characteristics were associated with participation in cognitive/social and physical activities in older African Americans with MCI. All participants had clinically meaningful deficits in memory and most had deficits in other cognitive domains. The participants were not, however, globally cognitively impaired or dependent in basic activities of daily living, and therefore did not meet criteria for dementia.⁴² Cognitive/social activities were independently associated with instrumental activities of daily living, education, depression, and verbal learning. Physical activities were independently associated with depression. All of these factors are known to contribute to the risk of progression from MCI to dementia, and their association with low activity levels may magnify that risk. In fact, low activity levels may represent an early sign of progression in this sample.^38–41

The results of this study must be understood in the context of the study’s limitations. The participants were recruited from the community and primary care clinics and enrolled in a randomized controlled trial. These characteristics limit generalizability. We conducted many statistical comparisons and thus there is a risk of finding statistically significant relationships by chance that are not clinically meaningful. Our use of the HVLT to identify potential participants constrained test scores, but HVLT Total Recall scores had substantial variability and range. By contrast, Delayed Recall scores were uniformly low as expected, given the study eligibility criteria. We relied on self-reports of functional abilities and activity levels, which may be biased by faulty recall. Some studies suggest, however, that individuals with MCI can provide accurate appraisals of their functional abilities.⁴³ We did not assess some variables which likely relate to activity participation, such as pain, personality traits (e.g., extroversion), access to transportation, and social support, and we used different metrics to rate participation in cognitive/social and physical activities (i.e., number of occasions and number of minutes, respectively). The FCAS and the US-HIS target different activity domains, however. Most FCAS activities are cognitive and social in nature although some FCAS activities (e.g., walking with a friend) require physical activity. The US-HIS more clearly assesses physical activity. Previous studies have noted the difficulty of deriving pure activity classifications; Schinka et al (2005) discuss this issue extensively.^44,45 Despite these limitations, our results provide new insights into the determinants of activity participation in this unique sample of older African Americans.

Depressive symptoms were common to low participation in cognitive/social as well as physical activities. Loss of interest, anhedonia, and anergia diminish motivation to pursue many different types of activities and, in persons with MCI, amplify the effects of apathy, which is the most frequently occurring behavioral symptom of MCI.⁴⁶ Whether depression is a risk factor for or prodromal sign of dementia is uncertain but depression adds to apathy to impede activity engagement, and may accelerate progression to dementia.

The bivariate analyses for cognitive/social activities revealed positive associations with various aspects of executive function, including attention, cognitive flexibility, perceptual-scanning skills, and processing speed. These subcomponents of executive function were not independent predictors of cognitive/social activity, however, after controlling for HVLT Total Recall. The latter involves both learning, an indicator of hippocampal function, and working memory, a prefrontal executive function. To the extent that Total Recall recruits the latter, and to the extent that cognitive/social activities depend on planning, initiating, and sustaining activity, then worse performance on Total Recall may understandably be linked to lower levels of participation in these types of activities. Moreover, as a sensitive indicator of Alzheimer’s disease pathology, worse Total Recall may be a proxy for more pervasive underlying disease.^{24, 47}

Education was also associated with cognitive/social activity levels. Persons with less education and low literacy have less opportunity to pursue cognitive and leisure activities due to financial and intellectual limitations and, on that basis, may be at higher risk for dementia.^{10, 48} Literacy, as an indicator of quality of education, may provide a more precise indicator of cognitive reserve than years of education, and determine whether someone with underlying brain pathology expresses symptoms of dementia.⁴⁹

We also found that performance of instrumental activities of daily living was independently associated with cognitive/social activity levels. Persons with MCI often have difficulty with or take longer completing such higher order activities (e.g., traveling, managing finances, and taking medications) and these difficulties increase the risk of dementia.^{41, 50}

Regarding physical activity participation, the unadjusted analyses revealed positive associations with depression, gender, literacy, physical functioning, and IADLs but no association with neuropsychological test performance. The latter finding suggests that cognitive function per se in this sample of older African Americans with MCI is unrelated to physical activities. Sturman et al (2005), using the same instrument we used to assess physical activity participation, and some of the same neuropsychological tests, in a sample of older persons (61.1% African American), similarly found no association between physical activity participation and neuropsychological test performance.³³ They hypothesized that the absence of an independent relationship may reflect the moderately low levels of physical activity in older urban populations.

To our knowledge, no studies similar to ours have examined relationships between activity levels and demographic, personal, and cognitive characteristics of older persons with MCI. Our results are consistent with studies that have examined these relationships in representative populations of older persons. Dotson et al (2008), studying activity participation using the FCAS in a large community sample of older African Americans, found small negative correlations between higher activity levels and age and depressive symptoms, and moderate positive correlations with years of education and overall cognition, memory, and executive function.³¹ Gow et al (2012) and Wilson et al (2003) found participation in cognitive, leisure, or physical activity was associated with higher levels of cognitive ability, which reflected prior ability and more years of education.^{8, 48} Together these results suggest both a protective effect of activity on cognition and a facilitative effect of pre-existing cognitive strengths on later activity levels. The converse, as we found, is that low activity participation reflects deficits in mood, memory, executive function, education, and everyday functioning, which represent five unique domains of mental and physical life. These characteristics may moderate treatment outcomes in the clinical trial we are now conducting, and identify participants who respond best, because the intervention reverses activity loss in participants at greatest risk, or who respond less well because they have more advanced disease.

The mechanism by which activity participation may be neuroprotective and reduce cognitive decline, if it does, in persons with MCI, is unknown. Older African Americans, however, comprise one of the fastest growing minority groups in the United States and nearly 25% have MCI.^{4, 51} These demographic and clinical realities, coupled with the uncertain efficacy and safety of current investigational medications for Alzheimer’s disease and their limited accessibility to older African Americans, necessitate culturally relevant behavioral interventions to prevent cognitive decline in this high risk population.⁵²