The Resilience Index: A Quantifiable Measure of Brain Health and Risk of Cognitive Impairment and Dementia

Abstract

Background:

There is increasing interest in lifestyle modification and integrative medicine approaches to treat and/or prevent mild cognitive impairment (MCI) and Alzheimer’s disease and related dementias (ADRD).

Objective:

To address the need for a quantifiable measure of brain health, we created the Resilience Index (RI).

Methods:

This cross-sectional study analyzed 241 participants undergoing a comprehensive evaluation including the Clinical Dementia Rating and neuropsychological testing. Six lifestyle factors including physical activity, cognitive activity, social engagements, dietary patterns, mindfulness, and cognitive reserve were combined to derive the RI (possible range of scores: 1–378). Psychometric properties were determined.

Results:

The participants (39 controls, 75 MCI, 127 ADRD) had a mean age of 74.6±9.5 years and a mean education of 15.8±2.6 years. The mean RI score was 138.2±35.6. The RI provided estimates of resilience across participant characteristics, cognitive staging, and ADRD etiologies. The RI showed moderate-to-strong correlations with clinical and cognitive measures and very good discrimination (AUC: 0.836; 95% CI: 0.774–0.897) between individuals with and without cognitive impairment (diagnostic odds ratio = 8.9). Individuals with high RI scores (> 143) had better cognitive, functional, and behavioral ratings than individuals with low RI scores. Within group analyses supported that controls, MCI, and mild ADRD cases with high RI had better cognitive, functional, and global outcomes than those with low RI.

Conclusion:

The RI is a brief, easy to administer, score and interpret assessment of brain health that incorporates six modifiable protective factors. Results from the RI could provide clinicians and researchers with a guide to develop personalized prevention plans to support brain health.

Keywords

Alzheimer’s disease brain health cognitive activity cognitive reserve dementia diet lifestyle mild cognitive impairment mindfulness physical activity resilience risk factors

INTRODUCTION

Alzheimer’s disease and related dementias (ADRD) currently affects 6.2 million Americans [1] and over 50 million people worldwide [2]. By year 2050, the number of ADRD cases is expected to increase 3-fold as the number of people over 65 years old grows by 62% and the number over 85 years of age is expected to grow by 84% [1 , 3–5]. Community detection of mild cognitive impairment (MCI) [6] and the early stages of ADRD is a challenge to clinicians and researchers alike, requiring in-depth evaluations that can be time-consuming and are typically based on assessment of cognitive performance to identify symptomatic individuals [3]. However, assessment of brain health in asymptomatic individuals and the understanding and quantification of resilience in older adults at-risk for MCI and ADRD is currently difficult to complete.

Of particular interest to ADRD prevention efforts, a large number of modifiable (e.g., exposures, lifestyle, and social habits) and nonmodifiable (e.g., age, sex, genetics) risk factors have been identified [5 , 7–10]. For primary prevention, up to 30% of ADRD cases could potentially be preventable through modification of risk factors and behavioral changes to mitigate the effect of unmodifiable risk factors [9, 10]. Multiple lines of evidence from epidemiological and longitudinal observational studies exist that suggest that the risk of ADRD appears to be reduced in individuals who are physically [11 –13] and cognitively active [14 –16], socially engaged [17, 18], practice mindfulness [19, 20], have higher educational attainment [1], and eat a heart- and brain-healthy diet [21, 22]. There is increasing interest in lifestyle modification [23 –25], integrative medicine [26], and personalized medicine [27] approaches to treat and/or prevent dementia. In terms of secondary and tertiary prevention, these same modifiable risk factors could lessen symptoms, alleviate aspects disease burden, and limit disability [5 , 10].

Another frequently discussed, but difficult to measure risk factor is cognitive reserve. Cognitive reserve is a latent construct that represents an individual’s ability to maintain cognitive functioning despite the presence of underlying neurodegenerative pathology [28, 29]. There is no consensus on determinants of cognitive reserve; however, two pragmatic metrics include educational and occupational attainment [30]. Educational attainment is usually recorded as the number of years of formal schooling. However, the number of years of schooling may not be representative of the quality of the educational experience, and opportunities for advanced education may not be equal across different racial, ethnic, and socioeconomic groups [31 –35]. Occupational attainment may provide a better measure of an individual’s practical cognitive abilities, as job opportunities and advancement may not necessarily match educational attainment [36]. Other potential measures of cognitive reserve include assessment of verbal IQ [28 –30], although this is a time intense assessment requiring expertise in neuropsychological testing and may not be readily available in clinical practice or in certain research situations.

There are a number of existing measures of disease and ADRD risk such as vascular risk factors, but a paucity of instruments to capture brain health and define resilience to developing ADRD. To address the need for a brief, easy-to-score quantifiable measure of brain health and resilience that could be applied in research projects and clinical practice, we created the Resilience Index (RI). The goal was to create an instrument that could capture factors associated with reduced risk of ADRD and resilience to accumulating pathology that would be easy to interpret, provide readily actionable information, and complement the traditional neurocognitive diagnostic evaluation. We hypothesized that 1) individuals with a healthy brain would have greater resilience than those experiencing cognitive impairment, and 2) within diagnostic groups (controls, MCI, ADRD), individuals with greater resilience would have better cognitive and functional performance. We conducted a cross-sectional study examining the psychometric properties RI compared with Gold Standard assessments to determine its utility as a measure of brain health and resilience.

METHODS

Study participants

We evaluated 241 participant-study partner dyads attending our center for clinical care or participating in cognitive aging research. During the visit, the participant and study partner underwent a comprehensive evaluation including the Clinical Dementia Rating (CDR) and its sum of boxes (CDR-SB) [37], physical and neurological examination, assessment of mood, physical performance, neuropsychological testing, and caregiver ratings of patient cognitive abilities, behavior, and function. All components are part of standard of care at our center [15 , 38] with patients and research participants undergoing identical evaluations. A waiver of consent was obtained for retrospective review of individuals attending clinic while research participants provided written informed consent. This study was approved by the University of Miami Institution Review Board.

Participant characteristics

Demographic information including age, sex, education, socioeconomic status, race, ethnicity, past medical history, medications, alcohol, tobacco, and substance use history, co-morbidities, and family history were collected. Socioeconomic status (SES) was calculated with combined education and head-of-household occupation scores using the Hollingshead Two-Factor Index of Social Status [39] and specified as upper, middle, or lower class.

Clinical assessment

The clinical assessments were modelled after the Uniform Dataset Set v3.0 from the National Institute of Aging Alzheimer Disease Research Center program [40, 41]. The CDR [37] was used to determine the presence or absence of dementia and to stage its severity. In most cases, a global CDR 0 indicates no dementia; CDR 0.5 represents MCI or very mild dementia; CDR 1, 2, or 3 correspond to mild, moderate, or severe dementia. The CDR-SB was calculated by adding up the individual CDR categories giving a score from 0–18 with higher scores supporting more severe stages. Extrapyramidal features were assessed with the Movement Disorders Society-Unified Parkinson’s Disease Rating Scale, motor subscale part III (UPDRS) [42]. The Charlson Comorbidity Index [43] and Functional Comorbidity Index (FCI) [44] were used to measure overall health and medical comorbidities. Global physical performance was captured with the mini-Physical Performance Test (mPPT) [45] and frailty was assessed with the Fried Frailty Scale [46]. Vascular contributions to dementia were assessed with the modified Hachinski scale [47] and a modified form of the Cardiovascular Risk Factors, Aging, and Incidence of Dementia scale (mCAIDE) [48].

Study partner ratings of cognition, function, and behavior

Study partners completed the informant version of the Quick Dementia Rating System (QDRS) [49] to provide a global rating of cognitive, functional, and behavioral domains with scores greater than 1.5 supporting cognitive impairment. Activities of daily living were captured with the Functional Activities Questionnaire (FAQ) [50]. Dementia-related behaviors were measured with the Neuropsychiatric Inventory (NPI) [51]. Health-related quality of life was captured with the Health Utilities Index Mark-3 (HUI-3) [52]. The participant’s overall nutritional status was capture with the Mini Nutrition Scale [53].

Cognitive assessment

Each patient was administered a 45-min test battery modeled after the UDSv3.0 battery [41] and supplemented with additional measures. The psychometrist was unaware of the diagnosis, CDR, or RI. Verbal IQ was determined with the Test of Premorbid Function (Pearson Assessments, San Antonio, TX) that tests the individual’s ability to read a list of 70 words with atypical or irregular grapheme to phoneme pronunciations. The Montreal Cognitive Assessment [54] was used for a global cognitive screen. The rest of the battery included: 15-item Multilingual Naming Test (naming) [41]; Animal naming (verbal fluency) [41]; Hopkins Verbal Learning Task (episodic memory for word lists – immediate, delayed, and recognition) [55]; Number forward and backward (working memory) [41]; Trailmaking A and B (attention, processing, and executive function) [56]; and the Number Symbol Coding Task (attention, set switching, and executive function) [30]. A cognitive z-score was computed as a summary of performance across the nine cognitive tests. Mood was assessed with the Hospital Anxiety Depression Scale [57] providing subscale scores for depression (HADS-D) and anxiety (HADS-A). Participants also completed the self-reported version of the QDRS [58] for a self-rating of cognitive abilities with scores greater than 1.5 supporting cognitive impairment.

Determination of consensus research diagnoses

Global rating scales (CDR) were combined with cognitive performance, the neurologic examination, and laboratory tests to assign individuals to the following diagnostic categories after consensus: Cognitively normal controls, MCI [6] or Dementia. Dementia diagnoses were determined using standardized published criteria for Alzheimer’s disease (AD) [59], dementia with Lewy bodies (DLB) [60], vascular contributions to cognitive impairment and dementia (VCID) [61], and frontotemporal degeneration (FTD) [62].

Calculation of resilience index

The Resilience Index was calculated as the sum from the following 6 scales which individually were demonstrated to differ between individuals with normal cognition and individuals with any form of cognitive impairment:

The Cognitive Reserve Unit Scale (CRUS): The CRUS captures standardized measures of 7 levels educational attainment and 7 levels of occupational attainment to estimate cognitive reserve expressed in units with a range of possible scores 0–66 (higher scores representing greater cognitive reserve) [30].

Social Engagement Scale: This scale captures participation in social activities, socialization, and engagement using a Likert scale with anchors Poor, Fair, Good, Excellent with a range of possible scores 1–4 (higher scores representing greater social engagement) [15, 20]

The Quick Physical Activity Rating (QPAR): The QPAR is a measure of intensity, frequency, and duration in 10 categories of physical activity to provide a quantifiable “dose” with a range of possible scores 0–153 (higher scores representing greater physical activity) [63].

The Cognitive & Leisure Activity Scale (CLAS): The CLAS is a measure of frequency of participation in15 categories of cognitively stimulating activities to provide a quantifiable “dose” with a range of possible scores 0–80 (higher scores representing greater cognitive activity) [15].

Mediterranean-DASH Intervention for Neurodegenerative Delay (MIND) Diet Score Sheet: The MIND diet score is a checklist of 15 categories of foods with three frequencies of intake to provide adherence to the principles of the MIND diet with a range of possible scores 0–15 (higher scores representing greater adherence) [15, 20].

Applied Mindfulness Process Scale (AMPS): The AMPS is a measure of the practice of three aspects of mindfulness (decentering, positive emotional regulation, and negative emotional regulation) with a range of possible scores 0–60 (higher scores representing greater mindfulness) [20, 64].

The six subscales are totaled to provide the Resilience Index with a range of possible scores from 1–378 with higher scores representing greater resilience and better brain health.

Apolipoprotein E genotyping

Apolipoprotein E (APOE) genotyping was performed by True Health Diagnostics LLC (Richmond, VA). Six possible allelic combinations were obtained with individuals dichotomized as being APOE4 carriers or non-carriers.

Volumetric MRI

A subset of individuals (n = 55) underwent volumetric MRI with NeuroQuant software (CorTechs Labs, San Diego, CA), an FDA-approved automated quantitative analysis of brain MRI images with normative reference data adjusted for age, sex, and intracranial volume with high correlation to FreeSurfer [65] and visual assessment [66]. Two measures were used in this study as a measure of brain health and a biomarker of neurodegeneration: hippocampal volumes and hippocampal occupancy [67]. While hippocampal volume is often used as a predictor of conversion of MCI to AD, hippocampal occupancy measures the degree of hippocampal atrophy accounting for volume loss and compensatory inferior lateral ventricle expansion. It is calculated as a ratio of hippocampal volume to the sum of the hippocampal and inferior lateral ventricle volumes in each hemisphere separately, which are then averaged and normalized for age and sex. This measure may aid in differentiation of individuals with congenitally small hippocampi from those with small hippocampi due to a degenerative disorder [67].

Statistical analyses

Analyses were conducted with IBM SPSS Statistics v26 (Armonk, NY). Descriptive statistics were used to examine participant demographic characteristics, informant rating scales, health metrics, dementia staging, neuropsychological testing, and resilience factors. One-way analysis of variance (ANOVA) with LSD post-hoc tests were used for continuous data and Chi-square analyses were used for categorical data. To assess RI scale variability, the frequency distribution, range, and standard deviation were calculated, and data were examined for floor and ceiling effects. Kurtosis and skewness statistics were examined to characterize the shape, symmetry, and outliers of the distribution [58, 63]. Kurtosis is a measure of the extent to which there are outliers. For a normal distribution, the value of the kurtosis statistic is zero. Positive kurtosis indicates that the data exhibit more extreme outliers than a normal distribution. Negative kurtosis indicates that the data exhibit fewer extreme outliers than a normal distribution. Skewness is a measure of the asymmetry of a distribution. The normal distribution is symmetric and has a skewness value of 0. A distribution with a significant positive skewness has a long right-sided tail. A distribution with a significant negative skewness has a long left-sided tail. As a guideline, a skewness value more than twice its standard error is taken to indicate a departure from symmetry. Known-group validity was assessed by examining the RI scores and its components by diagnostic groups, specific dementia etiology, sociodemographic variables, APOE genotype, verbal IQ, and CDR stages [58, 63]. Strength of association was assessed comparing the mean performance on participant characteristics, rating scales, and neuropsychological test performance with the RI using Pearson correlation coefficients [58, 63]. Multiple comparisons were addressed using the Bonferroni correction.

Receiver operator characteristic (ROC) curves were used to assess the ability of the RI and its individual components to discriminate between individuals with and without cognitive impairment. Results are reported as area under the curve (AUC) with standard errors and 95% confidence intervals (CIs). A k-means cluster analysis targeting two, three, and five clusters with 10 iterations was performed to determine the risk threshold of the RI. The two-cluster target identified that a score of 143 best separates high resilience from low resilience with respect to the participant’s cognitive status while maximizing the area under the ROC curve. Both the three- and five-cluster targets failed to beat the two-cluster target threshold for the RI.

The major goal of the RI was to identify cognitively normal individuals with high measures of brain health. A secondary goal of the RI was to identify individuals with low measures of brain health to: (a) refer for further cognitive and functional assessment and (b) to provide actionable data for personalized primary, secondary, and tertiary prevention plans. Sensitivity, specificity, predictive values, likelihood ratios, and diagnostic odds ratio for the RI were calculated. Since positive (PPV) and negative (NPV) predictive values are not intrinsic to the RI and are highly dependent on disease prevalence, likelihood ratios were generated which are independent of disease prevalence. The likelihood ratio of any screening test is the probability that a positive test is found in persons with disease divided by the probability of the same finding in persons without disease [68, 69]. Likelihood ratios range from 0 to infinity, with larger numbers providing more convincing evidence of disease; smaller numbers argue that disease is less likely. Ratios close to 1 lack diagnostic value. When the positive likelihood ratio is > 5 or the negative likelihood ratio is < 0.2, the pre-test probability of a patient having the disease tested can be used to estimate a post-test probability of the disease state existing [68, 69]. The diagnostic odds ratio is a measure of the effectiveness of a diagnostic test [70]. It is defined as the ratio of the odds of the test being positive if the subject has the condition of interest relative to the odds of the test being positive if the subject does not have the condition of interest. The diagnostic odds ratio is a single indicator of test performance but is independent of prevalence and presented as an odds ratio, which is familiar to medical practitioners and clinical researchers. For a screening test to classify brain health in cognitively normal individuals who are less likely to have ADRD, sensitivity is preferred over specificity, and negative likelihood ratios are preferred over positive likelihood ratios [68, 71, 72].

A cut-off value providing the best combination of sensitivity and specificity was determined from ROC curves and confirmed by k-cluster analyses. Participant characteristics, rating scales, and neuropsychological test performance were compared by High versus Low RI using generalized linear models controlling for age. Finally, we assessed the ability of the RI to differentiate baseline stages of cognitive status using mean scores with standard deviations and 95% confidence intervals to provide risk profiles for healthy controls, very mild impairment, and mild impairment.

RESULTS

Sample characteristics

The participants had a mean age of 74.6±9.5 years (range 38–98 years) with a mean education of 15.8±2.6 years (range 8–20 years) of education, and 45.2% were female (Table 1). APOE4 carriers comprised 38.4% of the cohort with only 1 individual being homozygous for ɛ4/ɛ4. The sample was 92.5% non-Hispanic White, 2.5% African American, and 4.6% Hispanic. The patients had a mean CDR-SB of 4.2±4.2 (range 0–18) and a mean MoCA score of 19.5±6.7 (range 2–30). The sample covered a range of CDR 0 (n = 40), very mild impairment (CDR 0.5 = 98), mild dementia (CDR 1 = 55), moderate dementia (CDR 2 = 39), and severe dementia (CDR 3 = 9). Consensus diagnoses are 39 cognitively normal controls, 75 MCI, and 127 dementia cases with one CDR 0 case representing a non-AD cause of MCI. The dementia cases were further classified using published criteria as 46 AD, 64 DLB, 10 VCID, and 7 FTD cases. The mean RI score was 138.2±35.6, with a median of 138.0 and mode of 138.5. The minimum score was 64.5 (floor effect: 0%) and maximum score was 257 (ceiling effect: 0%) covering a wide range of possible scores. Distribution statistics showed skewness was 0.43 (standard error = 0.16) and kurtosis was 0.30 (standard error = 0.31) supporting that the RI follows a normal distribution (Fig. 1).

Table 1

Sample characteristics

	Control (n = 40)	MCI (n = 74)	Dementia (n = 127)	p
Age, y	67.3 (10.3)	73.1 (9.5)	77.8 (7.7)	<0.001
Education, y	16.1 (2.2)	16.0 (2.6)	15.5 (2.6)	0.272
Sex, % Female	66.7	42.7	40.2	0.013
APOE4, %	38.2	36.0	40.7	0.884
QDRS-Informant	0.7 (1.1)	3.0 (2.7)	8.5 (4.7)	<0.001
FAQ	0.1 (0.5)	2.8 (4.1)	14.6 (8.8)	<0.001
HUI-3	0.88 (0.2)	0.65 (0.2)	0.41 (0.3)	<0.001
Mini Nutrition Scale	12.9 (1.2)	11.7 (2.1)	10.7 (2.0)	<0.001
NPI	1.5 (2.0)	5.2 (4.8)	8.6 (5.7)	<0.001
mPPT	13.0 (2.1)	11.1 (2.9)	9.1 (3.1)	<0.001
QDRS – Patient	0.5 (1.0)	2.6 (2.6)	6.2 (4.6)	<0.001
MoCA	26.4 (2.7)	23.4 (2.9)	15.1 (5.9)	<0.001
z-Score	1.11 (0.4)	0.37 (0.6)	-0.85 (0.7)	<0.001
UPDRS-III	3.1 (3.9)	6.3 (8.8)	15.1 (14.9)	<0.001
Charlson	1.3 (1.6)	2.4 (1.5)	2.7 (1.7)	<0.001
FCI	2.7 (1.7)	4.2 (1.9)	4.1 (1.6)	<0.001
Hachinski	0.5 (0.7)	0.9 (1.4)	1.1 (1.4)	0.081
mCAIDE	5.2 (3.2)	7.6 (2.8)	8.2 (2.5)	<0.001
Fried Frailty	0.8 (1.0)	1.8 (1.3)	2.8 (1.3)	<0.001
CDR-SB	0.1 (0.2)	1.3 (0.7)	7.1 (3.8)	<0.001

Mean (SD) or %. QDRS, Quick Dementia Rating System; FAQ, Functional Activities Questionnaire; HUI-3, Health Utilities Index-Mark3; NPI, Neuropsychiatric Inventory; mPPT, Mini Physical Performance Test; MoCA, Montreal Cognitive Assessment; UPDRS-III, Unified Parkinson’s Disease Rating Scale-Motor Subscale; FCI, Functional Comorbidity Index; mCAIDE, Modified Cardiovascular Risk Factors, Aging, and Incidence of Dementia; CDR-SB, Clinical Dementia Rating Sum of Boxes.

Fig. 1

Histogram of Resilience Index. This histogram demonstrates that the Resilience Index follows a normal distribution with a mean of 138.2, standard deviation of 35.6, a median of 138.0, and a mode of 138.5.

Fig. 2

Conceptual Model of Resilience Index. This conceptual model depicts how the six components of the Resilience Index (RI) contribute to the quantification of resilience and provides a measurement of brain health. Cognitive Reserve is an early- to mid-life contribution taking into account education and occupational attainment. Mid- to late-life resilience factors include physical and cognitive activity, social engagement, dietary patterns, and mindfulness. In addition to factors that make the brain more resilient, there are risk factors that increase vulnerability (genetics, social determinants of health, comorbid medical conditions). Understanding brain health and modifiable resilience factors can help develop and tailor personalized prevention plans for cognitive normal individuals (primary prevention), and those with MCI (secondary prevention) and ADRD (tertiary prevention).

Resilience index scores by diagnostic groups

The performance of the RI and its individual components were first compared across diagnostic groups (controls, MCI, dementia) in Table 2. Mean RI scores were different between all three diagnostic groups as were social engagement, QPAR, CLAS, and AMPS scores (all p-values < 0.001). MIND scores between controls and MCI were not different but were lower in dementia cases. Cognitive reserve scores were numerically lower in dementia groups but did not reach statistical significance. The performance of the RI and its individual components were then compared across different dementia etiologies (AD, DLB, VCID, FTD) in Table 3. Participants with VCID tended to have the lowest scores in total RI, cognitive reserve, physical activity, and mindfulness. The performance of the RI and its individual components were then compared across different dementia stages using the CDR in Table 4. As expected, the RI was highest in CDR 0 individuals and decreased across dementia stages. The individual components had varied patterns of decrements across dementia stages with MIND diet not different between CDR 0 and CDR 0.5, while social engagement, physical and cognitive activities, and mindfulness were different between CDR 0 and CDR 0.5 but not different between CDR 1–3. Cognitive reserve was only different in CDR 3 individuals.

Table 2

Resilience index and components by diagnostic group

	Control (n = 39)	MCI (n = 75)	Dementia (n = 127)	p
Cognitive Reserve Units	45.1 (12.5)	44.6 (14.2)	41.4 (14.8)	0.189
Social Engagement	3.3 (0.6)	2.9 (0.9)	2.4 (0.8)	< 0.001 ^b
QPAR	40.3 (23.1)	24.1 (17.5)	14.8 (13.1)	< 0.001 ^b
CLAS	31.7 (10.7)	25.2 (9.1)	21.3 (7.9)	< 0.001 ^b
MIND	9.6 (2.1)	9.3 (2.0)	8.3 (2.1)	< 0.001 ^a
AMPS	43.9 (10.9)	39.4 (10.8)	34.6 (12.2)	< 0.001 ^b
Resilience Index	174.0 (31.1)	145.6 (29.3)	122.8 (30.8)	< 0.001 ^b

Table 3

Resilience index and components by dementia etiology

	AD (n = 46)	DLB (n = 64)	VCID (n = 10)	FTD (n = 7)	p
Cognitive Reserve Units	40.9 (16.3)	43.9 (12.8)	26.9 (17.0)	43.4 (7.7)	0.007 ^a
Social Engagement	2.5 (0.9)	2.3 (0.9)	2.7 (0.7)	2.3 (0.7)	0.382
QPAR	16.5 (13.9)	13.3 (12.3)	11.1 (11.6)	24.4 (14.6)	0.104^b
CLAS	22.6 (8.9)	19.9 (6.4)	20.5 (11.3)	25.9 (7.2)	0.137
MIND	7.9 (1.9)	8.5 (2.3)	8.0 (2.1)	8.7 (1.6)	0.500
AMPS	35.8 (13.1)	34.3 (11.1)	27.6 (12.3)	39.4 (12.9)	0.180^c
Resilience Index	126.2 (33.4)	122.4 (27.4)	96.8 (31.7)	144.1 (19.8)	0.010^a

Table 4

Resilience index and components by CDR stages

	CDR 0 (n = 40)	CDR 0.5 (n = 98)	CDR 1 (n = 55)	CDR 2 (n = 39)	CDR 3 (n = 9)	p
Cognitive Reserve Units	45.7 (12.8)	44.0 (13.5)	39.3 (15.4)	46.1 (13.3)	28.9 (17.6)	0.003 ^a
Social Engagement	3.3 (0.6)	2.9 (0.9)	2.4 (0.8)	2.2 (0.9)	2.3 (0.7)	< 0.001 ^b
QPAR	40.4 (22.8)	23.0 (17.2)	16.2 (12.2)	11.3 (11.1)	5.7 (7.0)	< 0.001 ^b
CLAS	31.7 (10.6)	25.3 (8.8)	20.5 (6.0)	20.3 (9.0)	19.0 (9.8)	< 0.001 ^b
MIND	9.6 (2.1)	9.1 (2.0)	8.2 (2.2)	8.3 (2.0)	6.9 (2.0)	< 0.001 ^c
AMPS	44.3 (11.1)	38.9 (10.8)	33.2 (12.2)	34.1 (11.9)	34.6 (13.7)	< 0.001 ^b
Resilience Index	175.0 (31.3)	143.4 (28.7)	120.1 (29.5)	122.3 (30.0)	97.3 (25.4)	< 0.001 ^d

Mean (SD). Bold signifies significance after correction for multiple comparisons (corrected p-value < 0.0083). QPAR, Quick Physical Activity Rating; CLAS, Cognitive & Leisure Activity Scale; MIND, Mediterranean-DASH Intervention for Neurodegenerative Delay; AMPS, Applied Mindfulness Process Scale. Post-hoc comparisons: ^aCDR 3 different from other stages; ^bCDR 0 different from CDR 0.5-3, CDR 0.5 different from CDR 1-3; CDR 1-3 not different from each other; ^c CDR 0 and 0.5 not different from each other but different from CDR 1-3, CDR 1-3 not different from each other; ^dCDR 0 different from CDR 0.5-3, CDR 0.5 different from CDR 1-3, CDR 1-2 different from CDR 3.

Resilience Index scores by sociodemographic variables

We examined differences in RI scores by sex, age strata (<60, 60–69, 70–79, 80+ years), socioeconomic status tertiles (upper, middle, lower), education (≤12, 13–16, > 16 years), APOE4 carrier status (carrier, non-carrier), and verbal IQ tertiles (< 100, 100–120, > 120) (Table 5). There was no difference in RI between men and women but there were differences by sex in individual components. Men had higher mean cognitive reserve, while women had higher mean social engagement and cognitive activities. When examined by age strata, individuals 80+ years had lower RI scores as well as lower physical activity and MIND diet scores compared with other groups. When examined by education, cognitive reserve units and RI scores were highest in individuals with the highest educational level. There were no differences in RI or its individual components by APOE4 carrier status. When examined by socioeconomic status tertiles as determined by the Hollingshead Index, RI scores were highest in individuals in the highest tertile while physical activity scores were lowest in individuals in the middle tertile. Cognitive reserve units decreased across SES tertiles. Last, we examined RI and its individual components by Verbal IQ tertiles. Total RI scores, cognitive reserve units, and cognitive leisure activities were highest in individuals with Verbal IQ scores > 120. Interestingly, MIND scores were lowest in individuals with Verbal IQ scores > 120 although the significance was lost when correcting for multiple comparisons. There were not sufficient Hispanic and African American individuals in the sample to make meaningful comparisons.

Table 5

Resilience index and components by sex, age, socioeconomic status, and verbal IQ

	Sex			Age
	Male	Female	p	<60 y	60–69 y	70–79 y	80+ y	p
Cognitive Reserve Units	48.2 (12.4)	36.8 (14.1)	< 0.001	42.8 (10.6)	42.0 (14.4)	43.6 (12.4)	42.9 (17.4)	0.949
Social Engagement	2.6 (0.9)	2.9 (0.8)	0.002	2.7 (0.9)	2.9 (0.8)	2.8 (0.9)	2.6 (0.9)	0.180
QPAR	19.8 (17.7)	24.3 (19.9)	0.064	31.3 (19.0)	29.1 (20.0)	24.1 (19.9)	12.9 (12.1)	< 0.001 ^a
CLAS	22.1 (7.9)	26.8 (10.6)	< 0.001	24.2 (9.4)	26.5 (7.7)	26.5 (7.7)	22.5 (9.1)	0.168
MIND	8.8 (2.1)	8.8 (2.2)	0.999	9.3 (1.8)	9.4 (2.4)	8.9 (2.1)	8.2 (2.1)	0.014^a
AMPS	36.5 (11.7)	38.8 (12.4)	0.140	41.3 (8.9)	36.8 (11.8)	38.7 (12.2)	35.6 (12.5)	0.170
Resilience Index	137.9 (33.2)	138.5 (38.4)	0.902	151.6 (34.2)	146.7 (39.1)	142.8 (33.5)	124.7 (33.3)	< 0.001 ^a
	Education				APOE4
	≤12 y	13–16 y	> 16 y	p	Non-carrier	Carrier	p
Cognitive Reserve Units	26.5 (10.0)	40.3 (10.2)	56.3 (7.7)	< 0.001 ^a	43.0 (14.3)	42.9 (12.9)	0.980
Social Engagement	2.7 (0.8)	2.7 (0.9)	2.8 (0.9)	0.442	2.9 (0.9)	2.8 (0.9)	0.660
QPAR	21.3 (23.6)	21.2 (16.9)	22.9 (18.1)	0.795	25.5 (20.2)	28.1 (23.5)	0.481
CLAS	21.5 (10.3)	24.2 (9.9)	25.9 (8.2)	0.037^b	25.1 (8.6)	27.5 (10.7)	0.166
MIND	8.3 (2.2)	8.6 (2.1)	9.3 (2.1)	0.014^b	8.9 (2.2)	8.9 (2.2)	0.971
AMPS	35.6 (11.8)	37.3 (12.8)	39.2 (12.0)	0.243	37.5 (12.0)	40.4 (12.4)	0.176
Resilience Index	115.7 (35.4)	134.4 (33.5)	156.5 (29.0)	< 0.001 ^a	143.0 (34.3)	150.8 (37.8)	0.216
	SES Class				Verbal IQ
	Upper	Middle	Lower	p	<100	100–120	> 120	p
Cognitive Reserve Units	51.2 (12.6)	39.3 (12.4)	26.7 (9.8)	< 0.001 ^a	39.0 (8.2)	43.6 (12.4)	51.9 (10.3)	0.004 ^a
Social Engagement	3.0 (0.8)	2.7 (0.9)	3.0 (0.8)	0.149	2.8 (1.1)	2.7 (0.9)	2.9 (0.8)	0.755
QPAR	28.8 (21.5)	21.7 (15.8)	34.3 (35.4)	0.043^b	21.2 (9.9)	22.5 (13.2)	29.1 (24.8)	0.386
CLAS	27.7 (8.1)	24.5 (9.4)	28.0 (14.9)	0.152	18.9 (6.4)	24.4 (8.6)	27.3 (7.6)	0.038^b
MIND	9.2 (2.2)	8.8 (2.0)	8.4 (2.3)	0.393	9.1 (2.0)	9.7 (1.9)	8.3 (1.6)	0.046^c
AMPS	37.9 (12.8)	38.5 (12.1)	38.1 (12.1)	0.879	33.6 (16.3)	39.6 (9.6)	43.3 (9.1)	0.084
Resilience Index	157.9 (35.9)	135.6 (31.8)	137.3 (47.5)	0.004 ^c	124.5 (26.5)	142.5 (27.7)	162.9 (35.7)	0.008 ^a

Mean (SD). Bold signifies significance after correction for multiple comparisons (corrected p-value < 0.0083). QPAR, Quick Physical Activity Rating; CLAS = Cognitive & Leisure Activity Scale; MIND, Mediterranean-DASH Intervention for Neurodegenerative Delay; AMPS, Applied Mindfulness Process Scale; SES, Socioeconomic Status. Post-hoc comparisons by age: ^a80+ significantly different than other age groups. Post-hoc comparisons by education: ^aAll groups different from each other; ^b≤12 y different from > 16 y. Post-hoc comparisons by SES: ^aAll groups different from each other; ^bMiddle class different from Lower class; ^cUpper class different from Middle class. Post-hoc comparisons by verbal IQ: ^a> 120 different from other groups; ^b> 120 different than < 100; ^c> 120 different than 100–120.

Strength of association between Resilience Index, clinical, and cognitive measures

Construct validity is demonstrated in Table 6 by examining the strength of association between the RI and clinical, functional, behavioral, informant ratings, and neuropsychological testing. The RI showed moderate-to-strong correlations with clinical measures but most strongly correlated (all ps < 0.001) with informant QDRS (R = –0.528), HUI-3 (R =0.525), FAQ (R = –0.513), Fried frailty scale (R =–0.489), patient QDRS (R = –0.478), and CDR-SB (R = –0.475). The RI showed moderate correlation with neuropsychological test scores but most strongly correlated (all ps < 0.001) with the cognitive z-score (R = 0.534), HVLT immediate recall (R = 0.525) and delay recall (R = 0.497) and verbal IQ (R = 0.487).

Table 6

Strength of association between resilience index, demographics, rating scales, and cognitive performance

Rating scales			Cognitive tests
Variable	R	p	Variable	R	p
Age	–0.246	< 0.001	MoCA	0.467	< 0.001
Education	0.418	< 0.001	Numbers Forward	0.180	0.005
QDRS-Informant	–0.528	< 0.001	Numbers Backward	0.378	< 0.001
FAQ	–0.513	< 0.001	Trailmaking A	–0.342	< 0.001
HUI-3	0.525	< 0.001	Trailmaking B	–0.429	< 0.001
Mini Nutrition Scale	0.340	< 0.001	HVLT – Immediate	0.525	< 0.001
NPI	–0.412	< 0.001	HVLT – Delay	0.497	< 0.001
MPPT	0.452	< 0.001	HVLT – Recognition	0.323	< 0.001
QDRS – Patient	–0.478	< 0.001	Number Symbol Coding	0.455	< 0.001
UPDRS-III	–0.245	< 0.001	Animal Naming	0.474	< 0.001
Charlson	–0.136	0.035	MINT	0.258	< 0.001
FCI	–0.204	0.003	TOPF	0.487	< 0.001
Hachinski	–0.210	0.001	z-Score	0.534	< 0.001
mCAIDE	–0.335	< 0.001	HADS-Anxiety	–0.286	0.006
Fried Frailty	–0.489	< 0.001	HADS-Depression	–0.314	< 0.001
CDR-SB	-0.475	< 0.001

R-value; p-value. Bold signifies significance after correction for multiple comparisons (corrected p-value < 0.0031). QDRS, Quick Dementia Rating System; FAQ, Functional Activities Questionnaire; HUI-3, Health Utilities Index-Mark3; NPI, Neuropsychiatric Inventory; mPPT, Mini Physical Performance Test; UPDRS-III, Unified Parkinson’s Disease Rating Scale-Motor Subscale; FCI, Functional Comorbidity Index; mCAIDE, Modified Cardiovascular Risk Factors, Aging, and Incidence of Dementia; HADS, Hospital Anxiety and Depression Scale; MoCA, Montreal Cognitive Assessment; HVLT, Hopkins Verbal Learning Test; MINT, Multilingual Naming Test; TOPF, Test of Premorbid Function; CDR-SB, Clinical Dementia Rating Sum of Boxes.

Discriminability of the Resilience Index

We tested the ability of the RI and its individual components to provide estimates of brain health by discriminating between individuals with and without cognitive impairment using ROC analyses to provide area under the curve (AUC) (Table 7). The RI provided very good discrimination (AUC: 0.836; 95% CI: 0.774–0.897) between cognitive normal controls and those with any form of cognitive impairment. Five of six individual components also provided discrimination with the QPAR providing the best discrimination (AUC: 0.827; 95% CI: 0.759–0.895). Cognitive reserve units alone did not discriminate. In Table 8, we compared RI cut-off scores with consensus diagnosis of non-impaired versus impaired (grouping together MCI and dementia). A cut-off score of 143 provided the best combination of sensitivity (0.821) and specificity (0.658). The positive likelihood ratio was 2.4 while the negative likelihood ratio was 0.27 with a diagnostic odds ratio of 8.9. The RI correctly identified 82% of cognitively normal individuals with excellent accuracy (NPV = 95%). When used for screening asymptomatic individuals for brain health, a negative result (i.e., an RI > 143) suggests a nearly 9-fold lower risk of having MCI or dementia. For asymptomatic individuals with low RI (i.e., an RI < 143), a further workup including screening for cognitive, functional, and mood impairments could be recommended, and actionable treatment goals can be developed for preserving brain health.

Table 7

Discriminability of the resilience index and its components

Test result variable(s)	Area	Std. Error	p	95% Confidence interval
				Lower bound	Upper bound
Cognitive reserve units	0.566	0.046	0.155	0.475	0.657
Social engagement	0.715	0.040	0.000	0.636	0.793
QPAR	0.827	0.035	0.000	0.759	0.895
CLAS	0.742	0.044	0.000	0.655	0.829
MIND	0.622	0.049	0.012	0.527	0.718
AMPS	0.673	0.046	0.000	0.582	0.764
Resilience index	0.836	0.031	0.000	0.774	0.897

QPAR, Quick Physical Activity Rating; CLAS, Cognitive & Leisure Activity Scale; MIND, Mediterranean-DASH Intervention for Neurodegenerative Delay; AMPS, Applied Mindfulness Process Scale.

Table 8

Resilience index properties for diagnostic testing

	Not impaired	Impaired		TOTAL
	Controls	MCI	Dementia
High resilience	32	40	29	101
Low resilience	7	35	98	140
TOTAL	39	75	127	241

Resilience Index cut-off = 143. MCI, mild cognitive impairment.

We then compared individuals with high and low RI across all clinical and cognitive measures in Table 9. As individuals with low RI were older than individuals with high RI, we adjusted analyses controlling for age. Since education is part of the CRUS, no further adjustment was performed. Individuals with high RI were significantly different from individuals with low RI (adjusted p-values < 0.001) in informant and participant rated QDRS, CDR-SB, measures of function, behavior and health-related quality of life, nutritional status, physical performance and frailty, vascular risk factors, mood, and most cognitive domains.

Table 9

Comparison of clinical and cognitive variables by resilience index cut-off

Variable	High resilience (n = 101)	Low resilience (n = 140)	Adjusted p^*
Clinical Measures
Age, y	71.7 (9.4)	76.7 (9.0)	<0.001
Education, y	16.8 (2.4)	15.0 (2.4)	< 0.001
Sex, % Female	48.5	42.9	0.384
APOE4 Carriers, %	44.3	32.4	0.150
QDRS-Informant	2.6 (2.9)	7.6 (5.1)	< 0.001
FAQ	3.8 (6.5)	12.0 (9.5)	< 0.001
HUI-3	0.72 (0.2)	0.44 (0.3)	< 0.001
Mini Nutrition Scale	12.0 (1.9)	10.9 (2.1)	< 0.001
NPI	4.0 (4.3)	8.0 (5.9)	< 0.001
mPPT	11.9 (2.7)	9.3 (3.2)	< 0.001
QDRS – Patient	2.3 (2.8)	5.5 (4.7)	< 0.001
UPDRS-III	6.3 (9.6)	13.4 (14.3)	< 0.001
Charlson	2.0 (1.7)	2.6 (1.6)	0.169
FCI	3.6 (1.8)	4.2 (1.8)	0.131
Hachinski	0.6 (0.9)	1.2 (1.5)	0.016
mCAIDE	6.5 (2.9)	8.2 (2.7)	0.012
Fried Frailty	1.4 (1.3)	2.7 (1.3)	< 0.001
HADS-Anxiety	4.9 (3.2)	6.1 (3.9)	< 0.001
HADS-Depression	5.2 (3.5)	6.9 (3.9)	< 0.001
Cognitive Measures
MoCA	22.8 (5.5)	17.2 (6.5)	< 0.001
Numbers Forward	7.0 (1.6)	6.5 (1.5)	0.031
Numbers Backward	4.9 (1.6)	3.9 (1.6)	< 0.001
Trailmaking A	42.7 (29.7)	73.0 (48.9)	< 0.001
Trailmaking B	99.6 (46.4)	134.9 (49.6)	< 0.001
HVLT – Immediate	19.0 (6.6)	12.1 (5.9)	< 0.001
HVLT – Delay	6.1 (3.8)	2.7 (2.9)	< 0.001
HVLT – Recognition	9.9 (2.9)	8.2 (3.3)	0.003
Number Symbol Coding	36.6 (12.8)	25.4 (12.3)	0.002
Animal Naming	17.3 (5.5)	11.3 (5.7)	< 0.001
MINT	14.3 (1.9)	13.1 (3.2)	0.016
TOPF	115.8 (10.4)	107.7 (17.2)	0.029
z-Score	0.51 (0.8)	-0.39 (0.9)	< 0.001
CDR-SB	2.2 (2.9)	5.6 (4.3)	< 0.001

Mean (SD) or %. Bold signifies significance after correction for multiple comparisons (corrected p-value < 0.0026). ^*All values (except for age) are adjusted for age. QDRS, Quick Dementia Rating System; FAQ, Functional Activities Questionnaire; HUI-3, Health Utilities Index-Mark3; NPI, Neuropsychiatric Inventory; mPPT, Mini Physical Performance Test; UPDRS-III, Unified Parkinson’s Disease Rating Scale-Motor Subscale; FCI, Functional Comorbidity Index; mCAIDE, Modified Cardiovascular Risk Factors, Aging, and Incidence of Dementia; HADS, Hospital Anxiety and Depression Scale; MoCA, Montreal Cognitive Assessment; HVLT, Hopkins Verbal Learning Test; MINT, Multilingual Naming Test; TOPF, Test of Premorbid Function; CDR-SB, Clinical Dementia Rating Sum of Boxes.

We then tested for within group differences in cognitive and functional outcomes by comparing individuals with high versus low RI separately in controls, MCI and ADRD individuals. As individuals with moderate-to-severe ADRD almost exclusively had low RI scores, we included only mild ADRD individuals for this analysis. In cognitively normal controls, global and functional differences were detected between those with high RI compared with those with low RI: informant QDRS (0.5±0.9 versus 1.6±1.3, p = 0.012), FAQ (0.03±0.2 versus 0.6±1.1, p = 0.011), NPI (1.1±1.8 versus 3.3±2.4, p = 0.008), mPPT (13.5±1.3 versus 10.7±3.4, p = 0.001), and Charlson comorbidity score (1.0±1.5 versus 2.4±1.9, p = 0.034). Performance on the Hopkins total recall was also greater in controls with high RI (25.5±3.3 versus 21.3±3.6, p = 0.004). In MCI, ratings in the informant QDRS (2.2±2.1 versus 3.8±3.1, p = 0.010) and patient QDRS (1.9±2.2 versus 3.3±2.9, p = 0.017) were better in individuals with high RI. Performance in the MoCA (24.1±2.7 versus 22.5±2.8, p = 0.021), Trailmaking B (94.2±41.4 versus 113.9±42.8, p = 0.047), animal naming (19.0±4.0 versus 15.3±3.5, p < 0.001), and cognitive z-score (0.58±0.53 versus 0.11±0.51, p < 0.001) was greater, and fewer depressive symptoms were reported on the HADS-D (5.4± 3.8 versus 7.9±4.1, p = 0.007) in high RI individuals. In persons with mild dementia, significant differences in global and functional scales were present across multiple domains in individuals with high RI: informant QDRS (4.7±2.5 versus 7.4±3.9, p = 0.003), FAQ (11.7±7.1 versus 6.7±6.2, p = 0.005), HUI-3 (0.65±0.2 versus 0.43±0.3, p = 0.002), Mini Nutrition Scale (11.6±1.5 versus 10.5±1.9, p = 0.024), mPPT (11.2±2.8 versus 9.5±2.9, p = 0.019), and Fried frailty scores (1.8±1.4 versus 2.7±1.2, p = 0.009). Performance in Hopkins total recall (13.7±3.8 versus 11.1±4.7, p = 0.023) and animal naming (13.1±4.3 versus 10.5±4.9, p = 0.034) was greater in individual with mild dementia who had high RI.

Applying Resilience Index risk profiles in a clinical setting

Lastly, to provide a framework for utilizing the RI in a clinical setting revolving around brain health and determining baseline risk of cognitive impairment, we developed a profile of scores by level of impairment based on consensus clinical diagnoses from No Impairment to Mild Impairment (Table 10). Mean RI scores and 95% confidence intervals are shown with corresponding global staging by CDR, participant characteristics, and MRI (in the 55 individuals who had imaging). Decreasing RI scores correspond to older age, lower MoCA scores, higher FAQ, NPI, CDR-SB, smaller hippocampal volumes, and smaller hippocampal occupancy scores.

Table 10

Clinical profiles to guide interpretation of the resilience index

	No Impairment	Very Mild Impairment	Mild Impairment	p
RI Mean (SD)	175.0 (31.3)	143.4 (28.7)	120.1 (29.7)	<0.001
RI 95% CI	164.9–185.0	137.7–149.2	112.1–128.1	<0.001
Corresponds to CDR	0	0.5	1
Participant Characteristics
Age	67.6 (10.3)	74.2 (9.1)	75.9 (8.1)	< 0.001
MoCA	26.5 (2.7)	22.0 (3.7)	17.7 (5.1)	< 0.001
CDR-SB	0.1 (0.2)	1.7 (1.1)	5.4 (1.3)	< 0.001
FAQ	0.1 (0.5)	3.7 (4.9)	12.1 (6.7)	< 0.001
NPI	1.4 (2.0)	5.3 (4.5)	8.7 (5.7)	< 0.001
Hippocampal Volume (cm³)	7.6 (1.4)	6.1 (1.2)	5.7 (1.1)	0.003
Hippocampal Occupancy Score	0.76 (0.12)	0.66 (0.12)	0.58 (0.09)	0.007

Mean (SD). Bold signifies significance after correction for multiple comparisons (corrected p-value < 0.0071). RI, Resilience Index; CDR, Clinical Dementia Rating; CDR-SB, CDR Sum of Boxes; MoCA, Montreal Cognitive Assessment; FAQ, Functional Activities Questionnaire; NPI, Neuropsychiatric Inventory.

DISCUSSION

The RI is a brief assessment of brain health that incorporates six modifiable protective or resilient factors against ADRD: cognitive reserve, social engagement, physical activity, cognitive and leisure activities, dietary patterns, and mindfulness. Each individual component scale of the RI takes ∼3 min to complete with most cognitively normal controls, MCI, and mild dementia being able to complete the forms themselves, while other individuals with mild through severe dementia may require assistance. There was very good data quality with a normal distribution, a range of scores encompassing the majority of possible scores, and no floor and ceiling effects. The RI performed well across different participant characteristics, stages of cognitive abilities, and ADRD etiologies. The RI worked equally well in women and men with mild differences in several components and was not subject to effects of APOE4 carrier status. Individuals with high education, SES and IQ scored highest on the RI and the RI worked well across different age strata with only individuals over age 80 showing differences. The RI showed moderate correlations with clinical and cognitive measures suggesting that the RI captures domains other than cognition and can be used in a complementary fashion. Individuals with high RI scores (> 143) had better health, functional, behavioral, cognitive, and quality of life ratings than individuals with low scores.

The RI maximal potential could be in assessment on brain health and resisting the full impact of causes of cognitive impairment among asymptomatic mid- and late-life adults. Our study supports that the RI could be used in clinical and research situations to assess brain health in older adults; this was accomplished by a simple assessment of validated brain resilience factors that nearly all adults should be able to complete. The sensitivity, diagnostic ratio, and negative predictive value suggest that asymptomatic individuals with high resilience are less likely to have an underlying neurodegenerative disease. In the absence of specific recommendations about dementia screening, using a risk analysis such as the RI could guide use of limited resources to stratify screening evaluations, first focusing on individuals with low RI. Asymptomatic individuals with low resilience should be further evaluated with brief cognitive screening tests (e.g., MoCA, Mini-Mental State Exam), a screen for a mood disturbance, a physical exam, and labs to rule out false positives and prepare actionable treatment recommendations. Asymptomatic individuals with high RI could be further evaluated based on other risk factors such as age, family history, or comorbid medical conditions. Symptomatic individuals would also benefit from completing the RI as it could provide additional information on staging and provide actionable targets for intervention.

Results from the RI could provide clinicians and researchers with a guide to develop personalized prevention plans to support brain health (Fig. 2). Primary, secondary, and tertiary dementia prevention strategies could include recommendations for health promotion and healthful behavioral change while also addressing aspects that increase vulnerability such as risk reduction and aggressive management of co-morbid disease [5 , 73]. While primary prevention strategies are by definition limited to individuals without symptoms or evidence of disease [10] and thus harder to initiate, secondary and tertiary prevention can be taken advantage of with the RI. Secondary prevention entails engaging individuals in activities that can potentially reduce prevalence and symptom presentation when disease pathology is already established [5, 10]. Potential target groups are preclinical ADRD, individuals with subjective cognitive impairment and early MCI cases. Tertiary prevention, on the other hand, involves activities for individuals who are already symptomatic with disease (e.g., late MCI, ADRD) to reduce disability and disablement [5, 10]. We demonstrate in within group analyses that control (Primary, Secondary), MCI (Secondary, Tertiary), and ADRD (Tertiary) individuals with high RI have better cognitive, functional, and global ratings than individuals with low RI. Examination of the individual components of the RI could provide a platform for personalizing prevention plans across the spectrum of cognitive aging and neurodegenerative disease.

Collectively, our findings indicate the importance of considering baseline resilience on the observed reporting of cognition, function, and behavior by participants and study partners, and for participant cognitive performance. The process of resilience in dementia and brain health research could be viewed as a collection of activities that occur across the lifespan. Individuals with greater investments in brain health (e.g., performing resilience activities) that may help to maintain brain structure and function could translate to a greater resistance to the effects of accumulations of amyloid, tau, and neurodegeneration pathologies later in life [74] even if already symptomatic. While we cannot make predictions in a cross-sectional sample, within-group analysis demonstrated that individuals with high RI scores had better cognitive, functional, and global scores than individuals with low RI scores—this was true for controls, MCI, and mild ADRD. If resilience is the ability of the brain to resist the effects of pathology, then individuals with high RI should perform better than individual with low RI, regardless of their cognitive status. Our clinical and research goals were to establish what resilience-type activities people do, how often they do them, and if they are open to the concept of lifestyle modification prior to offering any specific intervention. Greater self-reported cognitive complaints (i.e., participant QDRS), study partner complaints (i.e., informant QDRS), participant mood complaints (i.e., HADS) and participant cognitive performance correlated with measures of brain health and resilience. These findings suggest that the RI may be a useful baseline predictor of brain health and cognitive status in older adults. The RI could serve as an indicator of responses on cognitive, functional, and behavioral outcomes tools that typically are longer to administer and, in some cases, require additional training and certification. The RI and its individual components could serve as targets for interventions to prevent dementia in cognitively normal controls, reduce symptoms, and slow progression and/or conversion of cognitive impairment in individuals with MCI by promoting adoption of positive lifestyle changes, and lessen symptom burden in individuals with ADRD. In addition, our results support that multimodal approaches for lifestyle factors that are focused on brain resilience could serve as targets for intervention in primary, secondary, and tertiary prevention studies.

There is a growing body of literature regarding studies of preclinical Alzheimer’s disease where people with amyloid, tau, and neurodegeneration pathologies may live without overt clinical evidence of disease, as well as how some people living with dementia may progress slower and have better outcomes compare with other individuals who experience more symptomatology. Accumulating evidence from observational studies suggests that modifiable factors such as diet, physical, and cognitive activities, and social engagement may mediate cognitive decline and possible render individuals resistant to manifesting clinical expression of pathology [75, 76]. Education, occupation, leisure, and social activities are frequently used as measures of resilience [77]. Although some risk factors are not modifiable (e.g., age, APOE genotype), modifiable factors may elicit protective mechanisms of resilience to age-associated neurodegenerative disease [75]. Public health campaigns promoting healthy lifestyles and socially engaging activities might provide avenues to improve brain health and wellbeing in mid- and late-life adults. These same strategies are being employed in a number of dementia prevention trials [8 , 23–25]. Although readouts from these trials are not yet known, research into resilience and prevention is critical to better understanding the results of ongoing trials and in the design of new studies.

The biological underpinnings of resilience are an active area of investigation. Inflammation and microglial activation [75], endocytosis pathways [78], and interactions with non-Alzheimer pathologies (e.g., vascular, TDP-43) [74, 79] have been associated with greater resilience. In a recent genome-wide association study of 5,108 individuals, resilience was linked to chromosome 18 upstream of ATP8B1 with the top variant of this locus associated with methylation in prefrontal cortex [80]. In a recent proteome-wide association study of dorsolateral prefrontal cortex from 391 individuals, eight cortical proteins were identified [81]. In a study of 126 older adults with and without cognitive impairment, lower cognitive reserve (combined education and vocabulary scores) predicted less resilience to amyloid pathology [82].

A number of studies have attempted to quantify resilience. The Alzheimer’s Disease Cognitive Resilience Score (AD-CR) was developed as individual-level quantification using postmortem pathology [83] and was correlated with cognitive function, education, personality, APOE genotypes, and life activities. However, the AD-CR cannot be calculated in a living person and thus is not helpful for screening or designing personalized interventions. The mental health resilience (MHR) measure was developed to study depression, anxiety, and wellbeing in individuals living with dementia [84] but has not been studied in cognitively normal individuals. Our goal with this project was to develop a measure of resilience—that is factors that may be protective. This is distinctly different from examining risk factors that increase the likelihood of ADRD and there are number of such scales that examine vulnerability in existence. Instead, we focused on directly modifiable protective factors that could contribute to the brain resilience to the effects of accumulating pathology and give a quantifiable estimate of brain health.

The potential advantages of quantifying brain resilience are multifold. An abundance of research suggests that modifiable risk factors such as physical and cognitive activity are essential to healthy aging [14 –16] and interventions to promote these activities in older adults can have positive effects on health outcomes [18, 20]. Regular physical activity is essential to healthy aging and interventions to promote physical activity in older adults can have positive effects on health outcomes [85]. Recommendations for exercise and physical activity approach 150 min per week [86] but many older adults lead a sedentary life [85 –87], particularly with the development of cognitive impairment [88, 89]. Baseline mindfulness has been associated with better patient and caregiver outcomes [20], and mindfulness-based interventions may improve performance on memory, language, and executive functional tasks [90] and stabilize global scores such as the Mini-Mental State Exam [91]. Multiple ADRD prevention studies are already underway [8 , 23–25] and many more are planned, nearly all of which are multi-modal in nature [92] and contain some aspect of all components of the RI [8 , 23–25]. Designing interventions and quantifying their potential effect on outcomes requires that measurements of brain health and resilience are valid and reliable, that domains captured reflect the multidimensionality of the construct, and that sufficient responsiveness of items is necessary to accurately measure changes of brain health and the risk of cognitive impairment. The RI worked in cognitive normal controls, MCI, and AD but worked equally well in DLB and showed the greatest declines in VCID. This would suggest applications of the RI could extend beyond cognitive work but also be applied in studies of Parkinson’s disease, stroke, and other neurological conditions.

There are several limitations in this study. As this is a cross-sectional study, cause and effect cannot be established. The longitudinal properties of the RI still need to be elucidated to determine its predictive properties and response to interventions. Because the goal of the RI was to detect cognitively normal individuals with high measures of brain health who are less likely to have ADRD, sensitivity was preferred over specificity, and negative likelihood ratios are preferred over positive likelihood ratios [68 , 72]. Sensitivity of the RI was very good (82.1%), but specificity was fair (65.8%). This was likely due to the roughly equal distribution of MCI cases in low and high RI categories. Further examination of RI in different MCI etiologies and their risk of transition to ADRD is needed. However, within group analyses suggests that in controls, MCI, and mild ADRD, individuals with high RI had overall better cognitive, functional, and global performance than those with low RI. The RI was validated in the context of an academic research setting where the prevalence of MCI and dementia are high, and the patients tend to be highly educated and predominantly White. Validation of the RI in other clinical and research settings, other countries, and in individuals from diverse racial, ethnic, and cultural backgrounds is needed. Although the research sample used in this study was largely White, we were able to establish the use of the RI across different socioeconomic, education, and verbal IQ strata. The clinical profile and cut-off scores were developed in this convenience sample and presented as a guide. Future studies should include a more diverse community population to further refine the profiles. The majority of cases consisted of MCI, AD, and DLB with fewer VCID and FTD cases. Biomarker examination was limited to APOE genotypes and MRI. Although NeuroQuant and more commonly used research programs for volumetric analyses (i.e., Freesurfer) are similar [65, 66], the number of regions available from NeuroQuant are limited and only a subset of individuals from the sample had volumetric imaging. Analyses with other specific fluid and imaging biomarkers measuring amyloid-β protein and tau (AD), α-synuclein (DLB and PD), TDP-43 (FTD), NFL (neurodegeneration), and vascular pathology are needed.

Strengths of this study include the use of a comprehensive evaluation that is part of standard of care with extensive characterization of patients and measurement of cognitive, functional, and behavioral constructs using Gold Standard instruments. The diagnostic odds ratio of 8.9 supports the detection of better brain health and suggests a lower risk of having ADRD in individuals with an RI > 143. Another advantage of the RI is its brevity and can be done prior to the research evaluation or clinical office visit without the need for clinician or staff time. The RI could be administered regardless of cognitive status or dementia etiology and can be completed by individuals (with or without assistance from their family) through the moderate-to-severe stages of dementia. The RI may be a useful tool for an estimation of brain health, understanding the potential risk of cognitive impairment, designing primary, secondary, and tertiary dementia prevention research and clinical programs, and exploring vulnerability and resilience in community samples. Low scores on the RI (<143) could be used as a trigger for patients or research participants to be referred for a more extensive evaluation of potentially modifiable factors. The RI performed well across a wide array of sociodemographic variables and correlated with objective measures of cognitive and functional performance in a brief fashion that could facilitate its use in clinical care and research.

Footnotes

ACKNOWLEDGMENTS

This study was supported by grants to JEG from the National Institute on Aging (R01 AG071514, R01 AG069765, and R01 NS101483), the Harry T. Mangurian Foundation, and the Leo and Anne Albert Charitable Trust. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

JEG is the creator of the Resilience Index, Quick Physical Activity Rating, Cognitive & Leisure Activity Scale, Cognitive Reserve Unit Scale, Number Symbol Coding Task, and Quick Dementia Rating System scale used in this study.

Authors’ disclosures available online ().

References

Alzheimer Association (2021) 2021 Alzheimer’s Disease Facts and Figures. https://alz.org/alzheimers-dementia/facts-figures. Accessed on 12 July 2021.

World Health Organization (2020) https://www.who.int/news-room/fact-sheets/detail/dementia. Accessed on 12 July 2021.

Galvin

(2018) Using informant and performance screening methods to detect mild cognitive impairment and dementia. Curr Rep Gerontol 7, 19–25.

Galvin

, Tolea

, Chrisphonte

(2020) What do older adults do with results from dementia screening. PLoS One 15, e0235534.

Galvin

(2017) Prevention of Alzheimer’s disease: Lessons learned and applied. J Am Geriatr Soc 65, 2128–2133.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

, Snyder

, Carrillo

, Thies

, Phelps

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

Mitchell

, Ridley

, Sancho

, Norton

(2017) The future of dementia risk reduction research: Barriers and solutions. J Public Health 39, e275–e281.

O’Donnell

, Browne

, Pierce

, McConnachie

, Deckers

, van Boxtel

MPJ

, Manera

, Kohler

, Redmond

, Verhey

FRJ

, van den Akker

, Power

, Irving

, In-MINDD Team (2015) Reducing dementia risk by targeting modifiable risk factors in mid-life: Study protocol for the Innovative Midlife Intervention for Dementia Deterrence (In-MINDD) randomised controlled feasibility trial. Pilot Feasibility Stud 1, 40.

, Xu

, Tan

, Andrieu

, Suckling

, Evangelou

, Pan

, Zhang

, Jia

, Feng

, Kua

, Wang

, Tan

, Li

, Hou

, Wan

, Tan

, Mok

, Tan

, Dong

, Touchon

, Gauthier

, Aisen

, Vellas

(2020) Evidence-based prevention of Alzheimer’s disease: Systematic review and meta-analysis of 243 observational prospective studies and 153 randomised controlled trials. J Neurol Neurosurg Psychiatry 91, 1201–1209.

10.

Hodes

, Oakley

, O’Keefe

, Lu

, Galvin

, Saif

, Bellara

, Rahman

, Kaufman

, Hristov

, Rajji

, Fosnacht Morgan

, Patel

, Merrill

, Kaiser

, Melendez-Cabrero

, Melendez

, Krikorian

, Isaacson

(2019) Alzheimer’s “prevention” vs. “risk reduction”: Transcending semantic for clinical practice. Front Neurol 9, 1179.

11.

Gaitán

, Boots

, Dougherty

, Ma

, Edwards

, Mitchell

, Christian

, Cook

, Okonkwo

(2020) Protocol of Aerobic Exercise and Cognitive Health (REACH): A pilot study. J Alzheimers Dis Rep 4, 107–121.

12.

Blocker

, Fry

, Luebbers

, Burns

, Perales-Puchalt

, Hansen

, Vidoni

(2020) Promoting Alzheimer’s risk-reduction through community-based lifestyle education and exercise in Rural America: A pilot intervention. Kans J Med 13, 179–185.

13.

Vidoni

, Perales

, Alshehri

, Giles

, Siengsukon

, Burns

(2019) Aerobic exercise sustains performance of Instrumental Activities of Daily Living in early-stage Alzheimer disease. J Geriatr Phys Ther 42, E129–E134.

14.

Yates

, Ziser

, Spector

, Orrell

(2016) Cognitive leisure activities and future risk of cognitive impairment and dementia: Systematic review and metanalysis. Int Psychogeriatr 28, 1791–1806.

15.

Galvin

, Tolea

, Chrisphonte

(2021) The Cognitive & Leisure Activity Scale (CLAS): A new measure to quantify cognitive activities in older adults with and without cognitive impairment. (N Y) 7, e12134.

16.

Cheng

(2016) Cognitive reserve and the prevention of dementia: The role of physical and cognitive activities. Curr Psychiatry Rep 18, 85.

17.

Hackett

, Steptoe

, Cadar

, Fancourt

(2019) Social engagement before and after dementia diagnosis in the English Longitudinal Study of Ageing. PLoS One 14, e0220195.

18.

Biddle

, d’Oleire Uquillas

, Jacobs

HIL

, Zide

, Kirn

, Rentz

, Johnson

, Sperling

, Donovan

(2019) Social engagement and amyloid-β-related cognitive decline in cognitively normal older adults. Am J Geriatr Psychiatry 27, 1247–1256.

19.

Khalsa

(2015) Stress, meditation, and Alzheimer’s disease prevention: Where the evidence stands. J Alzheimers Dis 48, 1–12.

20.

Innis

, Tolea

, Galvin

(2021) The effect of baseline patient and caregiver mindfulness on dementia outcomes. J Alzheimers Dis 79, 1345–1367.

21.

Gustafson

, Bäckman

, Scarmeas

, Stern

, Manly

, Mayeux

, Gu

(2020) Dietary fatty acids and risk of Alzheimer’s disease and related dementias: Observations from the Washington Heights-Hamilton Heights-Inwood Columbia Aging Project (WHICAP). Alzheimers Dement 16, 1638–1649.

22.

Dhana

, Evans

, Rajan

, Bennett

, Morris

(2020) Healthy lifestyle and the risk of Alzheimer dementia: Findings from 2 longitudinal studies. Neurology 95, e374–e383.

23.

Kivipelto

, Mangialasche

, Snyder

, Allegri

, Andrieu

, Arai

, Baker

, Belleville

, Brodaty

, Brucki

, Calandri

, Caramelli

, Chen

, Chertkow

, Chew

, Choi

, Chowdhary

, Crivelli

, Torre

, Du

, Dua

, Espeland

, Feldman

, Hartmanis

, Hartmann

, Heffernan

, Henry

, Hong

, Håkansson

, Iwatsubo

, Jeong

, Jimenez-Maggiora

, Koo

, Launer

, Lehtisalo

, Lopera

, Martínez-Lage

, Martins

, Middleton

, Molinuevo

, Montero-Odasso

, Moon

, Morales-Pérez

, Nitrini

, Nygaard

, Park

, Peltonen

, Qiu

, Quiroz

, Raman

, Rao

, Ravindranath

, Rosenberg

, Sakurai

, Salinas

, Scheltens

, Sevlever

, Soininen

, Sosa

, Suemoto

, Tainta-Cuezva

, Velilla

, Wang

, Whitmer

, Xu

, Bain

, Solomon

, Ngandu

, Carrillo

(2020) World-Wide FINGERS Network: A global approach to risk reduction and prevention of dementia. Alzheimers Dement 16, 1078–1094.

24.

Ritchie

, Molinuevo

, Truyen

, Satlin

, Van der Geyten

, Lovestone

; European Prevention of Alzheimer’s Dementia (EPAD) Consortium (2016) Development of interventions for the secondary prevention of Alzheimer’s dementia: The European Prevention of Alzheimer’s Dementia (EPAD) project. Lancet Psychiatry 3, 179–186.

25.

Rosenberg

, Mangialasche

, Ngandu

, Solomon

, Kivipelto

(2020) Multidomain interventions to prevent cognitive impairment, Alzheimer’s disease, and dementia: From FINGER to World-Wide FINGERS. J Prev Alzheimers Dis 7, 29–36.

26.

Ahn

, Hyun

(2019) Effectiveness of integrative medicine program for dementia prevention on cognitive function and depression of elderly in a public health center. Integr Med Res 8, 133–137.

27.

Galvin

(2018) Advancing personalized treatment of Alzheimer’s disease: A call for the N-of-1 trial design. Future Neurol 13, 151–160.

28.

Stern

, Barulli

(2019) Cognitive reserve. Handb Clin Neurol 167, 181–190.

29.

Chapko

, McCormack

, Black

, Staff

, Murray

(2018) Life-course determinants of cognitive reserve (CR) in cognitive aging and dementia - a systematic literature review. Aging Ment Health 22, 915–926.

30.

Galvin

, Tolea

, Moore

, Chrisphonte

(2020) The Number Symbol Coding Task: A brief measure of executive function to detect dementia and cognitive impairment. PLoS One 15, e0242233.

31.

Vonk

JMJ

, Arce Rentería

, Avila

, Schupf

, Noble

, Mayeux

, Brickman

, Manly

(2019) Secular trends in cognitive trajectories of diverse older adults. Alzheimers Dement 15,, 1576–1587.

32.

Avila

, Rentería

, Jones

, Vonk

JMJ

, Turney

, Sol

, Seblova

, Arias

, Hill-Jarrett

, Levy

, Meyer

, Racine

, Tom

, Melrose

, Deters

, Medina

, Carrion

, Diaz-Santos

, Byrd

, Chesebro

, Colon

, Igwe

, Maas

, Brickman

, Schupf

, Mayeux

, Manly

(2021) Education differentially contributes to cognitive reserve across racial/ethnic groups. Alzheimers Dement 17, 70–80.

33.

Zahodne

, Mayeda

, Hohman

, Fletcher

, Racine

, Gavett

, Manly

, Schupf

, Mayeux

, Brickman

, Mungas

(2019) The role of education in a vascular pathway to episodic memory: Brain maintenance or cognitive reserve? Neurobiol Aging 84, 109–118.

34.

Zahodne

, Manly

, Smith

, Seeman

, Lachman

(2017) Socioeconomic, health, and psychosocial mediators of racial disparities in cognition in early, middle, and late adulthood. Psychol Aging 32, 118–130.

35.

Zahodne

, Stern

, Manly

(2015) Differing effects of education on cognitive decline in diverse elders with low versus high educational attainment. Neuropsychology 29, 649–657.

36.

Ghaffar

, Fiati

, Feinstein

(2012) Occupational attainment as a marker of cognitive reserve in multiple sclerosis. PLoS One 7,, e47206.

37.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43,, 2412–2414.

38.

Galvin

, Valois

, Zweig

(2104) Collaborative transdisciplinary team approach for dementia care. Neurodegener Dis Manag 4, 455–469.

39.

Hollingshead

(1957) Two factor index of social position. Yale University Press, New Haven.

40.

Beekly

, Ramos

, Lee

, Deitrich

, Jacka

, Wu

, Hubbard

, Koepsell

, Morris

, Kukull

, NIA Alzheimer’s Disease Centers (2007) The National Alzheimer’s Coordinating Center (NACC) database: The Uniform Data Set. Alzheimer Dis Assoc Disord 21, 249–258.

41.

Weintraub

, Besser

, Dodge

, Teylan

, Ferris

, Goldstein

, Giordani

, Kramer

, Loewenstein

, Marson

, Mungas

, Salmon

, Welsh-Bohmer

, Zhou

, Shirk

, Atri

, Kukull

, Phelps

, Morris

(2018) Version 3 of the Alzheimer Disease Centers’ Neuropsychological Test Battery in the Uniform Data Set (UDS). Alzheimer Dis Assoc Disord 32, 10–17.

42.

Goetz

, Tilley

, Shaftman

, Stebbine

, Fahn

, Martinez-Martin

, Poewe

, Sampaio

, Stern

, Dodel

, Dubois

, Holloway

, Jankovic

, Kulisevsky

, Lang

, Lees

, Leurgans

, LeWitt

, Nyenhuis

, Olanow

, Rascol

, Schrag

, Teresi

, van Hilten

, Lapelle

; Movement Disorder Society UPDRS Revision Task Force (2008) Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results. Mov Disord 23, 2129–2170.

43.

Charlson

, Pompei

, Ales

, MacKenzie

(1987) A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 40, 373–383.

44.

Groll

, Heland

, Caeser

, Wright

(2006) Assessment of long-term physical function in acute respiratory distress syndrome (ARDS) patients: Comparison of the Charlson Comorbidty index and the functional comorbidity index. Am J Phys Med 7, 574–581.

45.

Wilkins

, Roe

, Morris

(2010) A brief clinical tool to assess physical function: The mini-physical performance test. Arch Gerontol Geriatr 50, 96–100.

46.

Fried

, Tangen

, Walston

, Newman

, Hirsch

, Gottdiener

, Seeman

, Tracy

, Kop

, Burke

, McBurnie

; Cardiovascular Health Study Collaborative Research Group (2001) Frailty in older adults: Evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56, M146–156.

47.

Rosen

, Terry

, Fuld

, Katzman

, Peck

(1980) Pathological verification of ischemic score in differentiation of dementias. Ann Neurol 7, 486–488.

48.

Tolea

, Heo

, Chrisphonte

, Galvin

(2021) A modified CAIDE risk score as a screening tool for cognitive impairment in older adults. J Alzheimers Dis 82, 1755–1768.

49.

Galvin

(2015) The Quick Dementia Rating System (QDRS): A rapid dementia staging tool. Alzheimers Dement (Amst) 1, 249–259.

50.

Tappen

, Rosselli

, Engstrom

(2010) Evaluation of the Functional Activities Questionnaire (FAQ) in cognitive screening across four American ethnic groups. Clin Neuropsychol 24, 646–661.

51.

Kaufer

, Cummings

, Ketchel

, Smith

, MacMillan

, Shelley

, Lopez

, DeKosky

(2000) Validation of the NPI-Q a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci 12, 233–239.

52.

Feeny

, Furlong

, Torrance

, Goldsmith

, Zhu

, DePauw

, Denton

, Boyle

(2002) Multi-attribute and single-attribute utility functions for the Health Utilities Index Mark 3 System. Medical Care 40, 113–128.

53.

Nestle Nutrition Institute, Mini Nutrition Assessment Form. https://www.mna-elderly.com/mna_forms.html. Accessed on 12 July 2021.

54.

Nasreddine

, Phillips

, Bédirian

, Charbonneau

, Whitehead

, Collin

, Cummings

, Chertkow

(2005) The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53, 695–699.

55.

Shapiro

, Benedict

, Schretlen

, Brandt

(1999) Construct and concurrent validity of the Hopkins Verbal Learning Test-revised. . Clin Neuropsychol 13, 348–358.

56.

Reitan

(1958) Validity of the trail making test as an indication of organic brain damage. Perceptual Motor Skills 8, 271–276.

57.

Snaith

(2003) The Hospital Anxiety and Depression Scale. Health Qual Life Outcomes 1, 29.

58.

Galvin

, Tolea

, Chrisphonte

(2020) Using a patient-reported outcome to improve detection of cognitive impairment and dementia: The patient version of the Quick Dementia Rating System (QDRS). PLoS One 15, e0240422.

59.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Many

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging–Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

60.

McKeith

, Boeve

, Dickson

, Halliday

, Taylor

, Weintraub

, Aarsland

, Galvin

, Attems

, Ballard

, Bayston

, Beach

, Blanc

, Bohnen

, Bonanni

, Bras

, Brundin

, Burn

, Chen-Plotkin

, Duda

, El-Agnaf

, Feldman

, Ferman

, Ffytche

, Fujishiro

, Galasko

, Goldman

, Gomperts

, Graff-Radford

, Honig

, Iranzo

, Kantarci

, Kaufer

, Kukull

, Lee

VMY

, Leverenz

, Lewis

, Lippa

, Lunde

, Masellis

, Masliah

, McLean

, Mollenhauer

, Montine

, Moreno

, Mori

, Murray

, O’Brien

, Orimo

, Postuma

, Ramaswamy

, Ross

, Salmon

, Singleton

, Taylor

, Thomas

, Tiraboschi

, Toledo

, Trojanowski

, Tsuang

, Walker

, Yamada

, Kosaka

(2017) Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology 89, 88–100.

61.

Skrobot

, O’Brien

, Black

, Chen

, DeCarli

, Erkinjuntti

, Ford

, Kalaria

, Pantoni

, Pasquier

, Roman

, Wallin

, Sachdev

, Skoog

; VICCCS group, Ben-Shlomo

, Passmore

, Love

, Kehoe

(2017) The Vascular Impairment of Cognition Classification Consensus Study. Alzheimers Dement 13, 624–633.

62.

Olney

, Spina

, Miller

(2017) Frontotemporal dementia. Neurol Clin 35, 339–374.

63.

Galvin

, Tolea

, Rosenfeld

, Chrisphonte

(2020) The Quick Physical Activity Rating (QPAR): A brief assessment of physical activity in older adults with and without cognitive impairment. PLoS One 15, e0241641.

64.

, Black

, Garland

(2016) The Applied Mindfulness Process Scale (AMPS): A process measure for evaluating mindfulness-based interventions. Pers Individ Dif 93, 6–15.

65.

Ross

, Ochs

, Tate

, Tokac

, Seabaugh

, Abildskov

, Bigler

(2018) High correlations between MRI brain volume measurements based on NeuroQuant(®) and FreeSurfer. Psychiatry Res Neuroimaging 278, 69–76.

66.

Persson

, Barca

, Cavallin

, Brækhus

, Knapskog

, Selbæk

, Engedal

(2018) Comparison of automated volumetry of the hippocampus using NeuroQuant® and visual assessment of the medial temporal lobe in Alzheimer’s disease. Acta Radiol 59, 997–1001.

67.

Heister

, Brewer

, Magda

, Blennow

, McEvoy

(2011) Predicting MCI outcome with clinically available MRI and CSF biomarkers. Neurology 77, 1619–1628.

68.

Galvin

, Fagan

, Holtzman

, Mintun

, Morris

(2010) Relationship of dementia screening tests with biomarkers of Alzheimer’s disease. Brain 133, 3290–3300.

69.

McGee

(2002) Simplifying likelihood ratios. . J Gen Intern Med 17, 646–649.

70.

Glas

, Lijmer

, Prins

, Bonsel

, Bossuyt

(2003) The diagnostic odds ratio: A single indicator of test performance. J Clin Epidemiology 56, 1129–1135.

71.

Harrell

, Califf

, Pryor

, Lee

, Rosati

(1982) Evaluating the yield of medical tests. JAMA 247, 2543–2546.

72.

Boston University School of Public Health, Screening for Disease. https://sphweb.bumc.bu.edu/otlt/MPH-Modules/EP/EP713_Screening/EP713_Screening_print.html. Accessed on 12 July 2021.

73.

Grande

, Qiu

, Fratiglioni

(2020) Prevention of dementia in an ageing world: Evidence and biological rationale. Ageing Res Rev 64, 01045.

74.

Arenaza-Urquijo

, Vemuri

(2020) Improving the resistance and resilience framework for aging and dementia studies. Alzheimers Res Ther 12, 41.

75.

Duggan

, Parikh

(2021) Microglia and modifiable life factors: Potential contributions to cognitive resilience in aging. Behav Brain Res 405, 113207.

76.

Fratiglioni

, Marseglia

, Dekhtyar

(2020) Ageing without dementia: Can stimulating psychosocial and lifestyle experiences make a difference? Lancet Neurol 19, 533–543.

77.

Bartrés-Faz

, Arenaza-Urquijo

, Ewers

, Belleville

, Chételat

, Franzmeier

, Gonneaud

, de Echevarri

JMG

, Okonkwo

, Schultz

, Valenzuela

, Stern

, Vemuri

(2020) Theoretical frameworks and approaches used within the Reserve, Resilience and Protective Factors professional interest area of the Alzheimer’s Association International Society to Advance Alzheimer’s Research and Treatment. Alzheimers Dement 12, e12115.

78.

Tesi

, van der Lee

, Hulsman

, Jansen

, Stringa

, van Schoor

, Scheltens

, van der Flier

, Huisman

, Reinders

MJT

, Holstege

(2020) Immune response and endocytosis pathways are associated with the resilience against Alzheimer’s disease. Transl Psychiatry 10, 332.

79.

Rosa

, Giannotti

, Martella

, Massa

, Serafini

, Pardini

, Nobili

, Monacelli

, Disease Management Team on Dementia of the IRCCS Ospedale Policlinico San Martino (Genoa, I) (2020) Brain aging, cardiovascular diseases, mixed dementia, and frailty in the oldest old: From brain phenotype to clinical expression. J Alzheimers Dis 75, 1083–1103.

80.

Dumitrescu

, Mahoney

, Mukherjee

, Lee

, Bush

, Engelman

, Lu

, Fardo

, Trittschuh

, Mez

, Kaczorowski

, Saucedo

, Widaman

, Buckley

, Properzi

, Mormino

, Yang

, Harrison

, Hedden

, Nho

, Andrews

, Tommet

, Hadad

, Sanders

, Ruderfer

, Gifford

, Moore

, Cambronero

, Zhong

, Raghavan

, Vardarajan

, Alzheimer Disease Neuroimaging Initiative (ADNI); Alzheimer’s Disease Genetics Consortium (ADGC), A4 Study Team; Pericak-Vance

, Farrer

, Wang

, Cruchaga

, Schellenberg

, Cox

, Haines

, Keene

, Saykin

, Larson

, Sperling

, Mayeux

, Bennett

, Schneider

, Crane

, Jefferson

, Hohman

(2020) Genetic variants and functional pathways associated with resilience to Alzheimer’s disease. Brain 143, 2561–2575.

81.

, Tasaki

, Schneider

, Arfanakis

, Duong

, Wingo

, Kearns

, Thatcher

GRJ

, Seyfried

, Levey

, De Jager

, Bennett

(2020) Cortical proteins associated with cognitive resilience in community-dwelling older persons. JAMA Psychiatry 77, 1–9.

82.

Busatto

, de Gobbi Porto

, Faria

, Squarzoni

, Coutinho

, Garcez

, Rosa

PGP

, da Costa

, Carvalho

, Torralbo

, de Almeida Hernandes

, Ono

, Brucki

SMD

, Nitrini

, Buchpiguel

, Souza Duran

, Forlenza

(2020) In vivo imaging evidence of poor cognitive resilience to Alzheimer’s disease pathology in subjects with very low cognitive reserve from a low-middle income environment. Alzheimers Dement 12, e12122.

83.

Yao

, Sweeney

, Nagorski

, Shulman

, Allen

(2020) Quantifying cognitive resilience in Alzheimer’s disease: The Alzheimer’s Disease Cognitive Resilience Score. PLoS One 15, e0241707.

84.

Windle

, Hoare

, Woods

, Huisman

, Burholt

(2021) A longitudinal exploration of mental health resilience, cognitive impairment, and loneliness. Int J Geriatr Psychiatry 36, 1020–1028.

85.

CDC Active People Healthy Nation, https://www.cdc.gov/physicalactivity/downloads/Active_People_Healthy_Nation_at-a-glance_082018_508.pdf. Accessed on 12 July 2021.

86.

Physical Activity Guidelines Advisory Committee (2008) Physical activity guidelines advisory committee report. US Department of Health and Human Services, 2008, A1-H14, Washington, DC.

87.

Falck

, McDonald

, Beets

, Brazendale

, Liu-Ambrose

(2016) Measurement of physical activity in older adult interventions: A systematic review. Br J Sports Med 50, 464–470.

88.

Tolea

, Morris

, Galvin

(2015) Longitudinal associations between physical and cognitive function among community-dwelling older adults. PLoS One 10, e0122878.

89.

Tolea

, Morris

, Galvin

(2016) Trajectory of physical functional decline by type of dementia. Alzheimer Dis Assoc Disord 30, 60–66.

90.

Newberg

, Wintering

, Khalsa

, Roggenkamp

, Waldman

(2010) Meditation effects on cognitive function and cerebral blood flow in subjects with memory loss: A preliminary study. J Alzheimers Dis 20, 517–526.

91.

Quintana-Hernandez

, Miro-Barrachina

, Ibanez-Fernandez

, Santana-DelPino

, Quintana-Montesdeoca

, Rodriguez-de Vera

, Morales-Casanova

, Perez-Vieitez

MDC

, Rodriguez-Garcia

, Bravo-Caraduje

(2016) Mindfulness in the maintenance of cognitive capacities in Alzheimer’s disease: A randomized clinical trial. J Alzheimers Dis 50, 217–232.

92.

Montero-Odasso

, Ismail

, Livingston

(2020) One third of dementia cases can be prevented within the next 25 years by tackling risk factors. The case “for” and “against”. Alzheimers Res Ther 12, 81.