Abstract

Higher occupational attainment has previously been associated with increased Alzheimer's disease (AD) neuropathology when individuals are matched for cognitive function, indicating occupation could provide cognitive reserve. We examined whether occupational complexity (OCC) associates with decreased hippocampal volume and increased whole-brain atrophy given comparable cognitive function in middle-aged adults at risk for AD. Participants (n = 323) underwent structural MRI, cognitive evaluation, and work history assessment. Three complexity ratings (work with data, people, and things) were obtained, averaged across up to 3 reported jobs, weighted by years per job, and summed to create a composite OCC rating. Greater OCC was associated with decreased hippocampal volume and increased whole-brain atrophy when matched for cognitive function; results remained substantively unchanged after adjusting for several demographic, AD risk, vascular, mental health, and socioeconomic characteristics. These findings suggest that, in people at risk for AD, OCC may confer resilience to the adverse effects of neuropathology on cognition.

Introduction

The cognitive reserve (CR) hypothesis posits that individuals with higher CR are able to better maintain cognitive function despite Alzheimer's disease (AD) pathology, likely due to compensatory approaches or preexisting cognitive strategies (Stern, 2012). These compensatory approaches and strategies are thought to be a direct result of various life exposures—such as high occupational or educational attainment—which lead to lessened adverse effects of AD pathology on cognition (Garibotto et al., 2008; Lo & Jagust, 2013; Stern et al., 1995).

One approach to testing the CR hypothesis is to investigate the association between a CR proxy (e.g., occupation) and brain pathology while holding cognitive function constant (Stern, Alexander, Prohovnik, & Mayeux, 1992). Holding cognitive function constant effectively matches individuals for variations in cognitive symptoms and allows for statistical comparison of individuals at the “same” level of cognitive functioning. In this paradigm, CR would be in effect if better scores on the CR proxy are associated with increased brain pathology, indicating that those with higher reserve are able to compensate for identified brain pathology and maintain a comparable level of cognitive function to individuals with less pathology (Stern, 2012; Stern et al., 1992). This method for testing CR has been used in prior studies with occupation as the CR proxy (Garibotto et al., 2008, 2012; Stern et al., 1995). In a seminal report, Stern and colleagues (1995) demonstrated an association between higher occupation and decreased parietal blood flow in AD patients while controlling for clinical dementia severity. More recently, Garibotto and colleagues (2008, 2012) showed similar results in both mild cognitive impairment (MCI) and AD patients, where higher occupational attainment was associated with cerebral hypometabolism while adjusting for cognitive function. Findings of a similar inverse nature have also been reported using other proxies for CR, including education (Kemppainen et al., 2008; Perneczky et al., 2006; Stern et al., 1992), socioeconomic status (SES; Fotenos, Mintun, Snyder, Morris, & Buckner, 2008), and cognitively stimulating activities (Scarmeas et al., 2003).

Occupational complexity (OCC) is a unique way to assess “occupation” because—in contrast to mere job title—it captures particular skills that, in turn, may reflect specific cognitive processes used in diverse occupations. Our group (Jonaitis et al., 2013) and others (Andel, Kareholt, Parker, Thorslund, & Gatz, 2007; Smart, Gow, & Deary, 2014) have shown that OCC is associated with better cognitive performance. However, no studies to date have examined the association between OCC and brain health given matched cognitive status in the fashion described earlier. Additionally, this concept has not been tested within a cohort of middle-aged adults who harbor specific risk factors for AD, including parental history of AD and apolipoprotein E4 (APOE4) genotype. This is a critical knowledge gap, seeing as a cohort at risk for AD is a choice population for implementing preventative strategies to protect against AD.

Therefore, the objective of this study was to determine whether higher OCC is associated with (a) comparatively lower hippocampal volume given the same level of memory function and (b) comparatively greater whole-brain atrophy given the same level of global cognitive function in a middle-aged cohort at risk for AD.

Methods

Participants

The Wisconsin Registry for Alzheimer's Prevention (WRAP) is an ongoing longitudinal study consisting of ∼1,500 cognitively healthy, middle-aged adults aged 40–65 at study entry (Sager, Hermann, & La Rue, 2005). For the present study, 332 participants were initially selected based on having available T1 MRI data. Two were subsequently excluded due to missing occupational data and seven were excluded due to missing covariate information. Thus, 323 participants were included in this study. As with the overall WRAP cohort, this study's sample was enriched with persons who had parental history of AD (74.0%) and who were APOE4 positive (41.2%). Additional relevant background characteristics are listed in Table 1. The University of Wisconsin Institutional Review Board approved all study procedures and informed consent was obtained from all participants included in the study.

Table 1.

Participant characteristics

Characteristics Valuea 
Demographics 
 Age (years) 60.38 (6.09) 
 Women (%) 68.1 
 Education (years) 16.05 (2.33) 
 Family history positive (%) 74.0 
 APOE4 status 
  Noncarrier (%) 58.8 
  Heterozygote %) 38.7 
  Homozygote (%) 2.5 
 Household income $50,000 or greater (%)b 79.9 
Occupational complexity 
 Total OCC 9.80 (3.15) 
 Complexity with data 4.17 (1.16) 
 Complexity with people 3.51 (2.18) 
 Complexity with things 2.11 (2.30) 
Cognitive and mood measures 
 MMSE 29.30 (1.05) 
 RAVLT composite score 75.42 (12.14) 
 Mean cognitive factor score 0.12 (0.64) 
 CES-D 6.00 (6.12) 
 Stress 3.14 (2.83) 
Vascular risk indices 
 Hypertension (%) 22.0 
 Diabetes (%) 3.7 
 Smoker (current) (%) 2.8 
Characteristics Valuea 
Demographics 
 Age (years) 60.38 (6.09) 
 Women (%) 68.1 
 Education (years) 16.05 (2.33) 
 Family history positive (%) 74.0 
 APOE4 status 
  Noncarrier (%) 58.8 
  Heterozygote %) 38.7 
  Homozygote (%) 2.5 
 Household income $50,000 or greater (%)b 79.9 
Occupational complexity 
 Total OCC 9.80 (3.15) 
 Complexity with data 4.17 (1.16) 
 Complexity with people 3.51 (2.18) 
 Complexity with things 2.11 (2.30) 
Cognitive and mood measures 
 MMSE 29.30 (1.05) 
 RAVLT composite score 75.42 (12.14) 
 Mean cognitive factor score 0.12 (0.64) 
 CES-D 6.00 (6.12) 
 Stress 3.14 (2.83) 
Vascular risk indices 
 Hypertension (%) 22.0 
 Diabetes (%) 3.7 
 Smoker (current) (%) 2.8 

Notes: OCC = occupational complexity; APOE4 = apolipoprotein E4 allele; RAVLT = Rey Auditory Verbal Learning Test; MMSE = Mini-Mental State Exam; CES-D = Center for Epidemiologic Studies—Depression scale.

aValues are mean (SD) unless otherwise indicated.

bMedian Wisconsin household income from 2009 to 2013 = $52,413; http://quickfacts.census.gov/qfd/states/55000.html, retrieved June 2, 2015.

Occupational Complexity

Occupational data were gathered from participants via questionnaires and/or in-person interviews. Participants were asked to report up to three main occupations held in their adult life and the number of years spent on each job. Each reported occupation was then coded using O*NET according to the 1970 U.S. Census Dictionary of Occupational Titles (United States Employment Service, 1991). Occupations were coded into three complexity ratings: complexity of work with data, complexity of work with people, and complexity of work with things (Roos & Treiman, 1980). Scores ranged from 0 (most complex) to 6 (least complex) for complexity of work with data, 0 to 8 for complexity of work with people, and 0 to 7 for complexity of work with things. Additional details about the coding process have been previously described (Jonaitis et al., 2013).

OCC scores for data, people, and things were reverse coded so that higher scores corresponded with greater job complexity. Next, following standard procedure (Jonaitis et al., 2013), we created separate weighted complexity ratings for data, people, and things for each participant by averaging ratings across their reported jobs, weighted by years on each job. Finally, a total OCC variable was created by summing the three weighted complexity ratings (data, people, and things; possible range, 0–21). Mean weighted reverse-coded ratings for total OCC and complexity of work with data, people, and things are shown in Table 1.

Neuroimaging Protocol

All participants underwent MRI scanning on a GE ×750 3.0-T scanner with an eight-channel phased array head coil (General Electric, Waukesha, WI). Scanning was completed after a 4-hr fast from food, caffeine, tobacco, and medications with vasomodulatory properties. Three-dimensional T1-weighted inversion recovery prepared SPGR anatomical sequences were collected using the following parameters: inversion time/echo time/relaxation time = 450 ms/3.2 ms/8.2 ms, flip angle = 12°, slice thickness = 1 mm no gap, field of view = 256, matrix size = 256 × 256, yielding a voxel resolution of 1 × 1 × 1 mm.

Hippocampal volume, ventricular volumes, and total gray matter (GM) volume were extracted from the T1 images using FreeSurfer image analysis suite version 5.1.0 (http://surfer.nmr.mgh.harvard.edu/, retrieved June 2, 2015), as has been previously described (Dale, Fischl, & Sereno, 1999; Fischl, Sereno, & Dale, 1999). Right and left hippocampal volumes were averaged in order to obtain a single hippocampal volumetric measure. Total ventricular volume was calculated by summing the lateral, inferior lateral, third, fourth, and fifth ventricles. Our measure of whole-brain atrophy, ventricle-to-brain ratio (VBR; Bigler et al., 2004), was computed as [(total ventricular volume/total GM volume) × 100], with higher values signifying greater brain atrophy. VBR can be considered as a measure of brain atrophy because “in the ex vacuo state enlarged ventricular space occurs only in proportion to dissolute brain parenchyma” (Bigler et al., 2004). Although VBR is a cross-sectional proxy for whole-brain atrophy, which is typically measured longitudinally, it has previously been shown to be a viable marker of AD progression (Bigler et al., 2004; Okonkwo et al., 2010), exhibits robust associations with neuropsychological performance (Bigler et al., 2004), discriminates between those with dementia and their cognitively healthy peers (Bigler & Tate, 2001), and is reliably associated with other AD biomarkers (Ott et al., 2010).

Neuropsychological Assessment

All WRAP participants complete a comprehensive neuropsychological test battery (Sager et al., 2005) consisting of global measures, such as the Mini-Mental State Exam (MMSE), as well as psychometric measures for specific cognitive domains, such as memory and executive function. For this study, test scores on the Rey Auditory Verbal Learning Test (RAVLT; Schmidt, 1996) Total Learning, Long Delay, and Recognition indices were summed to create a composite memory test score for each participant. As a measure of global cognition, the mean of four cognitive factors, which have been previously described (Dowling, Hermann, La Rue, & Sager, 2010; Koscik et al., 2014), was used. Briefly, these four cognitive factors include—Immediate Memory: RAVLT learning trials 1 and 2 (Schmidt, 1996); Verbal Learning & Memory: RAVLT learning trials 3–5 and Delayed Recall (Schmidt, 1996); Working Memory: Digit Span and Letter-Number Sequencing subtests from the Wechsler Adult Intelligence Scale, third edition (Wechsler, 1997); Speed & Flexibility: Stroop Color–Word Test Interference Trial (Trenerry, Crosson, DeBoe, & Leber, 1989) and Trail-Making Test A and B (Reitan & Wolfson, 1993). The average time interval between MRI scan and cognitive testing was 0.55 years (SD = 0.49 years).

Cardiovascular Risk Factors

Participants completed a health history questionnaire that inquired into current and past medical history of various cardiovascular diseases and risk factors, including hypertension, diabetes, and smoking (see Table 1).

Mental Health Measures

Stress was assessed using a 12-item questionnaire adapted from the Women's Health Initiative (Michael et al., 2009). This questionnaire asks whether or not the individual experienced specific stressful events in the past year (e.g., Did your spouse or partner die? Did you or a family member or close friend lose their job or retire?). If no, the participant was to rate the event as (0), and if yes, to rate the event as either mildly stressful (1), stressful (2), or very stressful (3). Responses from all 12 items were summed to create a total stress score (possible range, 0–36). Depression was measured using the Center for Epidemiological Studies—Depression (CES-D) scale (Radloff, 1977).

Socioeconomic Status

Participant SES was indexed by their total household income, which included salaries, wages, social security, retirement income, investments, and all other sources of income. Participants indicated their total household income by choosing from the following categories: <$10,000; $10,000–$19,999; $20,000–$34,999; $35,000–$49,999; $50,000–$74,999; $75,000–$99,999; $100,000–$149,999; and $150,000 or more.

Participants also reported the job that each of their parents performed for the majority of the parent's life. Because early-life SES may influence later cognitive functioning (Zahodne, Stern, & Manly, 2015), we captured SES in the participant's family of origin by computing parental total OCC based on the job held by the parent who had the more complex occupation. This parental total OCC measure was created in the same fashion as participant total OCC (i.e., summing complexity of work with data, people, and things); however, it was not weighted by years on the job because only one job was recorded per parent.

Statistical Analyses

Multiple linear regression was used to examine the relationship between total weighted OCC and either hippocampal volume or VBR. We fitted five models that incrementally adjusted for several potentially confounding measures in order to test the robustness of initial findings. Specifically, in assessing the relationship between total OCC and hippocampal volume, the following models were tested: Model 1—adjusted for age, sex, years of education, MRI scan-cognitive testing time interval, intracranial volume, and composite RAVLT score (i.e., Total Learning + Delayed Recall + Recognition); Model 2—adjusted for Model 1 covariates in addition to AD risk factors including family history and APOE4 status (i.e., 0 = noncarrier, 1 = heterozygote, 2 = homozygote); Model 3—adjusted for Model 2 covariates in addition to vascular risk factors including history of hypertension, diabetes, and smoking; Model 4—adjusted for Model 3 covariates in addition to mental health measures including CES-D and total stress; Model 5—adjusted for Model 4 covariates in addition to participant and parental SES. In assessing the relationship between total OCC and VBR, the same models were used with the exception that mean cognitive factor score, our measure of global cognition, was used in place of the RAVLT composite memory score and that intracranial volume was not included as a covariate in the models. Education was entered as a covariate in our models to account for its inherent relationship with both cognitive performance and CR. The design implemented allowed us to evaluate the relationship between OCC and brain measures while holding cognition constant and excluding the effects of other potentially confounding covariates.

Secondary analyses were conducted using weighted complexity of work with data, or people, or things—instead of total weighted OCC—in relation to either hippocampal volume or VBR in order to understand whether any of these work components differentially drove the associations found with total weighted OCC. These ancillary analyses were conducted using only our fully adjusted model (i.e., Model 5, see earlier) in order to comprehensively account for potentially confounding factors. Finally, although we have previously reported on the relationship between OCC and cognition in the full WRAP cohort (Jonaitis et al., 2013), for completeness sake, we also tested the association between total OCC and composite RAVLT score, mean cognitive factor score, and MMSE while adjusting for age, sex, education, family history of AD, APOE4 status, hypertension, diabetes, smoking, CES-D, total stress, and participant and parental SES. Only findings with p ≤ .05 (two tailed) were considered significant. All analyses were conducted using IBM SPSS, version 21.0.

Results

Background Characteristics

Participant background characteristics are detailed in Table 1. The average age was 60.38 ± 6.09 years at MRI scan and 68.1% were women. This cohort is well educated (mean years of education = 16.05 ± 2.33) with 79.9% of participants reporting a mean household income of $50,000 or greater. Mean total OCC was 9.80 ± 3.15. The mean MMSE score was 29.30 ± 1.05 (range = 23–30) and the mean CES-D score was 6.00 ± 6.12 (range = 0–34), indicating a cognitively healthy cohort.

OCC and Hippocampal Volume

The results of the linear regression showed that higher OCC was significantly associated with lower hippocampal volume, holding RAVLT composite test score constant, and additionally controlling for age, sex, education, intracranial volume, and MRI scan-cognitive testing time interval (Model 1, see Table 2). This result remained significant when additively adjusting for confounding factors, which successively included AD risk factors (Model 2), vascular risk factors (Model 3), mental health measures (Model 4), and participant/parental SES (Model 5; see Table 2 and Fig. 1). Vascular risk factors and SES seemed to have the greatest attenuating effects on the relationship between OCC and hippocampal volume, although the initial finding remained significant for all subsequent models.

Table 2.

Association between OCC and hippocampal volume

 β (SE) t p 
Model 1 −14.69 (6.96) −2.11 .036 
Model 2 −14.82 (6.98) −2.12 .034 
Model 3 −14.16 (6.95) −2.04 .042 
Model 4 −14.61 (6.97) −2.10 .037 
Model 5 −14.21 (7.14) −1.99 .048 
 β (SE) t p 
Model 1 −14.69 (6.96) −2.11 .036 
Model 2 −14.82 (6.98) −2.12 .034 
Model 3 −14.16 (6.95) −2.04 .042 
Model 4 −14.61 (6.97) −2.10 .037 
Model 5 −14.21 (7.14) −1.99 .048 

Notes: The regression statistics reported here are for the association between total OCC and hippocampal volume while adjusting for the following covariates: Model 1 = age, sex, years of education, MRI scan-cognitive testing time interval, intracranial volume, and composite RAVLT score; Model 2 = Model 1 plus FH positivity and APOE4 status; Model 3 = Model 2 plus hypertension, diabetes, and smoking; Model 4 = Model 3 plus stress and CES-D; Model 5 = Model 4 plus participant/parental SES. OCC = occupational complexity; RAVLT = Rey Auditory Verbal Learning Test; FH = family history of Alzheimer's disease; APOE4 = apolipoprotein E4 allele; CES-D = Center for Epidemiological Studies—Depression scale; SES = socioeconomic status.

Fig. 1.

Inverse relationship between OCC and hippocampal volume. Higher OCC is associated with lower hippocampal volumes given same memory test scores and adjusted for age, sex, education, ICV, MRI scan-cognitive testing time interval, AD risk factors, vascular risk factors, mental health factors, and SES. For graphing purposes, OCC was dichotomized at the median into the depicted low versus high groups. OCC = occupational complexity; ICV = intracranial volume; MRI = magnetic resonance imaging; AD = Alzheimer's disease; SES = socioeconomic status.

Fig. 1.

Inverse relationship between OCC and hippocampal volume. Higher OCC is associated with lower hippocampal volumes given same memory test scores and adjusted for age, sex, education, ICV, MRI scan-cognitive testing time interval, AD risk factors, vascular risk factors, mental health factors, and SES. For graphing purposes, OCC was dichotomized at the median into the depicted low versus high groups. OCC = occupational complexity; ICV = intracranial volume; MRI = magnetic resonance imaging; AD = Alzheimer's disease; SES = socioeconomic status.

In order to understand if specific occupational complexities (complexity of work with data, people, or things) were contributing differentially to the association between total OCC and hippocampal volume, we fitted fully adjusted models (Model 5) for each of the three specific occupational complexities. There were no associations between either complexity of work with data or complexity of work with things and hippocampal volume (p ≥ .545); however, complexity of work with people was negatively associated with hippocampal volume while adjusting for RAVLT memory test score, demographics, MRI scan-cognitive testing time interval, AD risk factors, vascular risk factors, mental health measures, and SES (β (SE) = −23.95 (10.77); t = −2.22; p = .027).

OCC and VBR

Regression analyses revealed that higher total OCC was significantly associated with greater brain atrophy while adjusting for mean cognitive factor score in addition to age, sex, education, and MRI scan-cognitive test time interval (Model 1, see Table 3). Incremental adjustment for AD risk factors, vascular risk factors, mental health measures, and parental/participant SES in subsequent models did not change this association (see Table 3 and Fig. 2). Parental/participant SES seemed to have the greatest attenuating effect on the relationship between total OCC and VBR, but again, the original association remained significant even while accounting for these additional potentially confounding factors. When these analyses were repeated with MMSE as our measure of global cognition, the results remained substantively unchanged (data not shown).

Table 3.

Association between OCC and brain atrophy

 β (SE) t p 
Model 1 0.094 (0.039) 2.43 .016 
Model 2 0.095 (0.039) 2.44 .015 
Model 3 0.096 (0.039) 2.47 .014 
Model 4 0.097 (0.039) 2.48 .014 
Model 5 0.088 (0.040) 2.21 .028 
 β (SE) t p 
Model 1 0.094 (0.039) 2.43 .016 
Model 2 0.095 (0.039) 2.44 .015 
Model 3 0.096 (0.039) 2.47 .014 
Model 4 0.097 (0.039) 2.48 .014 
Model 5 0.088 (0.040) 2.21 .028 

Notes: The regression statistics reported here are for the association between total OCC and VBR while adjusting for the following covariates: Model 1 = age, sex, years of education, MRI scan-cognitive testing time interval, and mean cognitive factor score; Model 2 = Model 1 plus FH positivity and APOE4 status; Model 3 = Model 2 plus hypertension, diabetes, and smoking; Model 4 = Model 3 plus stress and CES-D; Model 5 = Model 4 plus participant/parental SES. OCC = occupational complexity; VBR = ventricle-to-brain ratio; FH = family history of Alzheimer's disease; APOE4 = apolipoprotein E4 allele; CES-D = Center for Epidemiological Studies—Depression scale; SES = socioeconomic status.

Fig. 2.

Higher OCC is associated with increased brain atrophy. Higher OCC is associated with greater brain atrophy given same global cognitive performance and adjusted for age, sex, education, MRI scan-cognitive testing time interval, AD risk factors, vascular risk factors, mental health factors, and SES. For graphing purposes, OCC was dichotomized at the median into the depicted Low versus High groups. OCC = occupational complexity; MRI = magnetic resonance imaging; AD = Alzheimer's disease; SES = socioeconomic status.

Fig. 2.

Higher OCC is associated with increased brain atrophy. Higher OCC is associated with greater brain atrophy given same global cognitive performance and adjusted for age, sex, education, MRI scan-cognitive testing time interval, AD risk factors, vascular risk factors, mental health factors, and SES. For graphing purposes, OCC was dichotomized at the median into the depicted Low versus High groups. OCC = occupational complexity; MRI = magnetic resonance imaging; AD = Alzheimer's disease; SES = socioeconomic status.

Additionally, we tested whether complexity of work with data, people, or things would exhibit differential associations with VBR. Similar to our hippocampal volume results, there were no significant associations between either complexity of work with data or complexity of work with things and VBR (p ≥ .143). However, complexity of work with people was significantly associated with VBR, with higher complexity corresponding to greater VBR values (β (SE) = 0.23 (0.06); t = 3.86; p < .001).

OCC and Cognition

Linear regressions indicated that total OCC was positively associated with RAVLT composite score (β (SE) = 0.41 (0.20); t = 2.02; p = .044), adjusting for age, sex, education, family history of AD, APOE4 status, hypertension, diabetes, smoking, CES-D, total stress, and participant and parental SES. Total OCC was also associated with mean cognitive factor score (β (SE) = 0.03 (0.01); t = 2.34; p = .020) and trended toward significance with MMSE score (β (SE) = 0.04 (0.02); t = 1.81; p = .071). These results correspond with previous findings from the larger WRAP cohort demonstrating relationships between OCC and cognitive function (Jonaitis et al., 2013).

Discussion

In this study, we found that even after adjusting for potentially confounding variables that captured SES, AD risk, vascular health, and mental health, OCC (a CR proxy) was associated with decreased hippocampal volume and increased brain atrophy when participants were matched for cognitive function. These results suggest that in middle-aged persons at risk for AD, those with higher OCC are able to better tolerate AD-like pathology and maintain a similar level of cognitive performance compared with those with less pathology. Additionally, we found that of the three OCC components (i.e., data, people, and things), only complexity of work with people was significantly associated with decreased hippocampal volume and increased brain atrophy while controlling for cognitive function and other covariates, indicating that social components of occupation may have the greatest relevance to CR.

In a pivotal study, Stern and colleagues (1995) showed that in 51 AD patients (mean age = 74 years) who were matched for clinical severity, those whose primary lifetime occupation involved higher-level interpersonal skills exhibited decreased parietal blood flow. After additionally controlling for education, more advanced interpersonal skills in occupation were still associated with decreased parietal blood flow, suggesting that occupation was independently contributing to CR (Stern et al., 1995). More recently, Garibotto and colleagues (2008) demonstrated that in both AD (n = 242; mean age = 71 years) and amnestic MCI patients (n = 72; mean age = 68), greater occupational attainment as measured by job title was associated with decreased glucose metabolism in regions associated with AD, including temporal and parietal cortices and posterior cingulate cortex. To our knowledge, our study is the first to show comparable results in a cohort of individuals who are at risk for AD, but who do not currently show any clinical manifestations of the disease. The present study also aligns with previous research using different CR proxies, such as the study by Kemppainen and colleagues (2008), which found increased amyloid deposition and reduced glucose metabolism in highly educated participants compared with low-educated participants when matched for cognitive function. In addition to providing support for previous research, our findings suggest that OCC is providing protection in preclinical AD, indicating the potential benefit of partaking in CR-strengthening activities earlier in life. Importantly, our results also pinpoint that OCC can maintain memory function in the face of hippocampal atrophy, which is salient given that the hippocampus is critical for memory formation and is a primary brain structure affected in the AD cascade (Braak & Braak, 1997).

Although previous reports have investigated CR in a similar fashion to our analysis (Fotenos et al., 2008; Kemppainen et al., 2008; Perneczky et al., 2006), perhaps the more “traditional” method for investigating CR involves examining a CR proxy, such as education or occupation, as a modifier of the link between brain pathology and cognitive function. For example, Rentz and colleagues (2010) showed that CR, as measured by education, modified the relationship between amyloid deposition and neuropsychological performance such that those with greater CR did not exhibit poorer neuropsychological performance with increased amyloid deposition. A similarly designed study by Roe and colleagues (2008) found comparable results, wherein higher education ameliorated the deleterious effects of amyloid deposition on cognition. Our results, although presented in a different statistical fashion, also support the CR hypothesis because both approaches demonstrate CR buffering the otherwise adverse effect of AD pathology on cognitive function.

Interestingly, our study showed that when OCC was broken down to its individual components, complexity of work with people seemed to have the greatest impact on brain health. Prior research has indicated that social interaction is thought to help curb the progression of AD (Fratiglioni, Wang, Ericsson, Maytan, & Winblad, 2000; Zuelsdorff et al., 2013). Additionally, two studies that examined the effects of occupation and leisure activities on cognition found that both occupation and participating in social activities exerted similar protective effects on cognitive function (Adam, Bonsang, Grotz, & Perelman, 2013; Andel, Silverstein, & Kareholt, 2015). Furthermore, in Andel and colleagues (2015), the influence of complexity of work with people on later life cognition was attenuated when participation in social activity was taken into account. These findings are clinically significant, especially for individuals with jobs involving little social interaction, because they suggest that participating in social activities, such as clubs or organizations, could have as beneficial an impact on cognitive health as engaging in complex interactions with others in an occupational setting.

Some limitations of this study include its cross-sectional nature, which does not allow for understanding the effect OCC may have on neuropathological accumulation over time. Future studies are needed to test this relationship using longitudinal measures of hippocampal and whole-brain atrophy, as opposed to cross-sectional proxies such as VBR. Additionally, many of our measures, including OCC, SES, and vascular risk factors, rely on self-report, which could affect the accuracy of our results. Finally, OCC ratings were determined from the 1970 U.S. Census Dictionary of Occupational Titles, which may not be congruent with current occupational classifications; for example, homemakers were excluded from the Dictionary.

A major strength of this study was our ability to adjust for a variety of factors that had not been controlled for as rigorously in previous studies, including education, vascular risk factors, mental health factors, and SES. These factors are all highly correlated with either occupation or AD (Garibotto et al., 2008; Ngandu et al., 2007; Wilson et al., 2002), and it was necessary to adjust for them in order to discern the unique association between OCC and neuropathology. Even so, it was interesting that SES had the greatest attenuating effect on the relationship between OCC and neuropathology, possibly reflecting the intrinsic similarities between SES and OCC. More research is needed to elucidate which CR proxies have the greatest protective function against cognitive decline. Another strength of this study was the use of a robust composite OCC measure and, more importantly, the ability to explicate which components of OCC (specifically, complexity of work with people) made the greatest individual contribution to CR.

In conclusion, this study found that greater OCC is associated with decreased hippocampal volume and increased brain atrophy in a middle-aged cohort at risk for AD while holding cognition constant. This indicates that OCC is protective against cognitive deterioration in the face of AD pathology, providing support for the CR hypothesis and demonstrating the potential for a modifiable lifestyle factor to slow the clinical phenotype of AD.

Funding

This work was supported by the National Institute on Aging (K23 AG045957 to OCO, R01 AG021155 to SCJ, R01 AG027161 to SCJ, P50 AG033514 to SA, and P50 AG033514-S1 to OCO), Veterans Administration Merit Review Grant (I01CX000165 to SCJ), and Clinical and Translational Science Award (UL1RR025011) to the University of Wisconsin, Madison. Portions of this research were supported by the Wisconsin Alumni Research Foundation, the Helen Bader Foundation, Northwestern Mutual Foundation, Extendicare Foundation, and from the Veterans Administration including facilities and resources at the Geriatric Research Education and Clinical Center of the William S. Middleton Memorial Veterans Hospital, Madison, WI.

Conflict of Interest

None declared.

Acknowledgements

We thank Caitlin A. Cleary, Sandra Harding, Jennifer Bond, Nancy Davenport-Sis, Janet Rowley, Amy Hawley, and the WRAP psychometrists for helping with study data collection; researchers and staff at the Waisman Center, University of Wisconsin-Madison, where the brain scans took place; and finally, we thank study participants in the WRAP for their continued dedication.

References

Adam
S.
,
Bonsang
E.
,
Grotz
C.
,
Perelman
S.
(
2013
).
Occupational activity and cognitive reserve: Implications in terms of prevention of cognitive aging and Alzheimer's disease
.
Clinical Interventions in Aging
 ,
8
,
377
390
.
Andel
R.
,
Kareholt
I.
,
Parker
M. G.
,
Thorslund
M.
,
Gatz
M.
(
2007
).
Complexity of primary lifetime occupation and cognition in advanced old age
.
Journal of Aging and Health
 ,
19
,
397
415
.
Andel
R.
,
Silverstein
M.
,
Kareholt
I.
(
2015
).
The role of midlife occupational complexity and leisure activity in late-life cognition
.
Journals of Gerontology Series B: Psychological Sciences and Social Sciences
 ,
70
,
314
321
.
Bigler
E. D.
,
Neeley
E. S.
,
Miller
M. J.
,
Tate
D. F.
,
Rice
S. A.
,
Cleavinger
H.
et al
. (
2004
).
Cerebral volume loss, cognitive deficit and neuropsychological performance: Comparative measures of brain atrophy: I. Dementia
.
Journal of the International Neuropsychological Society
 ,
10
,
442
452
.
Bigler
E. D.
,
Tate
D. F.
(
2001
).
Brain volume, intracranial volume, and dementia
.
Investigative Radiology
 ,
36
,
539
546
.
Braak
H.
,
Braak
E.
(
1997
).
Frequency of stages of Alzheimer-related lesions in different age categories
.
Neurobiology of Aging
 ,
18
,
351
357
.
Dale
A. M.
,
Fischl
B.
,
Sereno
M. I.
(
1999
).
Cortical surface-based analysis. I. Segmentation and surface reconstruction
.
NeuroImage
 ,
9
,
179
194
.
Dowling
N. M.
,
Hermann
B.
,
La Rue
A.
,
Sager
M. A.
(
2010
).
Latent structure and factorial invariance of a neuropsychological test battery for the study of preclinical Alzheimer's disease
.
Neuropsychology
 ,
24
,
742
756
.
Fischl
B.
,
Sereno
M. I.
,
Dale
A. M.
(
1999
).
Cortical surface-based analysis. II: Inflation, flattening, and a surface-based coordinate system
.
NeuroImage
 ,
9
,
195
207
.
Fotenos
A. F.
,
Mintun
M. A.
,
Snyder
A. Z.
,
Morris
J. C.
,
Buckner
R. L.
(
2008
).
Brain volume decline in aging: Evidence for a relation between socioeconomic status, preclinical Alzheimer disease, and reserve
.
Archives of Neurology
 ,
65
,
113
120
.
Fratiglioni
L.
,
Wang
H. X.
,
Ericsson
K.
,
Maytan
M.
,
Winblad
B.
(
2000
).
Influence of social network on occurrence of dementia: A community-based longitudinal study
.
Lancet
 ,
355
,
1315
1319
.
Garibotto
V.
,
Borroni
B.
,
Kalbe
E.
,
Herholz
K.
,
Salmon
E.
,
Holtoff
V.
et al
. (
2008
).
Education and occupation as proxies for reserve in aMCI converters and AD: FDG-PET evidence
.
Neurology
 ,
71
,
1342
1349
.
Garibotto
V.
,
Borroni
B.
,
Sorbi
S.
,
Cappa
S. F.
,
Padovani
A.
,
Perani
D.
(
2012
).
Education and occupation provide reserve in both ApoE epsilon4 carrier and noncarrier patients with probable Alzheimer's disease
.
Neurological Sciences
 ,
33
,
1037
1042
.
Jonaitis
E.
,
La Rue
A.
,
Mueller
K. D.
,
Koscik
R. L.
,
Hermann
B.
,
Sager
M. A.
(
2013
).
Cognitive activities and cognitive performance in middle-aged adults at risk for Alzheimer's disease
.
Psychology and Aging
 ,
28
,
1004
1014
.
Kemppainen
N. M.
,
Aalto
S.
,
Karrasch
M.
,
Nagren
K.
,
Savisto
N.
,
Oikonen
V.
et al
. (
2008
).
Cognitive reserve hypothesis: Pittsburgh Compound B and fluorodeoxyglucose positron emission tomography in relation to education in mild Alzheimer's disease
.
Annals of Neurology
 ,
63
,
112
118
.
Koscik
R. L.
,
La Rue
A.
,
Jonaitis
E. M.
,
Okonkwo
O. C.
,
Johnson
S. C.
,
Bendlin
B. B.
et al
. (
2014
).
Emergence of mild cognitive impairment in late middle-aged adults in the Wisconsin Registry for Alzheimer's Prevention
.
Dementia and Geriatric Cognitive Disorders
 ,
38
,
16
30
.
Lo
R. Y.
,
Jagust
W. J.
(
2013
).
Effect of cognitive reserve markers on Alzheimer pathologic progression
.
Alzheimer Disease and Associated Disorders
 ,
27
,
343
350
.
Michael
Y. L.
,
Carlson
N. E.
,
Chlebowski
R. T.
,
Aickin
M.
,
Weihs
K. L.
,
Ockene
J. K.
et al
. (
2009
).
Influence of stressors on breast cancer incidence in the Women's Health Initiative
.
Health Psychology
 ,
28
,
137
146
.
Ngandu
T.
,
von Strauss
E.
,
Helkala
E. L.
,
Winblad
B.
,
Nissinen
A.
,
Tuomilehto
J.
et al
. (
2007
).
Education and dementia: What lies behind the association?
Neurology
 ,
69
,
1442
1450
.
Okonkwo
O. C.
,
Alosco
M. L.
,
Jerskey
B. A.
,
Sweet
L. H.
,
Ott
B. R.
,
Tremont
G.
(
2010
).
Cerebral atrophy, apolipoprotein E varepsilon4, and rate of decline in everyday function among patients with amnestic mild cognitive impairment
.
Alzheimer's & Dementia
 ,
6
,
404
411
.
Ott
B. R.
,
Cohen
R. A.
,
Gongvatana
A.
,
Okonkwo
O. C.
,
Johanson
C. E.
,
Stopa
E. G.
et al
. (
2010
).
Brain ventricular volume and cerebrospinal fluid biomarkers of Alzheimer's disease
.
Journal of Alzheimer's Disease
 ,
20
,
647
657
.
Perneczky
R.
,
Drzezga
A.
,
Diehl-Schmid
J.
,
Schmid
G.
,
Wohlschlager
A.
,
Kars
S.
et al
. (
2006
).
Schooling mediates brain reserve in Alzheimer's disease: Findings of fluoro-deoxy-glucose-positron emission tomography
.
Journal of Neurology, Neurosurgery and Psychiatry
 ,
77
,
1060
1063
.
Radloff
L. S.
(
1977
).
The CES-D scale: A self-report depression scale for research in the general population
.
Applied Psychological Measurement
 ,
1
,
385
401
.
Reitan
R.
,
Wolfson
D.
(
1993
).
The Halstead-Reitan Neuropsychological Test Battery: Theory and clinical interpretation
 .
Tucson
,
AZ
:
Neuropsychology Press
.
Rentz
D. M.
,
Locascio
J. J.
,
Becker
J. A.
,
Moran
E. K.
,
Eng
E.
,
Buckner
R. L.
et al
. (
2010
).
Cognition, reserve, and amyloid deposition in normal aging
.
Annals of Neurology
 ,
67
,
353
364
.
Roe
C. M.
,
Mintun
M. A.
,
D'Angelo
G.
,
Xiong
C.
,
Grant
E. A.
,
Morris
J. C.
(
2008
).
Alzheimer disease and cognitive reserve: Variation of education effect with carbon 11-labeled Pittsburgh Compound B uptake
.
Archives of Neurology
 ,
65
,
1467
1471
.
Roos
P. A.
,
Treiman
D. J.
(
1980
).
DOT scales for the 1970 Census classification
. In
Miller
A. R.
,
Treiman
D. J.
,
Cain
P. S.
,
Roos
P. A.
(Eds.),
Work, jobs, and occupations: A critical review of occupational titles
  (pp.
336
389
).
Washington, DC
:
National Academy Press
.
Sager
M. A.
,
Hermann
B.
,
La Rue
A.
(
2005
).
Middle-aged children of persons with Alzheimer's disease: APOE genotypes and cognitive function in the Wisconsin Registry for Alzheimer's Prevention
.
Journal of Geriatric Psychiatry and Neurology
 ,
18
,
245
249
.
Scarmeas
N.
,
Zarahn
E.
,
Anderson
K. E.
,
Habeck
C. G.
,
Hilton
J.
,
Flynn
J.
et al
. (
2003
).
Association of life activities with cerebral blood flow in Alzheimer disease: Implications for the cognitive reserve hypothesis
.
Archives of Neurology
 ,
60
,
359
365
.
Schmidt
M.
(
1996
).
Rey Auditory Verbal Learning Test: A handbook
 .
Torrance, CA
:
Western Psychological Services
.
Smart
E. L.
,
Gow
A. J.
,
Deary
I. J.
(
2014
).
Occupational complexity and lifetime cognitive abilities
.
Neurology
 ,
83
,
2285
2291
.
Stern
Y.
(
2012
).
Cognitive reserve in ageing and Alzheimer's disease
.
Lancet Neurology
 ,
11
,
1006
1012
.
Stern
Y.
,
Alexander
G. E.
,
Prohovnik
I.
,
Mayeux
R.
(
1992
).
Inverse relationship between education and parietotemporal perfusion deficit in Alzheimer's disease
.
Annals of Neurology
 ,
32
,
371
375
.
Stern
Y.
,
Alexander
G. E.
,
Prohovnik
I.
,
Stricks
L.
,
Link
B.
,
Lennon
M. C.
et al
. (
1995
).
Relationship between lifetime occupation and parietal flow: Implications for a reserve against Alzheimer's disease pathology
.
Neurology
 ,
45
,
55
60
.
Trenerry
M.
,
Crosson
B.
,
DeBoe
J.
,
Leber
L.
(
1989
).
Stroop neuropsychological screening test
 .
Odessa, FL
:
Psychological Assessment Resources
.
United States Employment Service
. (
1991
).
Dictionary of occupational titles
  (
4th ed.
).
Washington, DC
:
U.S. Government Printing Office
.
Wechsler
D.
(
1997
).
WAIS-III: Wechsler Adult Intelligence Scale
 
—3rd edition
 .
San Antonio, TX
:
The Psychological Corporation
.
Wilson
R. S.
,
Barnes
L. L.
,
Mendes de Leon
C. F.
,
Aggarwal
N. T.
,
Schneider
J. S.
,
Bach
J.
et al
. (
2002
).
Depressive symptoms, cognitive decline, and risk of AD in older persons
.
Neurology
 ,
59
,
364
370
.
Zahodne
L. B.
,
Stern
Y.
,
Manly
J. J.
(
2015
).
Differing effects of education on cognitive decline in diverse elders with low versus high educational attainment
.
Neuropsychology
 ,
29
,
649
657
.
Zuelsdorff
M. L.
,
Engelman
C. D.
,
Friedman
E. M.
,
Koscik
R. L.
,
Jonaitis
E. M.
,
Rue
A. L.
et al
. (
2013
).
Stressful events, social support, and cognitive function in middle-aged adults with a family history of Alzheimer's disease
.
Journal of Aging and Health
 ,
25
,
944
959
.