adni_nardeen_analyses

Description

The following analyses including a correlation matrix and two multiple linear regressions are performed on the ADNI dataset to consider whether occupation dimensions extracted from the ADNI dataset using the VOLCANO tool predict greater cognitive functioning as measured by the Mini Mental State Exam (MMSE). Only baseline measures of the MMSE were included in the current analyses. Sample size total was 2429 participants.

Distribution of Data

Correlation Matrix

##                         MMSE RAVLT_Immediate RAVLT_learning RAVLT_forgetting
## MMSE              1.00000000      0.60300327    0.490883194    -0.0801876749
## RAVLT_Immediate   0.60300327      1.00000000    0.641364387    -0.1717166395
## RAVLT_learning    0.49088319      0.64136439    1.000000000    -0.0074482142
## RAVLT_forgetting -0.08018767     -0.17171664   -0.007448214     1.0000000000
## Complexity       -0.13525931     -0.12004092   -0.075318130     0.0006768223
## STEM             -0.01798079     -0.06647532   -0.044793278    -0.0355076024
## MathVSHealth      0.04121008      0.08914927    0.046999657    -0.0439897256
## Age              -0.17271387     -0.26211314   -0.159493107     0.0411025374
##                     Complexity         STEM MathVSHealth          Age
## MMSE             -0.1352593080 -0.017980792   0.04121008 -0.172713874
## RAVLT_Immediate  -0.1200409225 -0.066475323   0.08914927 -0.262113136
## RAVLT_learning   -0.0753181301 -0.044793278   0.04699966 -0.159493107
## RAVLT_forgetting  0.0006768223 -0.035507602  -0.04398973  0.041102537
## Complexity        1.0000000000 -0.003855196   0.02832930 -0.004337896
## STEM             -0.0038551961  1.000000000   0.03674071  0.144560529
## MathVSHealth      0.0283292985  0.036740709   1.00000000 -0.039360945
## Age              -0.0043378955  0.144560529  -0.03936094  1.000000000

Analysis with MMSE and Occupation

Regression with MMSE without age covariate

Dimension 1 is occuaptional complexity. Dimension 2 is whether a field is in STEM or not. Dimension 3 is whether a field is more in a mathematical or health domain (health domain would result in negative relationship). The findings suggest that greater occupational complexity predicts higher MMSE scores. Having a job that is more in the health field would result in a negative relationship. This is indicated in the result suggesting have a job in the health field is related to higher MMSE scores.

## 
## Call:
## lm(formula = MMSE ~ dim1_all + dim2_all + dim3_all, data = data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -10.4549  -1.4809   0.7914   2.0722   3.7363 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 27.11023    0.07522 360.401  < 2e-16 ***
## dim1_all     0.07417    0.01164   6.373 2.27e-10 ***
## dim2_all     0.01189    0.01262   0.942   0.3463    
## dim3_all    -0.03157    0.01478  -2.136   0.0328 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.577 on 2132 degrees of freedom
##   (293 observations deleted due to missingness)
## Multiple R-squared:  0.02073,    Adjusted R-squared:  0.01935 
## F-statistic: 15.05 on 3 and 2132 DF,  p-value: 1.084e-09

## R-squared value: 0.02073198

## [1] "n = 2429"

Regression with MMSE with age covariate

When age is added to the model as a covariate, dimension 3 (mathvshealth field) is no longer significant. Only dimenision 1 (occupational complexity) remains significant in predicting higher MMSE scores.

## 
## Call:
## lm(formula = MMSE ~ dim1_all + dim2_all + dim3_all + AGE, data = df.adni.bl)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -10.4612  -1.3845   0.7434   1.9147   4.2421 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 31.509886   0.555641  56.709  < 2e-16 ***
## dim1_all    -0.073845   0.011494  -6.424 1.63e-10 ***
## dim2_all     0.002519   0.012585   0.200    0.841    
## dim3_all     0.026265   0.014589   1.800    0.072 .  
## AGE         -0.060291   0.007548  -7.988 2.22e-15 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.54 on 2128 degrees of freedom
##   (296 observations deleted due to missingness)
## Multiple R-squared:  0.04894,    Adjusted R-squared:  0.04715 
## F-statistic: 27.38 on 4 and 2128 DF,  p-value: < 2.2e-16

```