AV weekly meeting report 01 GBSS of PACTMD Data

meeting date: 2019-03-13
days until:
last run: 2019-03-25
link: http://rpubs.com/janderz8/469731

---

Summary

This week I have looked at the updated information from the baseline PACt-MD dataset. Previously, using the provisional diagnoses, I had analyzed 183 individuals. Using the updated baseline information, We now have access to 283 individuals. Five of those individuals have no current consensus diagnosis “sub-CMH0100009” “sub-CMH3060027” “sub-CMHH3060187” “sub-CMH3070023” “sub-CMH3070112”, and one individual does not appear in the Excel files “sub-CMH0350083”. I have included the five subjects without confirmed diagnoses by substituting in their preliminary diagnosis - I can exclude these individuals later if necessary.

I am currently in the process of updating the PLS and univariate analyses and will include descriptive statistics for the included sample here as well.

Finally, I have spoken to Jerry about pushing the NODDI metrics into surface space using CIFTIFY since the grey matter skeletons are fairly poor quality & smoothing on the surface has the advantage of increasing power & not smearing across sulci/gyri.

It may be useful to quantify overall level of free water using fslmeants since free water may not be distributed similarly by group.

---

Participant Numbers By Group

0 = Normal 1 = MCI 2 = MDD-MCI 3 = MDD+MCI 4 = Dementia

NOTE: sub-CMH0350083 was missing from the most recent release of the data. I have preprocessed their MRI images, but don’t have a diagnosis. Follow up with Neda.

Demographics for Included Subjects

first look at the control subjects

Check DOB for PAC01_CMH_0950016_01, is likely supposed to be May 1925 Now the patients

Now to join the controls and the patients & then subset them by to include only those with MRI scans…

Imputing missing data using the MICE package:

drop the one person’s MoCA score of 0 (they had an MMSE of 26) - follow up with this person PAC01_CMH_0950019_01

describeBy(completedData, group="Diagnosis", skew=FALSE, range=FALSE,quant=NULL, IQR=FALSE)

## 
##  Descriptive statistics by group 
## group: HC
##                          vars  n  mean    sd   se
## demo_id*                    1 25   NaN    NA   NA
## MRI*                        2 25   NaN    NA   NA
## Diagnosis*                  3 25  1.00  0.00 0.00
## all_dx*                     4 25  1.00  0.00 0.00
## cc_mci_a2_cog_d_lm*         5 25  1.12  0.33 0.07
## demo_gender                 6 25  1.64  0.49 0.10
## sdemo_country_birth*        7 25   NaN    NA   NA
## sdemo_age_immigration       8 25  9.56 15.95 3.19
## sdemo_prime_language*       9 25   NaN    NA   NA
## sdemo_employment           10 25  1.72  1.67 0.33
## sdemo_marital_status       11 25  2.28  1.10 0.22
## sdemo_children             12 25  1.28  0.46 0.09
## sdemo_religion             13 25  2.56  1.39 0.28
## sdemo_sbjct_highest_ed     14 25  6.24  0.88 0.18
## sdemo_sbjct_highest_occ    15 25  7.52  1.29 0.26
## sdemo_father_ed            16 25  4.84  1.89 0.38
## sdemo_father_highest_occ   17 25  5.40  2.36 0.47
## sdemo_mother_ed            18 25  3.64  1.75 0.35
## sdemo_mother_highest_occ   19 25  4.20  2.61 0.52
## moca_total_score           20 25 27.80  1.35 0.27
## age                        21 25 69.73  6.02 1.20
## -------------------------------------------------------- 
## group: NonAmnestic_MCI
##                          vars  n  mean    sd   se
## demo_id*                    1 35   NaN    NA   NA
## MRI*                        2 35   NaN    NA   NA
## Diagnosis*                  3 35  2.00  0.00 0.00
## all_dx*                     4 35  2.00  0.00 0.00
## cc_mci_a2_cog_d_lm*         5 35  1.00  0.00 0.00
## demo_gender                 6 35  1.57  0.50 0.08
## sdemo_country_birth*        7 35   NaN    NA   NA
## sdemo_age_immigration       8 35 13.71 14.90 2.52
## sdemo_prime_language*       9 35   NaN    NA   NA
## sdemo_employment           10 35  1.71  1.86 0.31
## sdemo_marital_status       11 35  2.40  1.24 0.21
## sdemo_children             12 35  1.26  0.44 0.07
## sdemo_religion             13 35  3.14  1.91 0.32
## sdemo_sbjct_highest_ed     14 35  6.00  1.00 0.17
## sdemo_sbjct_highest_occ    15 35  7.09  1.36 0.23
## sdemo_father_ed            16 35  4.71  1.67 0.28
## sdemo_father_highest_occ   17 35  5.97  1.93 0.33
## sdemo_mother_ed            18 35  3.80  1.86 0.31
## sdemo_mother_highest_occ   19 35  4.37  2.73 0.46
## moca_total_score           20 35 24.66  2.10 0.35
## age                        21 35 71.73  6.48 1.09
## -------------------------------------------------------- 
## group: MDD-MCI
##                          vars  n  mean    sd   se
## demo_id*                    1 48   NaN    NA   NA
## MRI*                        2 48   NaN    NA   NA
## Diagnosis*                  3 48  3.00  0.00 0.00
## all_dx*                     4 48  3.00  0.00 0.00
## cc_mci_a2_cog_d_lm*         5 48  1.27  0.45 0.06
## demo_gender                 6 48  1.67  0.48 0.07
## sdemo_country_birth*        7 48   NaN    NA   NA
## sdemo_age_immigration       8 48  8.98 14.62 2.11
## sdemo_prime_language*       9 48   NaN    NA   NA
## sdemo_employment           10 48  1.42  1.18 0.17
## sdemo_marital_status       11 48  2.67  1.24 0.18
## sdemo_children             12 48  1.33  0.48 0.07
## sdemo_religion             13 48  3.52  1.54 0.22
## sdemo_sbjct_highest_ed     14 48  6.02  1.12 0.16
## sdemo_sbjct_highest_occ    15 48  7.10  1.29 0.19
## sdemo_father_ed            16 48  3.79  1.98 0.29
## sdemo_father_highest_occ   17 48  5.69  2.21 0.32
## sdemo_mother_ed            18 48  3.81  1.79 0.26
## sdemo_mother_highest_occ   19 48  4.19  2.48 0.36
## moca_total_score           20 48 27.54  1.95 0.28
## age                        21 48 70.55  4.96 0.72
## -------------------------------------------------------- 
## group: NonAmnestic_MDD+MCI
##                          vars  n  mean    sd   se
## demo_id*                    1 28   NaN    NA   NA
## MRI*                        2 28   NaN    NA   NA
## Diagnosis*                  3 28  4.00  0.00 0.00
## all_dx*                     4 28  3.96  0.19 0.04
## cc_mci_a2_cog_d_lm*         5 28  1.00  0.00 0.00
## demo_gender                 6 28  1.71  0.46 0.09
## sdemo_country_birth*        7 28   NaN    NA   NA
## sdemo_age_immigration       8 28 16.64 20.93 3.95
## sdemo_prime_language*       9 28   NaN    NA   NA
## sdemo_employment           10 28  1.50  1.04 0.20
## sdemo_marital_status       11 28  2.64  1.34 0.25
## sdemo_children             12 28  1.36  0.49 0.09
## sdemo_religion             13 28  3.36  1.54 0.29
## sdemo_sbjct_highest_ed     14 28  6.11  0.83 0.16
## sdemo_sbjct_highest_occ    15 28  7.29  1.30 0.25
## sdemo_father_ed            16 28  4.54  2.08 0.39
## sdemo_father_highest_occ   17 28  6.46  2.30 0.43
## sdemo_mother_ed            18 28  3.36  2.08 0.39
## sdemo_mother_highest_occ   19 28  4.25  2.41 0.46
## moca_total_score           20 28 25.57  2.71 0.51
## age                        21 28 70.93  4.77 0.90
## -------------------------------------------------------- 
## group: Dementia
##                          vars n  mean    sd  se
## demo_id*                    1 2   NaN    NA  NA
## MRI*                        2 2   NaN    NA  NA
## Diagnosis*                  3 2  5.00  0.00 0.0
## all_dx*                     4 2  5.00  0.00 0.0
## cc_mci_a2_cog_d_lm*         5 2  2.00  0.00 0.0
## demo_gender                 6 2  2.00  0.00 0.0
## sdemo_country_birth*        7 2   NaN    NA  NA
## sdemo_age_immigration       8 2  0.00  0.00 0.0
## sdemo_prime_language*       9 2   NaN    NA  NA
## sdemo_employment           10 2  1.00  0.00 0.0
## sdemo_marital_status       11 2  4.00  0.00 0.0
## sdemo_children             12 2  1.00  0.00 0.0
## sdemo_religion             13 2  5.00  0.00 0.0
## sdemo_sbjct_highest_ed     14 2  5.00  1.41 1.0
## sdemo_sbjct_highest_occ    15 2  7.50  2.12 1.5
## sdemo_father_ed            16 2  1.00  0.00 0.0
## sdemo_father_highest_occ   17 2  7.00  0.00 0.0
## sdemo_mother_ed            18 2  2.00  1.41 1.0
## sdemo_mother_highest_occ   19 2  5.00  2.83 2.0
## moca_total_score           20 2 16.50  0.71 0.5
## age                        21 2 78.25 10.90 7.7
## -------------------------------------------------------- 
## group: Amnestic_MCI
##                          vars  n  mean    sd   se
## demo_id*                    1 94   NaN    NA   NA
## MRI*                        2 94   NaN    NA   NA
## Diagnosis*                  3 94  6.00  0.00 0.00
## all_dx*                     4 94  2.02  0.21 0.02
## cc_mci_a2_cog_d_lm*         5 94  1.98  0.15 0.01
## demo_gender                 6 94  1.54  0.50 0.05
## sdemo_country_birth*        7 94   NaN    NA   NA
## sdemo_age_immigration       8 94 10.96 13.65 1.41
## sdemo_prime_language*       9 93   NaN    NA   NA
## sdemo_employment           10 94  1.79  1.78 0.18
## sdemo_marital_status       11 94  2.78  1.36 0.14
## sdemo_children             12 94  1.20  0.40 0.04
## sdemo_religion             13 94  2.85  1.67 0.17
## sdemo_sbjct_highest_ed     14 94  5.59  1.35 0.14
## sdemo_sbjct_highest_occ    15 94  7.01  1.65 0.17
## sdemo_father_ed            16 94  3.16  2.19 0.23
## sdemo_father_highest_occ   17 94  5.80  2.27 0.23
## sdemo_mother_ed            18 94  2.85  1.94 0.20
## sdemo_mother_highest_occ   19 94  3.78  2.43 0.25
## moca_total_score           20 94 23.02  2.54 0.26
## age                        21 94 72.96  7.61 0.78
## -------------------------------------------------------- 
## group: Amnestic_MDD+MCI
##                          vars  n  mean    sd   se
## demo_id*                    1 50   NaN    NA   NA
## MRI*                        2 50   NaN    NA   NA
## Diagnosis*                  3 50  7.00  0.00 0.00
## all_dx*                     4 50  4.00  0.00 0.00
## cc_mci_a2_cog_d_lm*         5 50  1.98  0.14 0.02
## demo_gender                 6 50  1.52  0.50 0.07
## sdemo_country_birth*        7 50   NaN    NA   NA
## sdemo_age_immigration       8 50 14.08 15.37 2.17
## sdemo_prime_language*       9 50   NaN    NA   NA
## sdemo_employment           10 50  1.40  1.16 0.16
## sdemo_marital_status       11 50  2.52  1.05 0.15
## sdemo_children             12 50  1.24  0.43 0.06
## sdemo_religion             13 50  3.12  1.85 0.26
## sdemo_sbjct_highest_ed     14 50  5.66  1.12 0.16
## sdemo_sbjct_highest_occ    15 50  6.54  1.45 0.20
## sdemo_father_ed            16 50  3.30  2.05 0.29
## sdemo_father_highest_occ   17 50  4.92  2.58 0.36
## sdemo_mother_ed            18 50  3.46  1.82 0.26
## sdemo_mother_highest_occ   19 50  4.08  2.41 0.34
## moca_total_score           20 50 24.00  2.64 0.37
## age                        21 50 71.51  4.60 0.65

Now to calculate a “reserve” factor score

To do this, I will need to calculate the residual of every variable of interest from age, standardize them, and then feed them into a factor model.

library(psych)
library(GPArotation)
data_for_factor <- subset(completedData, select=c("age","sdemo_sbjct_highest_ed","sdemo_sbjct_highest_occ", "sdemo_father_ed", "sdemo_mother_ed","sdemo_father_highest_occ","sdemo_mother_highest_occ")) %>% scale()

#after running this once, only education and occupation scores covaried).  Gender, marital status, religion and age of immigration were dropped ,"sdemo_marital_status","sdemo_religion","sdemo_children","sdemo_age_immigration","demo_gender",

age_factor <- data_for_factor[,1]
data_for_factor <- data_for_factor[,-1]
#residualize age from the above...
age_resids <- lm(as.matrix(data_for_factor[,1:6]) ~ age_factor)$resid
heatmap(cor(age_resids))

parallel <- fa.parallel(age_resids, fm = 'ml', fa = 'fa')

## Parallel analysis suggests that the number of factors =  3  and the number of components =  NA

two <- fa(age_resids,nfactors = 2,rotate = "varimax",fm="ml")
print(two)

## Factor Analysis using method =  ml
## Call: fa(r = age_resids, nfactors = 2, rotate = "varimax", fm = "ml")
## Standardized loadings (pattern matrix) based upon correlation matrix
##                           ML2  ML1   h2    u2 com
## sdemo_sbjct_highest_ed   0.11 0.99 1.00 0.005 1.0
## sdemo_sbjct_highest_occ  0.13 0.53 0.30 0.700 1.1
## sdemo_father_ed          0.61 0.23 0.43 0.570 1.3
## sdemo_mother_ed          0.84 0.16 0.73 0.275 1.1
## sdemo_father_highest_occ 0.44 0.09 0.20 0.799 1.1
## sdemo_mother_highest_occ 0.63 0.03 0.40 0.602 1.0
## 
##                        ML2  ML1
## SS loadings           1.70 1.35
## Proportion Var        0.28 0.23
## Cumulative Var        0.28 0.51
## Proportion Explained  0.56 0.44
## Cumulative Proportion 0.56 1.00
## 
## Mean item complexity =  1.1
## Test of the hypothesis that 2 factors are sufficient.
## 
## The degrees of freedom for the null model are  15  and the objective function was  1.58 with Chi Square of  439.99
## The degrees of freedom for the model are 4  and the objective function was  0.23 
## 
## The root mean square of the residuals (RMSR) is  0.07 
## The df corrected root mean square of the residuals is  0.14 
## 
## The harmonic number of observations is  282 with the empirical chi square  47.15  with prob <  1.4e-09 
## The total number of observations was  282  with Likelihood Chi Square =  64.06  with prob <  4.1e-13 
## 
## Tucker Lewis Index of factoring reliability =  0.467
## RMSEA index =  0.233  and the 90 % confidence intervals are  0.183 0.283
## BIC =  41.49
## Fit based upon off diagonal values = 0.95
## Measures of factor score adequacy             
##                                                    ML2  ML1
## Correlation of (regression) scores with factors   0.89 1.00
## Multiple R square of scores with factors          0.80 0.99
## Minimum correlation of possible factor scores     0.60 0.98

fa.diagram(two)

#now to estimate factor scores per subject
FS <- as.data.frame(factor.scores(age_resids, two, Phi = NULL, method = "Thurstone")$scores)
FS$Cumulative <-FS$ML2 * .31 + FS$ML1*.22
completedData <- cbind(completedData, FS)
#plot "reserve factor by group"
pd <- position_dodge(0.1) # move them .05 to the left and right
completedData$Cumulative <- as.numeric(completedData$Cumulative)

ggplot(completedData, aes(x = factor(all_dx), fill = factor(all_dx), y = Cumulative)) +
  geom_dotplot(binaxis = "y", stackdir = "center", position = "dodge",colour="black",size=.5)+facet_grid(.~cc_mci_a2_cog_d_lm)+
    stat_summary(fun.y = mean, fun.ymin = mean, fun.ymax = mean,
                 geom = "crossbar", width = 0.5)

## Warning: Ignoring unknown parameters: size

## `stat_bindot()` using `bins = 30`. Pick better value with `binwidth`.

reserve_info <- summarySE(data = subset(completedData, Diagnosis !="Dementia" & Diagnosis !="Amnestic_HC"),measurevar = "Cumulative",groupvars =c("all_dx","cc_mci_a2_cog_d_lm"),na.rm = TRUE)
#reserve_info$all_dx <- factor(reserve_info$all_dx, levels = c("HC", "MDD-MCI", "MCI", "MDD+MCI"))

ggplot(reserve_info,aes(x=all_dx, y=Cumulative, group=all_dx, fill =all_dx))+geom_bar(stat="identity", colour="black")+ 
    geom_errorbar(aes(ymin=Cumulative-se, ymax=Cumulative+se), width=.1, position=pd, colour="black")+labs(title= "Level of Reserve by Group (error bars are +/- 1SE)")+geom_hline(yintercept = 0, colour="black", size = 1)+facet_grid(.~cc_mci_a2_cog_d_lm)

#adding in the one case who is missing from the baseline data so that I have the same number of rows in the dataframe.
missing_mri <- data.frame(MRI="sub-CMH0350083")
completedData <- merge(completedData, missing_mri, by ="MRI", all=TRUE)
completedData <- completedData[order(completedData$MRI),]
completedData$MRI_Name <- paste(paste(completedData$all_dx, completedData$cc_mci_a2_cog_d_lm,"${i}", completedData$MRI,sep="_"),".nii.gz", sep="")
completedData <- subset(completedData, MRI != "sub-CMHH3070015")
write.csv(completedData, "/projects/janderson/PACTMD/data/demographic/PACt-MD_T0_Baseline_Data/Demo_Data_with_Reserve_Score.csv")

Now plotting “reserve factors against MOCA scores”

cor.test(completedData$Cumulative, completedData$moca_total_score)

## 
##  Pearson's product-moment correlation
## 
## data:  completedData$Cumulative and completedData$moca_total_score
## t = 4.4954, df = 279, p-value = 1.019e-05
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.1473510 0.3657719
## sample estimates:
##       cor 
## 0.2598825

ggplot(completedData, aes(x=Cumulative, y=moca_total_score))+geom_point()+geom_smooth(method="rlm")#+facet_grid(all_dx~cc_mci_a2_cog_d_lm)

## Warning: Removed 1 rows containing non-finite values (stat_smooth).

## Warning: Removed 1 rows containing missing values (geom_point).

cog_reserve_mod <- rlm(moca_total_score ~ Cumulative, data = completedData)
f.robftest(cog_reserve_mod, var = "Cumulative")

## 
##  robust F-test (as if non-random weights)
## 
## data:  from rlm(formula = moca_total_score ~ Cumulative, data = completedData)
## F = 20.443, p-value = 9.091e-06
## alternative hypothesis: true Cumulative is not equal to 0

reserve_pred <- subset(completedData, select=c("demo_id","Diagnosis","sdemo_sbjct_highest_ed","sdemo_sbjct_highest_occ", "sdemo_father_ed", "sdemo_mother_ed","sdemo_father_highest_occ","sdemo_mother_highest_occ","Cumulative"))
reserve_out <- subset(completedData, select=c("demo_id","age","moca_total_score"))

reserve_pred_long <- melt(data=reserve_pred, id.vars=c("demo_id","Diagnosis"))
reserve_pred_long <- merge(reserve_pred_long, reserve_out, by.x ="demo_id")
ggplot(reserve_pred_long, aes(x=value, y=moca_total_score))+geom_point()+geom_smooth(method="rlm")+facet_grid(.~variable, scales="free")+labs(title= "prediction of MoCA by various reserve scores")

## Warning: Removed 7 rows containing non-finite values (stat_smooth).

## Warning: Removed 7 rows containing missing values (geom_point).

ggplot(reserve_pred_long, aes(x=value, y=age))+geom_point()+geom_smooth(method="rlm")+facet_grid(.~variable,scales = "free")+labs(title= "prediction of age by various reserve scores")

## Warning: Removed 7 rows containing non-finite values (stat_smooth).

## Warning: Removed 7 rows containing missing values (geom_point).

Incorporating NODDI MRI data…

First let’s take a look at the relationship with age…

library(R.matlab)
library(readxl)
library(ggthemes)

# load in the data here...
# First white matter and age data
wm_CSF_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_CSF_wmskel_HC_STRUCTresult.mat")
wm_ODI_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_ODI_wmskel_HC_STRUCTresult.mat")
wm_NDI_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_NDI_wmskel_HC_STRUCTresult.mat")

# Then grey matter and age data
gm_CSF_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_CSF_gmskel_STRUCTresult.mat")
gm_ODI_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_ODI_gmskel_STRUCTresult.mat")
gm_NDI_Age_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/analysis_age_HcDepMciMciDep_NDI_gmskel_STRUCTresult.mat")

# Now white matter data
wm_CSF_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_wmskel_BOTH_CSF_STRUCTresult.mat")
wm_ODI_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_wmskel_BOTH_ODI_STRUCTresult.mat")
wm_NDI_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_wmskel_BOTH_NDI_STRUCTresult.mat")

# Now grey matter data
gm_CSF_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_gmskel_BOTH_CSF_STRUCTresult.mat")
gm_ODI_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_gmskel_BOTH_ODI_STRUCTresult.mat")
gm_NDI_Mat <-("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/Analysis_Group_gmskel_BOTH_NDI_STRUCTresult.mat")

descriptives <- read_excel("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/subset_with_controls/descriptives.xls")

## readxl works best with a newer version of the tibble package.
## You currently have tibble v1.4.2.
## Falling back to column name repair from tibble <= v1.4.2.
## Message displays once per session.

#function to pull data from Mean Behavioural PLS outputs...
Beh_PLS_Result <- function(x, LV, label, Invert){
input <- readMat(x)
res <- input$result
boot_result <- res[[14]]
orig_cors <- as.data.frame(boot_result[[6]][,LV])
upper <- as.data.frame(boot_result[[7]][,LV])
lower <- as.data.frame(boot_result[[8]][,LV])
bar <- cbind(orig_cors, upper, lower)
colnames(bar) <- c("Correlation", "upper","lower")
bar$Correlation <- bar$Correlation * Invert
bar$upper <- bar$upper * Invert
bar$lower <- bar$lower * Invert
bar$upper_adj <- abs(bar$upper - bar$Correlation)
bar$lower_adj <- abs(bar$Correlation - bar$lower)
group <- c("HC", "Dep", "MCI", "MCI+Dep")
#Amnesia <- c("Non Amnestic","Non Amnestic","Non Amnestic","Amnestic","Amnestic")
Label <- c(label, label, label, label)
bar <- cbind(Label, group, bar)
bar$group <- factor(bar$group, levels = c("HC", "Dep", "MCI", "MCI+Dep")) 
#bar$Amnesia <- factor(bar$Amnesia, levels = c("Non Amnestic", "Amnestic")) 
age <- as.data.frame(res[[10]])
subs <- as.data.frame(res[[11]])
brain_score <- as.data.frame(res[[8]][,1])
return(bar)
}

#Now creating the wm_age datamats
wm_CSF_Age_Dat <- Beh_PLS_Result(wm_CSF_Age_Mat,1,"CSF",1)
wm_NDI_Age_Dat <- Beh_PLS_Result(wm_NDI_Age_Mat,1,"NDI",1)
wm_ODI_Age_Dat <- Beh_PLS_Result(wm_ODI_Age_Mat,1,"ODI",1)

#Now creating the gm_age datamats
gm_CSF_Age_Dat <- Beh_PLS_Result(gm_CSF_Age_Mat,3,"CSF",1)
gm_NDI_Age_Dat <- Beh_PLS_Result(gm_NDI_Age_Mat,3,"NDI",1)
gm_ODI_Age_Dat <- Beh_PLS_Result(gm_ODI_Age_Mat,3,"ODI",1)

wm_age <- rbind(wm_ODI_Age_Dat,wm_NDI_Age_Dat,wm_CSF_Age_Dat)
gm_age <- rbind(gm_ODI_Age_Dat,gm_NDI_Age_Dat,gm_CSF_Age_Dat)

bar_age_wm <- ggplot(wm_age, aes(x = group, y = Correlation, fill= group))+geom_bar(stat = "identity")+
  geom_errorbar(aes(ymin=Correlation-lower_adj, ymax=Correlation + upper_adj), width=0,
                 position=position_dodge(.9), colour="black")  + facet_wrap(~Label)+ geom_hline(yintercept = 0, colour="black")+theme_gdocs() + scale_fill_gdocs()

bar_age_gm <- ggplot(gm_age, aes(x = group, y = Correlation, fill= group))+geom_bar(stat = "identity")+
  geom_errorbar(aes(ymin=Correlation-lower_adj, ymax=Correlation + upper_adj), width=0,
                 position=position_dodge(.9), colour="black")  + facet_wrap(~Label)+ geom_hline(yintercept = 0, colour="black")+theme_gdocs() + scale_fill_gdocs()
#################################################
# Plot the Mean Centered (Group Comparisons)
#################################################

#function to pull data from Mean Centered PLS outputs...
MC_PLS_Result <- function(x, LV, label, Invert, groups){
input <- readMat(x)
res <- input$result
boot_result <- res[[11]]
orig_usc <- as.data.frame(boot_result[[7]][,LV])
upper <- as.data.frame(boot_result[[8]][,LV])
lower <- as.data.frame(boot_result[[9]][,LV])
bar <- cbind(orig_usc, upper, lower)
colnames(bar) <- c("BSR", "upper","lower")
bar$BSR <- bar$BSR * Invert
bar$upper <- bar$upper * Invert
bar$lower <- bar$lower * Invert
bar$upper_adj <- bar$upper - bar$BSR 
bar$lower_adj <- bar$BSR - bar$lower
group <- seq(1:groups)
#Amnesia <- c("Non Amnestic","Non Amnestic","Non Amnestic","Amnestic","Amnestic")
Label <- c(label, label, label, label)
bar <- cbind(Label, group, bar)
bar$group <- factor(bar$group) 
#bar$Amnesia <- factor(bar$Amnesia, levels = c("Non Amnestic", "Amnestic")) 
return(bar)
}

#wm data
wm_CSF_Dat <- MC_PLS_Result(wm_CSF_Mat,1,"CSF",1, 4)
wm_NDI_Dat <- MC_PLS_Result(wm_NDI_Mat,1,"NDI",1, 4)
wm_ODI_Dat <- MC_PLS_Result(wm_ODI_Mat,1,"ODI",1, 4)
#gm data
gm_CSF_Dat <- MC_PLS_Result(gm_CSF_Mat,1,"CSF",1, 4)
gm_NDI_Dat <- MC_PLS_Result(gm_NDI_Mat,1,"NDI",1, 4)
gm_ODI_Dat <- MC_PLS_Result(gm_ODI_Mat,1,"ODI",1, 4)

#age <- as.data.frame(res[[10]])
#subs <- as.data.frame(res[[11]])
#brain_score <- as.data.frame(res[[8]][,1])

#toplot <- cbind(descriptives, brain_score)
#toplot$Group <- factor(toplot$Group, levels = c("HC", "Dep", "MCI", "MCI+Dep"))

#scatter_age_csf <- ggplot(toplot, aes(x=Age, y=`res[[8]][, 1]`, group=Group, fill=Group,colour=Group))+geom_point()+geom_smooth(method="rlm")+facet_grid(.~Group) + ylab("Brain Score /n Free Water")
age_csf_brain  <-ggdraw() + draw_image("/projects/janderson/PACTMD/pipelines/NODDI_mdt/my_gbss/stats/Figures/Age_by_CSF.png")

#combining the brain figure with the PLS data

##################################################
# Now looking at the boostrapped correlations
##################################################
#plot_grid(age_csf_brain,bar_age_csf, scatter_age_csf, labels ="AUTO", nrow = 3, ncol =1)

Forays into suface space…

## Parsed with column specification:
## cols(
##   ODI = col_double(),
##   NDI = col_double(),
##   CSF = col_double(),
##   label = col_character()
## )

## Joining, by = "label"

## Joining, by = "label"

Now look at surface correlations with age and with reserve…

## Parsed with column specification:
## cols(
##   ODI = col_double(),
##   NDI = col_double(),
##   CSF = col_double(),
##   label = col_character()
## )

## Joining, by = "label"
## Joining, by = "label"