Deformation Fields and Grey Matter vs Grey Matter and White Matter¶

Data Loading¶

suppressWarnings(suppressMessages(library("tidyverse")))
suppressWarnings(suppressMessages(library("ggpubr")))
suppressWarnings(suppressMessages(library("rstatix")))

library(tidyverse)
library(ggpubr)
library(rstatix)


map_abb <- list("df_gm"= "DF+ǴM", "gm_wm"= "GM+WM")

all_results <- c()

df_eval <- list("crossval"= data.frame(), "test"= data.frame())


for (evaluation in names(df_eval)){
    for (mod in names(map_abb)){

      filename <- sprintf("%s_%s.csv", mod, evaluation)
      path_col <- file.path(getwd(), "..", "..", "..", "resources", filename)
      
      df_mod <- read.csv(path_col)
      df_mod$fusion <- map_abb[[mod]]

      df_eval[[evaluation]] <- rbind(df_eval[[evaluation]], df_mod)
    }
}
df_val <- df_eval[["crossval"]]
df_test <- df_eval[["test"]]

Validation data¶

head(df_val)

A data.frame: 6 × 4
	fold	r2	MAE	fusion
	<int>	<dbl>	<dbl>	<chr>
1	0	0.8026467	5.740682	DF+ǴM
2	1	0.7473344	6.313860	DF+ǴM
3	2	0.7362875	5.625505	DF+ǴM
4	3	0.7456859	5.830985	DF+ǴM
5	4	0.7941903	5.643035	DF+ǴM
6	5	0.7565052	5.573205	DF+ǴM

aggregate(df_val[,c("r2", "MAE")], list(df_val$fusion), mean)

A data.frame: 2 × 3
Group.1	r2	MAE
<chr>	<dbl>	<dbl>
DF+ǴM	0.7884606	5.548079
GM+WM	0.7896436	5.445226

Test data¶

head(df_test)

A data.frame: 6 × 5
	subjectID	gender	age	y_hat	fusion
	<chr>	<int>	<dbl>	<dbl>	<chr>
1	sub-474	1	34.01	40.33901	DF+ǴM
2	sub-292	1	23.73	31.65292	DF+ǴM
3	sub-382	1	65.91	46.45294	DF+ǴM
4	sub-434	1	67.24	56.62014	DF+ǴM
5	sub-395	0	45.03	35.61988	DF+ǴM
6	sub-306	1	30.18	39.87024	DF+ǴM

df_test["abs_diff"] <- abs(df_test$y_hat - df_test$age)

result_mean <- aggregate(df_test[,c("age", "y_hat", "abs_diff")], list(df_test$fusion), mean)
result_mean

A data.frame: 2 × 4
Group.1	age	y_hat	abs_diff
<chr>	<dbl>	<dbl>	<dbl>
DF+ǴM	42.37309	45.59776	6.896464
GM+WM	42.37309	39.80127	7.919839

Statistics¶

Paired t-test¶

colors <-c("#AB1111", "#143F6B")

pwc <- df_val %>%
  pairwise_t_test(
    MAE ~ fusion, paired = TRUE, alternative="less"
  )
pwc

options(repr.plot.width = 7, repr.plot.height = 5)

bxp <- ggboxplot(df_val, x = "fusion", y = "MAE", add = "point", fill = colors,
                 xlab = "Fusion", ylab = "MAE") + theme(text = element_text(size=20))


pwc <- pwc %>% add_xy_position(x = "fusion")
bxp <- bxp + 
  stat_pvalue_manual(pwc, label = "p = {p.adj}") +
  labs(
    subtitle = get_test_label(pwc, detailed = TRUE),
  ) + geom_point(data = data.frame(x = factor(result_mean[["Group.1"]]), y = result_mean[["abs_diff"]]),
                   aes(x=x, y=y),
                   color = 'red', shape=8, size=5)

bxp 

A rstatix_test: 1 × 10
	.y.	group1	group2	n1	n2	statistic	df	p	p.adj	p.adj.signif
	<chr>	<chr>	<chr>	<int>	<int>	<dbl>	<dbl>	<dbl>	<dbl>	<chr>
1	MAE	DF+ǴM	GM+WM	30	30	0.6854774	29	0.751	0.751	ns

Warning message in if (fill %in% names(data) & is.null(add.params$fill)) add.params$fill <- fill:
“the condition has length > 1 and only the first element will be used”

Brain Age and Deformation Fields

Deformation Fields and Grey Matter vs Grey Matter and White Matter

Contents

Deformation Fields and Grey Matter vs Grey Matter and White Matter¶

Data Loading¶

Validation data¶

Test data¶

Statistics¶

Paired t-test¶