Multivariate_fMRI_Analysis/One_class_SVM_TPJ_across_Gestalt_Object.R at main · jorennig/Multivariate_fMRI_Analysis · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
# Clear environment
rm(list=ls())

pwd <- dirname(rstudioapi::getSourceEditorContext()$path)
setwd(pwd)

library(e1071)
library(caTools)
library(dplyr)


## Functions
norm_data <- function(x){
  (x-mean(x))/(sd(x))
}

t_test_col <- function(x,y=NA){

  t_test <- matrix(ncol=2, nrow=ncol(x))
  col_names_t <- c("t","p")

  for (t_index in 1:ncol(x)){
    col_c <- x[,t_index]
    t_test_c <- t.test(col_c, mu=0.5)
    t_test[t_index,1] <- t_test_c$statistic
    t_test[t_index,2] <- t_test_c$p.value
  }
  t_test <- data.frame(t_test)
  colnames(t_test) <- col_names_t
  if (!is.na(y)){
    rownames(t_test) <- y
  }
  return(t_test)
}


## Prepare
sub_dir <- dir('.')[file.info(dir('.',full.names=T))$isdir]

accuracy_lh <- matrix(ncol=3, nrow=length(sub_dir))
accuracy_rh <- accuracy_lh

set.seed(42)


sub_index <- 1
for (sub in sub_dir){

  message(sprintf("Subject %s\n", sub_index))

  #sub <- sub_dir[1]
  path_c <- paste(pwd, sub, sep='/')
  roi_list = list.files(path = path_c, pattern = "*02.txt")

  roi_index <- 1
  for (roi_sub in roi_list){

    #roi_sub <- roi_list[1]
    roi_c <- paste(path_c, roi_sub, sep='/')
    roi_data <- read.table(roi_c,sep=",",header=T)

    # 1 = Can/NonCan; 0/1 = can/non-can; 0/1 = side view, tilted/depth view, rotated; 0/1 = metal/non-metal
    # 2 = global shape; 2/8 = scrambling; 0 = filler; 0 = filler

    ## Data by experiment
    # Objects can-noncan
    exp_can <- roi_data[roi_data[,1]==1,]
    exp_can <- select(exp_can, -betas1, -betas3, -betas4)

    data_can <- exp_can[exp_can[,1]==0,]
    data_noncan <- exp_can[exp_can[,1]==1,]

    labels_can <- select(data_can, betas2)
    #labels_can[labels_can=="0"] <- "FALSE"
    labels_can[labels_can=="0"] <- "TRUE"
    labels_can <- as.character(labels_can$betas2)

    data_can <- select(data_can, -betas2)

    labels_noncan <- select(data_noncan, betas2)
    labels_noncan[labels_noncan=="1"] <- "TRUE"
    labels_noncan <- as.character(labels_noncan$betas2)

    data_noncan <- select(data_noncan, -betas2)

    # Global shapes
    exp_global <- roi_data[roi_data[,1]==2,]
    exp_global <- select(exp_global, -betas1, -betas3, -betas4)

    exp_global <- exp_global[exp_global[,1]==2,]
    exp_global$betas2 <- 1

    labels_global <- select(exp_global, betas2)
    labels_global[labels_global=="1"] <- "TRUE"
    labels_global <- as.character(labels_global$betas2)

    data_global <- select(exp_global, -betas2)

    # Split global into training and test data
    global_split <- sample.split(data_global, SplitRatio = 0.80)

    labels_global_train <- labels_global[global_split]
    data_global_train <- data_global[global_split,]

    labels_global_test <- labels_global[-global_split]
    data_global_test <- data_global[-global_split,]

    rm(exp_can, exp_global, roi_data)

    # Normalize features
    #data_can <- t(apply(data_can, 1, norm_data))
    #data_noncan <- t(apply(data_noncan, 1, norm_data))
    #data_global <- t(apply(data_global, 1, norm_data))


    ## SVM
    #svm.global <- svm(data_global,y=NULL, type='one-classification', nu = 0.10, scale=TRUE, kernel="radial", cross = 10)
    svm.global <- svm(data_global_train, y=NULL, type='one-classification', nu = 0.05, scale = TRUE, kernel="radial", cross = 10)

    svm.predglobal <- predict(svm.global, data_global_test)
    svm.prednoncan <- predict(svm.global, data_noncan)
    svm.predcan <- predict(svm.global, data_can)

    conf_global <- table(Predicted = svm.predglobal, Reference = labels_global_test)
    conf_noncan <- table(Predicted = svm.prednoncan, Reference = labels_noncan)
    conf_can <- table(Predicted = svm.predcan, Reference = labels_can)

    accuracy_global <- conf_global[2]/(conf_global[1] + conf_global[2])
    accuracy_noncan <- conf_noncan[2]/(conf_noncan[1] + conf_noncan[2])
    accuracy_can <- conf_can[2]/(conf_can[1] + conf_can[2])


    # Save data
    if (roi_index == 1){
      accuracy_lh[sub_index,] <- c(accuracy_global, accuracy_noncan, accuracy_can)
    } else {
      accuracy_rh[sub_index,] <- c(accuracy_global, accuracy_noncan, accuracy_can)
    }

    roi_index <- roi_index + 1
  }

  sub_index <- sub_index + 1

}


## Prepare results
col_names_tpj <- c("ACC_Global", "ACC_NonCan", "ACC_Can")

accuracy_lh  = data.frame(accuracy_lh )
colnames(accuracy_lh) <- col_names_tpj

accuracy_rh  = data.frame(accuracy_rh )
colnames(accuracy_rh) <- col_names_tpj

write.table(accuracy_lh, file = "SVM_OneClass_TPJ_Global_Obj_accuracy_lh.txt", sep = ",", row.names = FALSE, col.names = TRUE)
write.table(accuracy_rh, file = "SVM_OneClass_TPJ_Global_Obj_accuracy_rh.txt", sep = ",", row.names = FALSE, col.names = TRUE)