MLtools/vectorizeFC.R at main · CogBrainHealthLab/MLtools · 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
########################################################################################################
########################################################################################################

#' @title vectorizeFC
#'
#' @description vectorizing and concatenating and M x M matrices
#'
#' @details This function reads the M x M matrices (saved as .csv files) in a directory, vectorizes and concatenates them into a single N x ((M x M)-M)/2 matrix. N= number of subjects; M=number of nodes in the connectome
#'
#' @param FC_dir the directory containing the M x M matrices (saved as .csv files)
#' @param filename the output format of the concatenated group level matrix. Only filenames with the *.csv or *.rds extensions are accepted. The .rds format uses less disk space. Set to 'FCmat.rds' by default
#' @param task A string to select the .csv files with a specific taskname.
#' @param fisherz `TRUE` or `FALSE` option specifying if the edge values should undergo a fisher-Z transformation. This would be applicable in cases and functional connectivity is calculated as a pearson's correlation coefficient. Set to `TRUE` by default.
#' @param timeseries `TRUE` or `FALSE` option specifying if the parcellated timeseries are used as inputs. Set to `FALSE` by default.
#' @returns Depending on whether a filename with a .csv or .rds file extension is specified, outputs a .csv file containing the  N x ((M x M)-M)/2 matrix or a .rds file containing a list object: 1)N x ((M x M)-M)/2 matrix, 2) a vector of subject IDs.
#'
#' @examples
#' \dontrun{
#' vectorizeFC(FC_dir="FCmat", filename="FCmat.rds", fisherz=TRUE)
#' }
#' @importFrom stringr str_detect
#' @importFrom psych fisherz
#' @export
#'
#'
########################################################################################################
########################################################################################################
vectorizeFC=function(FC_dir, filename="FCmat.rds", task, fisherz=TRUE, timeseries=FALSE)
{
  ##check filename extension
  if(stringr::str_detect(string=filename, pattern=".csv")==F & stringr::str_detect(string=filename, pattern=".rds")==F)
  {
    stop("invalid filename extension")
  }
  if(missing("task"))  {subjlist=list.files(path=FC_dir)}
  else {subjlist=list.files(path=FC_dir,pattern=task)}

  if(timeseries==F)
  {
    FCformat=read.csv(paste(FC_dir,"/",subjlist[1], sep=""), header=F)

    #if subjects' FCmat is already vectorized
    if(dim(FCformat)[1]==1)
    {
      FCmat=matrix(NA,nrow=NROW(subjlist), ncol=dim(FCformat)[2])
      for(sub in 1:NROW(subjlist))
      {
        FCmat[sub,]=unlist(read.table(paste(FC_dir,"/",subjlist[sub], sep=""),header = F,sep=",",col.names = F))
        cat(paste(subjlist[sub],"\n"))
      }
      if(fisherz==T)
      {
        FCmat=psych::fisherz(FCmat)
      }
    }

    #if subjects' FCmat has not been vectorized
    if(dim(FCformat)[1]>1)
    {
      x=dim(FCformat)[1]
      FCmat=matrix(NA,nrow=NROW(subjlist), ncol=((x*x-x)/2))
      for(sub in 1:NROW(subjlist))
      {
        FCmat.temp=read.table(paste(FC_dir,"/",subjlist[sub], sep=""),header = F,sep=",")
        FCmat[sub,]=FCmat.temp[upper.tri(FCmat.temp, diag = FALSE)]
        cat(paste(subjlist[sub],"\n"))
        remove(FCmat.temp)
      }
      if(fisherz==T)
      {
        FCmat=psych::fisherz(FCmat)
      }
    }
  } else if(timeseries==T)
  {
    FCformat=read.csv(paste(FC_dir,"/",subjlist[1], sep=""), header=F)
    x=NCOL(FCformat)
    FCmat=matrix(NA,nrow=NROW(subjlist), ncol=((x*x-x)/2))
    for(sub in 1:NROW(subjlist))
    {
      FCmat.temp=read.table(paste(FC_dir,"/",subjlist[sub], sep=""),header = F,sep=",")
      FCmat[sub,]=FCmat.temp[upper.tri(cor(FCmat.temp), diag = FALSE)]
      cat(paste(subjlist[sub],"\n"))
      remove(FCmat.temp)
    }
    if(fisherz==T)
    {
      FCmat=psych::fisherz(FCmat)
    }
  }

  ##output file
  if(stringr::str_detect(string=filename, pattern=".csv"))
  {
    write.table(FCmat,file = filename,row.names = F,col.names = F, sep=",")
    cat(paste("Concatenated FC matrices saved to",filename,"\n",sep=" "))
  } else if(stringr::str_detect(string=filename, pattern=".rds"))
  {
    saveRDS(list(FCmat,gsub(".csv","",subjlist)),file = filename)
    cat(paste("Concatenated FC matrices saved to",filename,"\n",sep=" "))
  }
}

########################################################################################################
########################################################################################################