adding the task

code/stats/mfe_d_face_measure_computations_v2_mini.R

@@ -0,0 +1,146 @@
+# This script will load, and organize the pavlovia data. Then, computes measures of interest.
+# For each participant, a single, new, organized csv file that has all the necessary information will be generated.
+# Author: Kianoosh Hosseini at NDCLab @FIU (https://Kianoosh.info; https://NDClab.com)
+# Last Update: 2024-10-17 (YYYY-MM-DD)
+# This version computes hit rates by binning the face delay.
+### This version just computes number of errors for all participants without any exclusion.
+
+library(tidyverse)
+library(dplyr)
+library(stringr)
+library(psycho) # to compute d' measures, etc.
+
+#Working directory should be the Psychopy experiment directory.
+proje_wd <- "/Users/kihossei/Library/CloudStorage/[email protected]/My Drive/My Digital Life/Professional/Github_Repos/mfe-d-dataset"
+setwd(proje_wd)
+
+
+input_raw_path <- paste(proje_wd, "sourcedata", "checked", "psychopy", sep ="/", collapse = NULL) # input data directory
+input_organized_path <- paste(proje_wd, "derivatives", "psychopy", "csv_output", sep ="/", collapse = NULL) # input directory for data files generated by the organizer script!
+output_path <- paste(proje_wd, "derivatives", "psychopy", "stat_output", sep ="/", collapse = NULL) # output directory
+proc_fileName <- "processed_data_mfe_d_Proj_v3.csv" # output filename
+flanker_csv_fileName <- "_mfe_d_flankerDat_v1.csv" # each output csv file will have this on its filename
+surprise_csv_fileName <- "_mfe_d_surpriseDat_v1.csv" # each output csv file will have this on its filename
+
+
+## creating a list of all raw data csv files in the input folder.
+raw_datafiles_list <- c() # an empty list that will be filled in the next "for" loop!
+csvSelect <- list.files(input_raw_path, pattern = ".csv") # listing only csv files
+for (i in 1:length(csvSelect)){
+  temp_for_file <- ifelse (str_detect(csvSelect[i], "mfe_d", negate = FALSE), 1, 0)
+  if (temp_for_file == 1){
+    temp_list <- csvSelect[i]
+    raw_datafiles_list <- c(raw_datafiles_list, temp_list)
+  }
+}
+# Creating the main empty dataframe that will be filled with the data from the loop below:
+main_df <- setNames(data.frame(matrix(ncol = 25, nrow = 0)), c("participant_id", "congAcc", "incongAcc",
+                                                               "incongruent_dat_meanRT", "errorDat_meanRT", "congruent_dat_meanRT", "corrDat_meanRT",
+                                                               "congCorr_meanRT", "incongCorr_meanRT", "congCorr_logMeanRT",
+                                                               "congErr_meanRT", "incongErr_meanRT", "congErr_logMeanRT", "incongErr_logMeanRT",
+                                                               "incongCorr_logMeanRT", "flankEff_meanACC", "flankEff_meanRT", "flankEff_logMeanRT",
+                                                               "reported_errors", "committed_errors", "memoryBias_score", "num_incong_errorDat",
+                                                               "faceDuration_avg", "faceDuration_min", "faceDuration_max"))
+
+
+# Looping over all participants
+for (subject in 1:length(raw_datafiles_list)){
+  #for this participant, find the raw csv file
+  psychopy_file <- paste(input_raw_path,raw_datafiles_list[subject], sep = "/", collapse = NULL)
+
+  #read in the data for this participant, establish id, and remove extraneous variables
+  psychopyDat <- read.csv(file = psychopy_file, stringsAsFactors = FALSE, na.strings=c("", "NA"))
+  participant_id <- psychopyDat$id[1]
+
+  # Load this participant's flanker and surprise data frames
+  flanker_name <- paste0(participant_id, flanker_csv_fileName, sep = "", collapse = NULL)
+  surprise_name <- paste0(participant_id, surprise_csv_fileName, sep = "", collapse = NULL)
+  flanker_df <- read.csv(file = paste(input_organized_path, flanker_name, sep = "/", collapse = NULL), stringsAsFactors = FALSE, na.strings=c("", "NA"))
+  surprise_df <- read.csv(file = paste(input_organized_path, surprise_name, sep = "/", collapse = NULL), stringsAsFactors = FALSE, na.strings=c("", "NA"))
+
+
+  incong_flankerDat <- filter(flanker_df, current_trial_congruency == 0)
+  cong_flankerDat <- filter(flanker_df, current_trial_congruency == 1)
+
+  incongAcc <- mean(as.numeric(incong_flankerDat$current_trial_accuracy))
+  congAcc <- mean(as.numeric(cong_flankerDat$current_trial_accuracy))
+
+  # subset the data for correct and error trials, separately for congruent and incongruent trials, creating new data frames for each
+  corrDat <- flanker_df[flanker_df$current_trial_accuracy == 1,]
+  corrDat_meanRT <- mean(corrDat$current_trial_rt, na.rm = TRUE)
+
+  congruent_dat <- flanker_df[flanker_df$current_trial_congruency == 1,]
+  congruent_dat_meanRT <- mean(congruent_dat$current_trial_rt, na.rm = TRUE)
+
+  cong_corrDat <- corrDat[corrDat$current_trial_congruency == 1,]
+  incong_corrDat <- corrDat[corrDat$current_trial_congruency == 0,]
+
+  errorDat <- flanker_df[flanker_df$current_trial_accuracy == 0,]
+  errorDat_meanRT <- mean(errorDat$current_trial_rt, na.rm = TRUE)
+
+  incongruent_dat <- flanker_df[flanker_df$current_trial_congruency == 0,]
+  incongruent_dat_meanRT <- mean(incongruent_dat$current_trial_rt, na.rm = TRUE)
+
+
+  cong_errorDat <- errorDat[errorDat$current_trial_congruency == 1,]
+  incong_errorDat <- errorDat[errorDat$current_trial_congruency == 0,]
+  #for correct trials, compute mean RT (raw and log-corrected)
+  congCorr_meanRT <- mean(cong_corrDat$current_trial_rt, na.rm = TRUE)
+  incongCorr_meanRT <- mean(incong_corrDat$current_trial_rt, na.rm = TRUE)
+
+  congErr_meanRT <- mean(cong_errorDat$current_trial_rt, na.rm = TRUE)
+  incongErr_meanRT <- mean(incong_errorDat$current_trial_rt, na.rm = TRUE)
+
+  congCorr_logMeanRT <- mean(log((1+cong_corrDat$current_trial_rt)), na.rm = TRUE)
+  incongCorr_logMeanRT <- mean(log((1+incong_corrDat$current_trial_rt)), na.rm = TRUE)
+
+  congErr_logMeanRT <- mean(log((1+cong_errorDat$current_trial_rt)), na.rm = TRUE)
+  incongErr_logMeanRT <- mean(log((1+incong_errorDat$current_trial_rt)), na.rm = TRUE)
+
+  # compute flanker-effect scores for accuracy, RT, log-RT
+  flankEff_meanACC <- incongAcc - congAcc
+  flankEff_meanRT <- incongCorr_meanRT - congCorr_meanRT
+  flankEff_logMeanRT <- incongCorr_logMeanRT - congCorr_logMeanRT
+
+  #
+  # number of committed errors in the flanker task
+  committed_errors <- nrow(errorDat)
+  # number of reported errors
+  psychopyDatTrim <- psychopyDat[("errorNum_text_box.text")] # stores the number of reported errors by subjects
+  reported_errors <- subset(psychopyDatTrim, complete.cases(psychopyDatTrim$errorNum_text_box.text))
+  reported_errors <- reported_errors$errorNum_text_box.text # number of reported errors by participants
+  reported_errors <- str_extract_all(reported_errors, '\\d+\\.?\\d*') # to extract all sequences of digits from the input string
+  # There was a participant who had reported 70/100. values 70 and 100. To solve this issue, I replace this kind of values with NAs!
+  if (length(reported_errors) == 0){ # there is no reported errors
+    reported_errors <- NA
+    memoryBias_score <- NA
+  } else {
+    reported_errors <- parse_number(reported_errors[[1]]) # in cases like 70/100, we will have 70 100 at this stage. So, length(reported_errors) will
+    # higher than 1.
+    if (length(reported_errors) > 1){ # In case they have reported sth like 70/100
+      reported_errors <- NA
+      memoryBias_score <- NA
+    } else if (length(reported_errors) == 1){
+      reported_errors <- reported_errors[1]
+      memoryBias_score <- ((reported_errors - committed_errors)/ reported_errors) # percent bias score calculation
+    }
+  }
+
+
+  num_incong_errorDat <- nrow(incong_errorDat)
+  faceDuration_avg <- mean(flanker_df$current_trial_faceDuration)
+  faceDuration_min <- min(flanker_df$current_trial_faceDuration)
+  faceDuration_max <- max(flanker_df$current_trial_faceDuration)
+            #### filling the main data frame
+            main_df[nrow(main_df) + 1,] <-c(participant_id, congAcc, incongAcc,
+                                            incongruent_dat_meanRT, errorDat_meanRT, congruent_dat_meanRT, corrDat_meanRT,
+                                            congCorr_meanRT, incongCorr_meanRT, congCorr_logMeanRT,
+                                            congErr_meanRT, incongErr_meanRT, congErr_logMeanRT, incongErr_logMeanRT,
+                                            incongCorr_logMeanRT, flankEff_meanACC, flankEff_meanRT, flankEff_logMeanRT,
+                                            reported_errors, committed_errors, memoryBias_score, num_incong_errorDat,
+                                            faceDuration_avg, faceDuration_min, faceDuration_max)
+} # Closing the loop for each participant
+
+
+
+