R/cleaning_functions.R
In bramzandbelt/irmass: Independent Race Model Analysis of Selective Stopping
# Cleaning functions
# ==============================================================================
#' Identify analysis-level outliers
#'
#' @param df data frame with summary stats for each subject
#' @param group_vars variables for grouping
#' @param dep_var dependent variables
#' @param z_threshold outlier criterion z score (defaults to 2.5)
#' @export
identify_outlying_values <- function(df, group_vars, dep_var, z_threshold = 2.5) {
zscore <- function(x) (x - mean(x))/sd(x)
expr_z <- lazyeval::interp(~zscore(x), x = as.name(dep_var))
df %>%
dplyr::group_by_(.dots = group_vars) %>%
dplyr::mutate_(.dots = setNames(list(expr_z), "z_score")) %>%
dplyr::mutate(is_outlier = abs(z_score) > z_threshold) %>%
dplyr::ungroup()
}
# ==============================================================================
#' Identify analysis-level outliers
#'
#' @param df data frame with summary stats for each subject
#' @param analysis_id analysis identifier (e.g. 'a5'), corresponds to analysis ID in preregistration document
#' @param trial_alt_levels trial_alt levels to include
#' @export
identify_outliers_analysis_level <- function(df, analysis_id, trial_alt_levels) {
# Identify outlying values
# if (analysis_id %in% c('a5','a6') {
# df_outlying_vals_identified <-
# df %>%
# # Compute outlying values separately for each trial and delay type
# dplyr::group_by(trial_alt, t_d_alt) %>%
# dplyr::mutate(zscore = (mean_RT - mean(mean_RT)) / sd(mean_RT),
# is_outlier = abs(zscore) > 2.5
# )
# }
if (analysis_id == 'a5') {
dplyr::select(subjectIx, trial_alt, mean_RT)
} else if (analysis_id == 'a6') {
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
tidyr::spread(key = c(t_d_alt), value = mean_RT)
bla <-
df_outlying_vals_identified %>%
dplyr::filter(trial_alt %in% trial_alt_levels) %>%
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
# Identify outlying subjects
outlying_subjects <-
df_outlying_vals_identified
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
tidyr::spread(key = c(t_d_alt), value = mean_RT)
df_outlying_subs_removed
}
}
# ==============================================================================
#' Tidy block data
#'
#' Tidying involves multiple steps, including aggregating no-signal performance trial performance across go-signal levels, renaming variables, and assessing performance criteria on the basis of aggregate measures.
#' @export
tidy_block_data <- function(df, stage = 'expt') {
if (stage == 'prac') {
clean_df <-
df %>%
dplyr::mutate(block_type = ifelse(stringr::str_detect(blockId, '^e.*'),
"mixed",
ifelse(.$blockIx == 0,
"NS",
ifelse(((.$blockIx == 1) & (.$taskVersionId == "A")) |
((.$blockIx == 2) & (.$taskVersionId == "B")),
"AS",
ifelse(((.$blockIx == 2) & (.$taskVersionId == "A")) |
((.$blockIx == 1) & (.$taskVersionId == "B")),
"SS",
ifelse(.$blockIx == 3,
"mixed",
NA_character_
)
)
)
)
)
) %>%
dplyr::mutate(NS_accuracy = pmin(s1Acc_00,s1Acc_01),
NS_mean_RT = pmax(s1MeanRt_00,s1MeanRt_01),
NS_mean_RTdiff = pmax(s1MeanRtDiff_00,s1MeanRtDiff_01)
) %>%
dplyr::rename(SL_accuracy = s2Acc_00,
SR_accuracy = s2Acc_01,
SB_accuracy = s2Acc_02,
IG_accuracy = s2Acc_03
) %>%
dplyr::mutate(NS_accuracy_failed = NS_accuracy < 85,
NS_mean_RT_failed = NS_mean_RT > 650,
NS_mean_RTdiff_failed = NS_mean_RTdiff > 50,
SL_accuracy_failed = SL_accuracy < 20,
SR_accuracy_failed = SR_accuracy < 20,
SB_accuracy_failed = SB_accuracy < 20,
IG_accuracy_failed = IG_accuracy < 80
) %>%
dplyr::select(subjectIx, blockId, block_type,
NS_accuracy, NS_mean_RT, NS_mean_RTdiff,
SL_accuracy, SR_accuracy, SB_accuracy, IG_accuracy,
NS_accuracy_failed, NS_mean_RT_failed, NS_mean_RTdiff_failed,
SL_accuracy_failed, SR_accuracy_failed, SB_accuracy_failed, IG_accuracy_failed
) %>%
dplyr::group_by(subjectIx,blockId) %>%
dplyr::mutate(attempt = 1:dplyr::n()) %>%
dplyr::ungroup()
} else if (stage == 'expt') {
clean_df <-
df %>%
dplyr::group_by(subjectIx, blockId) %>%
dplyr::summarize(NS_accuracy = mean(c(s1Acc_00,s1Acc_01)),
NS_accuracy_failed = NS_accuracy < 85,
NS_mean_RT = mean(c(s1MeanRt_00,s1MeanRt_01)),
NS_mean_RT_failed = NS_mean_RT > 650,
NS_mean_RTdiff = mean(c(s1MeanRtDiff_00,s1MeanRtDiff_01)),
NS_mean_RTdiff_failed = NS_mean_RTdiff > 50,
SL_accuracy = mean(s2Acc_00),
SL_accuracy_failed = SL_accuracy < 20,
SR_accuracy = mean(s2Acc_01),
SR_accuracy_failed = SR_accuracy < 20,
SB_accuracy = mean(s2Acc_02),
SB_accuracy_failed = SB_accuracy < 20,
IG_accuracy = mean(s2Acc_03),
IG_accuracy_failed = IG_accuracy < 80)
}
# Make tidy ------------------------------------------------------------------
if (stage == 'prac') {
general_cols <- c("subjectIx","blockId","block_type","attempt")
performance_cols <- c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")
failure_cols <- colnames(clean_df) %>% .[stringr::str_detect(., ".*failed$")]
# Performance stats
tidy_performance_stats <-
clean_df %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "performance") %>%
dplyr::select_(.dots = c(general_cols,"criterion","performance"))
# Failure stats
tidy_failure_stats <-
clean_df %>%
dplyr::select_(.dots = c(general_cols,failure_cols)) %>%
setNames(gsub("_failed","",names(.))) %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "failed") %>%
dplyr::select_(.dots = c(general_cols,"criterion","failed"))
# Combine performance and failure stats in one data frame
tidy_data <-
dplyr::left_join(tidy_performance_stats,tidy_failure_stats,
by = c(general_cols,"criterion")) %>%
dplyr::arrange(subjectIx, blockId, attempt, criterion)
} else if (stage == 'expt') {
general_cols <- c("subjectIx","blockId")
performance_cols <- c("NS_accuracy","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy","NS_mean_RT", "NS_mean_RTdiff")
failure_cols <- colnames(clean_df) %>% .[stringr::str_detect(., ".*failed$")]
# Performance stats
tidy_performance_stats <-
clean_df %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "performance") %>%
dplyr::select_(.dots = c(general_cols,"criterion","performance"))
# Failure stats
tidy_failure_stats <-
clean_df %>%
dplyr::select_(.dots = c(general_cols,failure_cols)) %>%
setNames(gsub("_failed","",names(.))) %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "failed") %>%
dplyr::select_(.dots = c(general_cols,"criterion","failed"))
# Combine performance and failure stats in one data frame
tidy_data <-
dplyr::left_join(tidy_performance_stats,tidy_failure_stats,
by = c(general_cols,"criterion")) %>%
dplyr::arrange(subjectIx, criterion, blockId)
}
}
# ==============================================================================
#' Tidy trial data
#'
#' Tidying involves the following steps
#' - renaming and computing variables in line with preregistration
#'
#' For confirmatory analyses, only the necessary variables are returned. For exploratory analysis, all variables are returned.
#' @param df Data frame with minimally processed trial data
#' @param analysis_type Analysis type: 'confirmatory' or 'exploratory'
#' @export
tidy_trial_data <- function(df, analysis_type = 'confirmatory') {
# Manipulated variables (see §2.5.1. of the preregistration) =================
# s1 - go stimululs
# trial - trial type
# t_d - stop-signal delay
# t_d_alt - alternative stop-signal delay category
clean_df <-
df %>%
dplyr::mutate(s1 = ifelse(s1Ix == 0,
"M",
ifelse(s1Ix == 1,
"I",
NA_integer_)
),
trial = ifelse(is.na(s2Ix),
"NS",
ifelse(s2Ix == 0,
"SL",
ifelse(s2Ix == 1,
"SR",
ifelse(s2Ix == 2,
"SB",
ifelse(s2Ix == 3,
"IG",
NA_integer_)
)
)
)
),
trial_alt = forcats::fct_collapse(trial,
SAS = c("SL","SR"),
SSS = "SB",
NS = "NS"),
t_d = round((s2Ons - s2OnsDt) - (s1Ons - s1OnsDt), digits = 3),
t_d_alt = ifelse(soaIx %in% 0:2,
"short",
ifelse(soaIx == 3,
"intermediate",
ifelse(soaIx == 4,
"long",
NA_character_
)
)
)
) %>%
# Measured variables (see § 2.5.2. of the pre-registration) ==================
# Finger-level response count (RC_{finger})
# Finger-level response time (RT_{finger})
dplyr::rename(RC_LM = keyCount_f,
RC_RM = keyCount_b,
RC_LI = keyCount_e,
RC_RI = keyCount_a,
) %>%
dplyr::rename(RT_LM = rt1_f,
RT_RM = rt1_b,
RT_LI = rt1_e,
RT_RI = rt1_a
) %>%
# Indices (see § 2.5.3. of the pre-registration) ===========================
# Response category ----------------------------------------------------------
# - RB: Bimanual response, fingers compatible with the go-stimulus
# - RL: Left-hand response, finger compatible with the go-stimulus
# - RR: Right-hand response, finger compatible with the go-stimulus
# - RBO: Bimanual response, fingers compatible with the other go-stimulus
# - RLO: Left-hand response, finger compatible with the other go-stimulus
# - RRO: Right-hand response, finger compatible with the other go-stimulus
# - NR: No response
# - NOC: Not otherwise classified
dplyr::mutate(r = dplyr::case_when(
.$s1 == "M" & .$RC_LM == 1 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RB",
.$s1 == "M" & .$RC_LM == 1 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RL",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RR",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 1 ~ "RBO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 0 ~ "RLO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 1 ~ "RRO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "NR",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 1 ~ "RB",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 0 ~ "RL",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 1 ~ "RR",
.$s1 == "I" & .$RC_LM == 1 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RBO",
.$s1 == "I" & .$RC_LM == 1 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RLO",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RRO",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "NR"
)
) %>%
dplyr::mutate(r = ifelse(is.na(r),
"NOC",
r
)
) %>%
# Bimanual response recorded (r_{bi})
dplyr::mutate(r_bi = ifelse(r %in% c("RB","RBO"),
TRUE,
FALSE
)
) %>%
# Trial-level response time (RT_{trial}); also include RT difference for reporting of descriptive statistics
dplyr::mutate(RT_trial = rt1_mean,
RT_diff_trial = rtDiff1_mean) %>%
# Trial label - performance ($trial_{performance}$)
dplyr::mutate(trial_performance = dplyr::case_when(
.$trial == "NS" & .$r == "RB" ~ "NS_correct",
.$trial == "NS" & .$r == "RBO" ~ "NS_commission-error-bi",
.$trial == "NS" & .$r %in% c("RL","RR","RLO","RRO") ~ "NS_commission-error-uni",
.$trial == "NS" & .$r == "NOC" ~ "NS_commission-error-other",
.$trial == "NS" & .$r == "NR" ~ "NS_omission-error",
.$trial == "SL" & .$r %in% c("RB","RBO") ~ "SL_stop-respond-bi",
.$trial == "SL" & .$r %in% c("RL","RLO","RRO") ~ "SL_stop-respond-uni",
.$trial == "SL" & .$r %in% c("NR","NOC") ~ "SL_stop-respond-other",
.$trial == "SL" & .$r == "RR" ~ "SL_stop-inhibit",
.$trial == "SR" & .$r %in% c("RB","RBO") ~ "SR_stop-respond-bi",
.$trial == "SR" & .$r %in% c("RR","RLO","RRO") ~ "SR_stop-respond-uni",
.$trial == "SR" & .$r %in% c("NR","NOC") ~ "SR_stop-respond-other",
.$trial == "SR" & .$r == "RL" ~ "SR_stop-inhibit",
.$trial == "SB" & .$r %in% c("RB","RBO") ~ "SB_stop-respond-bi",
.$trial == "SB" & .$r %in% c("RL","RR","RLO","RRO") ~ "SB_stop-respond-uni",
.$trial == "SB" & .$r == "NOC" ~ "SB_stop-respond-other",
.$trial == "SB" & .$r == "NR" ~ "SB_stop-inhibit",
.$trial == "IG" & .$r == "RB" ~ "IG_correct",
.$trial == "IG" & .$r == "RBO" ~ "IG_commission-error-bi",
.$trial == "IG" & .$r %in% c("RL","RR","RLO","RRO") ~ "IG_commission-error-uni",
.$trial == "IG" & .$r == "NOC" ~ "IG_commission-error-other",
.$trial == "IG" & .$r == "NR" ~ "IG_omission-error"
)
)
# Select relevant variables (for confirmatory analysis) or return everything (for exploratory analysis)
if (analysis_type == 'confirmatory') {
RT_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, t_d, t_d_alt, RT_trial, RT_diff_trial)
resp_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, t_d, t_d_alt, r_bi, trialCorrect)
} else if (analysis_type == 'exploratory') {
RT_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, responseType, t_d, t_d_alt, RT_trial, RT_diff_trial, trialCorrect, trial_performance)
resp_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, responseType, t_d, t_d_alt, r_bi, trialCorrect, trial_performance)
}
list(RT_df = RT_df, resp_df = resp_df)
}
# ==============================================================================
#' Get column types for extracting data from csv files
#'
#' File type can be any of the following:
#'
#' expt_block_data - tidied block-level data from experimental session
#' expt_trial_resp_data - tidied trial-level response data from experimental session
#' expt_trial_rt_data - tidied trial-level resonse time data from experimental session
#' log_sess_cols - source data, session-related columns
#' log_block_cols - source data, block-related columns
#' log_trial_cols - source data, trial-related columns
#' log_trial_cols_sub00 - source data, trial-related columns for subject 00
#' prac_block_data - tidied block-level data from practice session
#'
#' @param file_type Type of file to be read
#' @export
get_col_types <- function(file_type){
switch(tolower(file_type),
expt_block_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
criterion = readr::col_factor(levels = c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")),
performance = readr::col_double(),
failed = readr::col_logical()
),
expt_trial_resp_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
trialCorrect = readr::col_logical()
),
expt_trial_resp_data_exploratory =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
trialCorrect = readr::col_logical(),
trial_performance = readr::col_character()
),
expt_trial_rt_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
RT_trial = readr::col_double()
),
expt_trial_rt_data_exploratory =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
RT_trial = readr::col_double(),
trialCorrect = readr::col_logical(),
trial_performance = readr::col_character()
),
log_sess_cols =
readr::cols_only(subjectIx = readr::col_integer(),
sessDate = readr::col_date(format = "%Y-%m-%d"),
studyId = readr::col_character(),
experimenterId = readr::col_character(),
taskVersionId = readr::col_character(),
responseDevice = readr::col_character(),
blockId = readr::col_character()
),
log_block_cols =
readr::cols_only(subjectIx = readr::col_integer(),
taskVersionId = readr::col_character(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
s1Acc_00 = readr::col_double(),
s1AccCritMet_00 = readr::col_logical(),
s1Acc_01 = readr::col_double(),
s1AccCritMet_01 = readr::col_logical(),
s2Acc_00 = readr::col_double(),
s2AccCritMet_00 = readr::col_logical(),
s2Acc_01 = readr::col_double(),
s2AccCritMet_01 = readr::col_logical(),
s2Acc_02 = readr::col_double(),
s2AccCritMet_02 = readr::col_logical(),
s2Acc_03 = readr::col_double(),
s2AccCritMet_03 = readr::col_logical(),
s1MeanRt_00 = readr::col_double(),
s1MeanRtCritMet_00 = readr::col_logical(),
s1MeanRt_01 = readr::col_double(),
s1MeanRtCritMet_01 = readr::col_logical(),
s1MeanRtDiff_00 = readr::col_double(),
s1MeanRtDiffCritMet_00 = readr::col_logical(),
s1MeanRtDiff_01 = readr::col_double(),
s1MeanRtDiffCritMet_01 = readr::col_logical()
),
log_trial_cols =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
fixIx = readr::col_integer(),
s1Ix = readr::col_integer(),
s2Ix = readr::col_double(),
soaIx = readr::col_double(),
tSession = readr::col_time(format = "%H:%M:%S"),
tBlock = readr::col_time(format = "%H:%M:%OS"),
trialOns = readr::col_double(),
trialDur = readr::col_double(),
fixOns = readr::col_double(),
fixOnsDt = readr::col_double(),
fixDur = readr::col_double(),
fixDurDt = readr::col_double(),
s1Ons = readr::col_double(),
s1OnsDt = readr::col_double(),
s1Dur = readr::col_double(),
s1DurDt = readr::col_double(),
s2Ons = readr::col_double(),
s2OnsDt = readr::col_double(),
s2Dur = readr::col_double(),
s2DurDt = readr::col_double(),
keyCount_f = readr::col_integer(),
keyCount_b = readr::col_integer(),
keyCount_e = readr::col_integer(),
keyCount_a = readr::col_integer(),
rt1_f = readr::col_double(),
rt1_b = readr::col_double(),
rt1_e = readr::col_double(),
rt1_a = readr::col_double(),
rt1_mean = readr::col_double(),
`rtDiff1_f-b` = readr::col_double(),
`rtDiff1_e-a` = readr::col_double(),
rtDiff1_mean = readr::col_double(),
trialCorrect = readr::col_logical(),
trialType = readr::col_character(),
responseType = readr::col_character(),
trialFeedback = readr::col_character()
),
log_trial_cols_sub00 =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
fixIx = readr::col_integer(),
s1Ix = readr::col_integer(),
s2Ix = readr::col_double(),
soaIx = readr::col_double(),
tSession = readr::col_time(format = "%H:%M:%S"),
tBlock = readr::col_time(format = "%H:%M:%OS"),
trialOns = readr::col_double(),
trialDur = readr::col_double(),
fixOns = readr::col_double(),
fixOnsDt = readr::col_double(),
fixDur = readr::col_double(),
fixDurDt = readr::col_double(),
s1Ons = readr::col_double(),
s1OnsDt = readr::col_double(),
s1Dur = readr::col_double(),
s1DurDt = readr::col_double(),
s2Ons = readr::col_double(),
s2OnsDt = readr::col_double(),
s2Dur = readr::col_double(),
s2DurDt = readr::col_double(),
keyCount_f = readr::col_integer(),
keyCount_h = readr::col_integer(),
keyCount_v = readr::col_integer(),
keyCount_b = readr::col_integer(),
rt1_f = readr::col_double(),
rt1_h = readr::col_double(),
rt1_v = readr::col_double(),
rt1_b = readr::col_double(),
rt1_mean = readr::col_double(),
`rtDiff1_f-h` = readr::col_double(),
`rtDiff1_v-b` = readr::col_double(),
rtDiff1_mean = readr::col_double(),
trialCorrect = readr::col_logical(),
trialType = readr::col_character(),
responseType = readr::col_character(),
trialFeedback = readr::col_character()
),
prac_block_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
block_type = readr::col_factor(levels = c("NS","AS","SS","mixed"),
ordered = TRUE),
attempt = readr::col_factor(levels = c(1, 2, 3, 4, 5),
ordered = TRUE),
criterion = readr::col_factor(levels = c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")),
performance = readr::col_double(),
failed = readr::col_logical()
),
sim_trial_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_double(),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
RT_trial = readr::col_double(),
SSRT_true = readr::col_double(),
trialCorrect = readr::col_logical(),
trial_performance = col_character()
)
)
}
#' Splits log files into a session table and a block/trial table
#'
split_log_in_multi_table <- function(df) {
session_data <-
df %>%
dplyr::select(studyId:sessTime, -rngSeed) %>%
dplyr::distinct()
}
# ==============================================================================
#' Verifies index columns
#'
#' Identifies index columns, replaces NaNs with NAs, and converts objects to integer
#' @export
verify_ix_cols <- function(df) {
col_vector <- stringr::str_subset(colnames(df),".*Ix$")
df[col_vector] <- df[col_vector] %>%
purrr::map(~ stringr::str_replace(.x, "nan", "NA")) %>%
purrr::map(~ as.integer(.x))
df
}
bramzandbelt/irmass documentation built on June 28, 2019, 9:07 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
# Cleaning functions
# ==============================================================================
#' Identify analysis-level outliers
#'
#' @param df data frame with summary stats for each subject
#' @param group_vars variables for grouping
#' @param dep_var dependent variables
#' @param z_threshold outlier criterion z score (defaults to 2.5)
#' @export
identify_outlying_values <- function(df, group_vars, dep_var, z_threshold = 2.5) {
zscore <- function(x) (x - mean(x))/sd(x)
expr_z <- lazyeval::interp(~zscore(x), x = as.name(dep_var))
df %>%
dplyr::group_by_(.dots = group_vars) %>%
dplyr::mutate_(.dots = setNames(list(expr_z), "z_score")) %>%
dplyr::mutate(is_outlier = abs(z_score) > z_threshold) %>%
dplyr::ungroup()
}
# ==============================================================================
#' Identify analysis-level outliers
#'
#' @param df data frame with summary stats for each subject
#' @param analysis_id analysis identifier (e.g. 'a5'), corresponds to analysis ID in preregistration document
#' @param trial_alt_levels trial_alt levels to include
#' @export
identify_outliers_analysis_level <- function(df, analysis_id, trial_alt_levels) {
# Identify outlying values
# if (analysis_id %in% c('a5','a6') {
# df_outlying_vals_identified <-
# df %>%
# # Compute outlying values separately for each trial and delay type
# dplyr::group_by(trial_alt, t_d_alt) %>%
# dplyr::mutate(zscore = (mean_RT - mean(mean_RT)) / sd(mean_RT),
# is_outlier = abs(zscore) > 2.5
# )
# }
if (analysis_id == 'a5') {
dplyr::select(subjectIx, trial_alt, mean_RT)
} else if (analysis_id == 'a6') {
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
tidyr::spread(key = c(t_d_alt), value = mean_RT)
bla <-
df_outlying_vals_identified %>%
dplyr::filter(trial_alt %in% trial_alt_levels) %>%
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
# Identify outlying subjects
outlying_subjects <-
df_outlying_vals_identified
dplyr::select(subjectIx, t_d_alt, mean_RT) %>%
tidyr::spread(key = c(t_d_alt), value = mean_RT)
df_outlying_subs_removed
}
}
# ==============================================================================
#' Tidy block data
#'
#' Tidying involves multiple steps, including aggregating no-signal performance trial performance across go-signal levels, renaming variables, and assessing performance criteria on the basis of aggregate measures.
#' @export
tidy_block_data <- function(df, stage = 'expt') {
if (stage == 'prac') {
clean_df <-
df %>%
dplyr::mutate(block_type = ifelse(stringr::str_detect(blockId, '^e.*'),
"mixed",
ifelse(.$blockIx == 0,
"NS",
ifelse(((.$blockIx == 1) & (.$taskVersionId == "A")) |
((.$blockIx == 2) & (.$taskVersionId == "B")),
"AS",
ifelse(((.$blockIx == 2) & (.$taskVersionId == "A")) |
((.$blockIx == 1) & (.$taskVersionId == "B")),
"SS",
ifelse(.$blockIx == 3,
"mixed",
NA_character_
)
)
)
)
)
) %>%
dplyr::mutate(NS_accuracy = pmin(s1Acc_00,s1Acc_01),
NS_mean_RT = pmax(s1MeanRt_00,s1MeanRt_01),
NS_mean_RTdiff = pmax(s1MeanRtDiff_00,s1MeanRtDiff_01)
) %>%
dplyr::rename(SL_accuracy = s2Acc_00,
SR_accuracy = s2Acc_01,
SB_accuracy = s2Acc_02,
IG_accuracy = s2Acc_03
) %>%
dplyr::mutate(NS_accuracy_failed = NS_accuracy < 85,
NS_mean_RT_failed = NS_mean_RT > 650,
NS_mean_RTdiff_failed = NS_mean_RTdiff > 50,
SL_accuracy_failed = SL_accuracy < 20,
SR_accuracy_failed = SR_accuracy < 20,
SB_accuracy_failed = SB_accuracy < 20,
IG_accuracy_failed = IG_accuracy < 80
) %>%
dplyr::select(subjectIx, blockId, block_type,
NS_accuracy, NS_mean_RT, NS_mean_RTdiff,
SL_accuracy, SR_accuracy, SB_accuracy, IG_accuracy,
NS_accuracy_failed, NS_mean_RT_failed, NS_mean_RTdiff_failed,
SL_accuracy_failed, SR_accuracy_failed, SB_accuracy_failed, IG_accuracy_failed
) %>%
dplyr::group_by(subjectIx,blockId) %>%
dplyr::mutate(attempt = 1:dplyr::n()) %>%
dplyr::ungroup()
} else if (stage == 'expt') {
clean_df <-
df %>%
dplyr::group_by(subjectIx, blockId) %>%
dplyr::summarize(NS_accuracy = mean(c(s1Acc_00,s1Acc_01)),
NS_accuracy_failed = NS_accuracy < 85,
NS_mean_RT = mean(c(s1MeanRt_00,s1MeanRt_01)),
NS_mean_RT_failed = NS_mean_RT > 650,
NS_mean_RTdiff = mean(c(s1MeanRtDiff_00,s1MeanRtDiff_01)),
NS_mean_RTdiff_failed = NS_mean_RTdiff > 50,
SL_accuracy = mean(s2Acc_00),
SL_accuracy_failed = SL_accuracy < 20,
SR_accuracy = mean(s2Acc_01),
SR_accuracy_failed = SR_accuracy < 20,
SB_accuracy = mean(s2Acc_02),
SB_accuracy_failed = SB_accuracy < 20,
IG_accuracy = mean(s2Acc_03),
IG_accuracy_failed = IG_accuracy < 80)
}
# Make tidy ------------------------------------------------------------------
if (stage == 'prac') {
general_cols <- c("subjectIx","blockId","block_type","attempt")
performance_cols <- c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")
failure_cols <- colnames(clean_df) %>% .[stringr::str_detect(., ".*failed$")]
# Performance stats
tidy_performance_stats <-
clean_df %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "performance") %>%
dplyr::select_(.dots = c(general_cols,"criterion","performance"))
# Failure stats
tidy_failure_stats <-
clean_df %>%
dplyr::select_(.dots = c(general_cols,failure_cols)) %>%
setNames(gsub("_failed","",names(.))) %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "failed") %>%
dplyr::select_(.dots = c(general_cols,"criterion","failed"))
# Combine performance and failure stats in one data frame
tidy_data <-
dplyr::left_join(tidy_performance_stats,tidy_failure_stats,
by = c(general_cols,"criterion")) %>%
dplyr::arrange(subjectIx, blockId, attempt, criterion)
} else if (stage == 'expt') {
general_cols <- c("subjectIx","blockId")
performance_cols <- c("NS_accuracy","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy","NS_mean_RT", "NS_mean_RTdiff")
failure_cols <- colnames(clean_df) %>% .[stringr::str_detect(., ".*failed$")]
# Performance stats
tidy_performance_stats <-
clean_df %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "performance") %>%
dplyr::select_(.dots = c(general_cols,"criterion","performance"))
# Failure stats
tidy_failure_stats <-
clean_df %>%
dplyr::select_(.dots = c(general_cols,failure_cols)) %>%
setNames(gsub("_failed","",names(.))) %>%
tidyr::gather_(gather_cols = performance_cols,
key_col = "criterion",
value_col = "failed") %>%
dplyr::select_(.dots = c(general_cols,"criterion","failed"))
# Combine performance and failure stats in one data frame
tidy_data <-
dplyr::left_join(tidy_performance_stats,tidy_failure_stats,
by = c(general_cols,"criterion")) %>%
dplyr::arrange(subjectIx, criterion, blockId)
}
}
# ==============================================================================
#' Tidy trial data
#'
#' Tidying involves the following steps
#' - renaming and computing variables in line with preregistration
#'
#' For confirmatory analyses, only the necessary variables are returned. For exploratory analysis, all variables are returned.
#' @param df Data frame with minimally processed trial data
#' @param analysis_type Analysis type: 'confirmatory' or 'exploratory'
#' @export
tidy_trial_data <- function(df, analysis_type = 'confirmatory') {
# Manipulated variables (see §2.5.1. of the preregistration) =================
# s1 - go stimululs
# trial - trial type
# t_d - stop-signal delay
# t_d_alt - alternative stop-signal delay category
clean_df <-
df %>%
dplyr::mutate(s1 = ifelse(s1Ix == 0,
"M",
ifelse(s1Ix == 1,
"I",
NA_integer_)
),
trial = ifelse(is.na(s2Ix),
"NS",
ifelse(s2Ix == 0,
"SL",
ifelse(s2Ix == 1,
"SR",
ifelse(s2Ix == 2,
"SB",
ifelse(s2Ix == 3,
"IG",
NA_integer_)
)
)
)
),
trial_alt = forcats::fct_collapse(trial,
SAS = c("SL","SR"),
SSS = "SB",
NS = "NS"),
t_d = round((s2Ons - s2OnsDt) - (s1Ons - s1OnsDt), digits = 3),
t_d_alt = ifelse(soaIx %in% 0:2,
"short",
ifelse(soaIx == 3,
"intermediate",
ifelse(soaIx == 4,
"long",
NA_character_
)
)
)
) %>%
# Measured variables (see § 2.5.2. of the pre-registration) ==================
# Finger-level response count (RC_{finger})
# Finger-level response time (RT_{finger})
dplyr::rename(RC_LM = keyCount_f,
RC_RM = keyCount_b,
RC_LI = keyCount_e,
RC_RI = keyCount_a,
) %>%
dplyr::rename(RT_LM = rt1_f,
RT_RM = rt1_b,
RT_LI = rt1_e,
RT_RI = rt1_a
) %>%
# Indices (see § 2.5.3. of the pre-registration) ===========================
# Response category ----------------------------------------------------------
# - RB: Bimanual response, fingers compatible with the go-stimulus
# - RL: Left-hand response, finger compatible with the go-stimulus
# - RR: Right-hand response, finger compatible with the go-stimulus
# - RBO: Bimanual response, fingers compatible with the other go-stimulus
# - RLO: Left-hand response, finger compatible with the other go-stimulus
# - RRO: Right-hand response, finger compatible with the other go-stimulus
# - NR: No response
# - NOC: Not otherwise classified
dplyr::mutate(r = dplyr::case_when(
.$s1 == "M" & .$RC_LM == 1 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RB",
.$s1 == "M" & .$RC_LM == 1 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RL",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RR",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 1 ~ "RBO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 0 ~ "RLO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 1 ~ "RRO",
.$s1 == "M" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "NR",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 1 ~ "RB",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 1 & .$RC_RI == 0 ~ "RL",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 1 ~ "RR",
.$s1 == "I" & .$RC_LM == 1 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RBO",
.$s1 == "I" & .$RC_LM == 1 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RLO",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 1 & .$RC_LI == 0 & .$RC_RI == 0 ~ "RRO",
.$s1 == "I" & .$RC_LM == 0 & .$RC_RM == 0 & .$RC_LI == 0 & .$RC_RI == 0 ~ "NR"
)
) %>%
dplyr::mutate(r = ifelse(is.na(r),
"NOC",
r
)
) %>%
# Bimanual response recorded (r_{bi})
dplyr::mutate(r_bi = ifelse(r %in% c("RB","RBO"),
TRUE,
FALSE
)
) %>%
# Trial-level response time (RT_{trial}); also include RT difference for reporting of descriptive statistics
dplyr::mutate(RT_trial = rt1_mean,
RT_diff_trial = rtDiff1_mean) %>%
# Trial label - performance ($trial_{performance}$)
dplyr::mutate(trial_performance = dplyr::case_when(
.$trial == "NS" & .$r == "RB" ~ "NS_correct",
.$trial == "NS" & .$r == "RBO" ~ "NS_commission-error-bi",
.$trial == "NS" & .$r %in% c("RL","RR","RLO","RRO") ~ "NS_commission-error-uni",
.$trial == "NS" & .$r == "NOC" ~ "NS_commission-error-other",
.$trial == "NS" & .$r == "NR" ~ "NS_omission-error",
.$trial == "SL" & .$r %in% c("RB","RBO") ~ "SL_stop-respond-bi",
.$trial == "SL" & .$r %in% c("RL","RLO","RRO") ~ "SL_stop-respond-uni",
.$trial == "SL" & .$r %in% c("NR","NOC") ~ "SL_stop-respond-other",
.$trial == "SL" & .$r == "RR" ~ "SL_stop-inhibit",
.$trial == "SR" & .$r %in% c("RB","RBO") ~ "SR_stop-respond-bi",
.$trial == "SR" & .$r %in% c("RR","RLO","RRO") ~ "SR_stop-respond-uni",
.$trial == "SR" & .$r %in% c("NR","NOC") ~ "SR_stop-respond-other",
.$trial == "SR" & .$r == "RL" ~ "SR_stop-inhibit",
.$trial == "SB" & .$r %in% c("RB","RBO") ~ "SB_stop-respond-bi",
.$trial == "SB" & .$r %in% c("RL","RR","RLO","RRO") ~ "SB_stop-respond-uni",
.$trial == "SB" & .$r == "NOC" ~ "SB_stop-respond-other",
.$trial == "SB" & .$r == "NR" ~ "SB_stop-inhibit",
.$trial == "IG" & .$r == "RB" ~ "IG_correct",
.$trial == "IG" & .$r == "RBO" ~ "IG_commission-error-bi",
.$trial == "IG" & .$r %in% c("RL","RR","RLO","RRO") ~ "IG_commission-error-uni",
.$trial == "IG" & .$r == "NOC" ~ "IG_commission-error-other",
.$trial == "IG" & .$r == "NR" ~ "IG_omission-error"
)
)
# Select relevant variables (for confirmatory analysis) or return everything (for exploratory analysis)
if (analysis_type == 'confirmatory') {
RT_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, t_d, t_d_alt, RT_trial, RT_diff_trial)
resp_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, t_d, t_d_alt, r_bi, trialCorrect)
} else if (analysis_type == 'exploratory') {
RT_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, responseType, t_d, t_d_alt, RT_trial, RT_diff_trial, trialCorrect, trial_performance)
resp_df <-
clean_df %>%
dplyr::select(subjectIx, blockIx, trialIx, trial, trial_alt, r, responseType, t_d, t_d_alt, r_bi, trialCorrect, trial_performance)
}
list(RT_df = RT_df, resp_df = resp_df)
}
# ==============================================================================
#' Get column types for extracting data from csv files
#'
#' File type can be any of the following:
#'
#' expt_block_data - tidied block-level data from experimental session
#' expt_trial_resp_data - tidied trial-level response data from experimental session
#' expt_trial_rt_data - tidied trial-level resonse time data from experimental session
#' log_sess_cols - source data, session-related columns
#' log_block_cols - source data, block-related columns
#' log_trial_cols - source data, trial-related columns
#' log_trial_cols_sub00 - source data, trial-related columns for subject 00
#' prac_block_data - tidied block-level data from practice session
#'
#' @param file_type Type of file to be read
#' @export
get_col_types <- function(file_type){
switch(tolower(file_type),
expt_block_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
criterion = readr::col_factor(levels = c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")),
performance = readr::col_double(),
failed = readr::col_logical()
),
expt_trial_resp_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
trialCorrect = readr::col_logical()
),
expt_trial_resp_data_exploratory =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
trialCorrect = readr::col_logical(),
trial_performance = readr::col_character()
),
expt_trial_rt_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
RT_trial = readr::col_double()
),
expt_trial_rt_data_exploratory =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_factor(levels = c(0.066, 0.166, 0.266, 0.366, 0.466),
ordered = TRUE),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
RT_trial = readr::col_double(),
trialCorrect = readr::col_logical(),
trial_performance = readr::col_character()
),
log_sess_cols =
readr::cols_only(subjectIx = readr::col_integer(),
sessDate = readr::col_date(format = "%Y-%m-%d"),
studyId = readr::col_character(),
experimenterId = readr::col_character(),
taskVersionId = readr::col_character(),
responseDevice = readr::col_character(),
blockId = readr::col_character()
),
log_block_cols =
readr::cols_only(subjectIx = readr::col_integer(),
taskVersionId = readr::col_character(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
s1Acc_00 = readr::col_double(),
s1AccCritMet_00 = readr::col_logical(),
s1Acc_01 = readr::col_double(),
s1AccCritMet_01 = readr::col_logical(),
s2Acc_00 = readr::col_double(),
s2AccCritMet_00 = readr::col_logical(),
s2Acc_01 = readr::col_double(),
s2AccCritMet_01 = readr::col_logical(),
s2Acc_02 = readr::col_double(),
s2AccCritMet_02 = readr::col_logical(),
s2Acc_03 = readr::col_double(),
s2AccCritMet_03 = readr::col_logical(),
s1MeanRt_00 = readr::col_double(),
s1MeanRtCritMet_00 = readr::col_logical(),
s1MeanRt_01 = readr::col_double(),
s1MeanRtCritMet_01 = readr::col_logical(),
s1MeanRtDiff_00 = readr::col_double(),
s1MeanRtDiffCritMet_00 = readr::col_logical(),
s1MeanRtDiff_01 = readr::col_double(),
s1MeanRtDiffCritMet_01 = readr::col_logical()
),
log_trial_cols =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
fixIx = readr::col_integer(),
s1Ix = readr::col_integer(),
s2Ix = readr::col_double(),
soaIx = readr::col_double(),
tSession = readr::col_time(format = "%H:%M:%S"),
tBlock = readr::col_time(format = "%H:%M:%OS"),
trialOns = readr::col_double(),
trialDur = readr::col_double(),
fixOns = readr::col_double(),
fixOnsDt = readr::col_double(),
fixDur = readr::col_double(),
fixDurDt = readr::col_double(),
s1Ons = readr::col_double(),
s1OnsDt = readr::col_double(),
s1Dur = readr::col_double(),
s1DurDt = readr::col_double(),
s2Ons = readr::col_double(),
s2OnsDt = readr::col_double(),
s2Dur = readr::col_double(),
s2DurDt = readr::col_double(),
keyCount_f = readr::col_integer(),
keyCount_b = readr::col_integer(),
keyCount_e = readr::col_integer(),
keyCount_a = readr::col_integer(),
rt1_f = readr::col_double(),
rt1_b = readr::col_double(),
rt1_e = readr::col_double(),
rt1_a = readr::col_double(),
rt1_mean = readr::col_double(),
`rtDiff1_f-b` = readr::col_double(),
`rtDiff1_e-a` = readr::col_double(),
rtDiff1_mean = readr::col_double(),
trialCorrect = readr::col_logical(),
trialType = readr::col_character(),
responseType = readr::col_character(),
trialFeedback = readr::col_character()
),
log_trial_cols_sub00 =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
fixIx = readr::col_integer(),
s1Ix = readr::col_integer(),
s2Ix = readr::col_double(),
soaIx = readr::col_double(),
tSession = readr::col_time(format = "%H:%M:%S"),
tBlock = readr::col_time(format = "%H:%M:%OS"),
trialOns = readr::col_double(),
trialDur = readr::col_double(),
fixOns = readr::col_double(),
fixOnsDt = readr::col_double(),
fixDur = readr::col_double(),
fixDurDt = readr::col_double(),
s1Ons = readr::col_double(),
s1OnsDt = readr::col_double(),
s1Dur = readr::col_double(),
s1DurDt = readr::col_double(),
s2Ons = readr::col_double(),
s2OnsDt = readr::col_double(),
s2Dur = readr::col_double(),
s2DurDt = readr::col_double(),
keyCount_f = readr::col_integer(),
keyCount_h = readr::col_integer(),
keyCount_v = readr::col_integer(),
keyCount_b = readr::col_integer(),
rt1_f = readr::col_double(),
rt1_h = readr::col_double(),
rt1_v = readr::col_double(),
rt1_b = readr::col_double(),
rt1_mean = readr::col_double(),
`rtDiff1_f-h` = readr::col_double(),
`rtDiff1_v-b` = readr::col_double(),
rtDiff1_mean = readr::col_double(),
trialCorrect = readr::col_logical(),
trialType = readr::col_character(),
responseType = readr::col_character(),
trialFeedback = readr::col_character()
),
prac_block_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockId = readr::col_character(),
block_type = readr::col_factor(levels = c("NS","AS","SS","mixed"),
ordered = TRUE),
attempt = readr::col_factor(levels = c(1, 2, 3, 4, 5),
ordered = TRUE),
criterion = readr::col_factor(levels = c("NS_accuracy","NS_mean_RT","NS_mean_RTdiff","SL_accuracy","SR_accuracy","SB_accuracy","IG_accuracy")),
performance = readr::col_double(),
failed = readr::col_logical()
),
sim_trial_data =
readr::cols_only(subjectIx = readr::col_integer(),
blockIx = readr::col_integer(),
trialIx = readr::col_integer(),
trial = readr::col_factor(levels = c("NS", "SL", "SR", "SB", "IG"),
ordered = FALSE
),
trial_alt = readr::col_factor(levels = c("SAS", "SSS", "NS"),
ordered = TRUE
),
r = readr::col_character(),
responseType = readr::col_character(),
t_d = readr::col_double(),
t_d_alt = readr::col_factor(levels = c("short", "intermediate", "long"),
ordered = TRUE
),
r_bi = readr::col_logical(),
RT_trial = readr::col_double(),
SSRT_true = readr::col_double(),
trialCorrect = readr::col_logical(),
trial_performance = col_character()
)
)
}
#' Splits log files into a session table and a block/trial table
#'
split_log_in_multi_table <- function(df) {
session_data <-
df %>%
dplyr::select(studyId:sessTime, -rngSeed) %>%
dplyr::distinct()
}
# ==============================================================================
#' Verifies index columns
#'
#' Identifies index columns, replaces NaNs with NAs, and converts objects to integer
#' @export
verify_ix_cols <- function(df) {
col_vector <- stringr::str_subset(colnames(df),".*Ix$")
df[col_vector] <- df[col_vector] %>%
purrr::map(~ stringr::str_replace(.x, "nan", "NA")) %>%
purrr::map(~ as.integer(.x))
df
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.