Update behavioral preprocessing notebook with relying on new script s…

…ource

R/behavioral_data_preprocessing.R

@@ -0,0 +1,59 @@
+#' @title Get hit times vector from signals and responses per participant
+#' @description This function uses input of both signal times and response times
+#'   from a single participant's extracted behavioral data to determine the hit
+#'   indices, if any, and in turn, the hit times.
+#' Get hit timestamp from a vector of signal times and a vector of response times
+#' @note Use the .interval variable if 8 seconds is not the desired interval
+#' @export
+get_hit_times <- function(signal_times, response_times, .interval = 8.0) {
+  signal_times %>%
+    map_dbl(function(signal_time) {
+
+      hit_index <- first(which(
+        response_times %>%
+          map_lgl(~ between(.x, signal_time, signal_time + .interval)),
+        arr.ind = TRUE
+      ))
+
+      hit_time <- response_times[hit_index]
+
+      return(hit_time)
+
+    })
+}
+
+#' @title Get all participants' hits with reaction times
+#' @description Creates a dataframe composed of each participant's hit times and
+#'   reaction times for those hits, row-by-row with signal times.
+#' @export
+get_all_hits_with_reaction_times <- function(participants, combined_df) {
+
+  # Extract only rows where a signal is present
+  all_signals_df <- combined_df %>%
+    filter(is_signal == 1) %>%
+    mutate(
+      signal_time = step_time
+    ) %>%
+    select(trial, id, image_index, signal_time)
+
+  # Extract only rows where a response attempt is present
+  all_responses_df <- combined_df %>%
+    filter(is_response == 1) %>%
+    select(trial, id, image_index, resp_time)
+
+  # Map over the unlisted participants' ids to get the per-participant
+  # signals and responses, then return a combined dataframe of all participant
+  # including trial rows for signals, and if it exists, hit time and reaction time
+  map_dfr(unlist(participants), function(participant) {
+    signals <- all_signals_df %>%
+      filter(id == participant)
+
+    responses <- all_responses_df %>%
+      filter(id == participant)
+
+    signals %>% mutate(
+      hit_time = get_hit_times(signals$signal_time, responses$resp_time),
+      reaction_time = hit_time - signal_time
+    )
+  })
+}

notebooks/behavioral_data_preprocessing.Rmd

@@ -74,75 +74,12 @@ combined_df %>%
 
 ## Function definitions
 
-The function to get each participant's hit time:
-
-```{r}
-HIT_INTERVAL <- 8
-
-# Get hit timestampe from a vector of signal times and a vector of response times
-# Use the .interval variable if HIT_INTERVAL is not defined or a different hit interval is desired
-get_hit_time <- function(signal_times, response_times, .interval = HIT_INTERVAL) {
-  signal_times %>% 
-    map_dbl(function(signal_time) {
-      
-      hit_index <- first(which(
-        response_times %>% 
-          map_lgl(~ between(.x, signal_time, signal_time + .interval)),
-        arr.ind = TRUE
-      ))
-      
-      hit_time <- response_times[hit_index]
-      
-      return(hit_time)
-      
-    })
-}
-```
-
-The function to get a dataframe of combined hits, including the hit time itself,
-and the reaction time between that hit time and the signal prompting that response.
-
-```{r}
-# Get combined hits dataframe composed of each participant's hit times and
-# reaction times for those hits, row-by-row with signal times.
-# Uses both combined signals and responses
-get_combined_hits <- function(participants, combined_df) {
-  
-  # Extract only rows where a signal is present
-  combined_signals_df <- combined_df %>%
-    filter(is_signal == 1) %>%
-    mutate(
-      signal_time = step_time
-    ) %>%
-    select(trial, id, image_index, signal_time)
-  
-  # Extract only rows where a response attempt is present
-  combined_responses_df <- combined_df %>%
-    filter(is_response == 1) %>%
-    select(trial, id, image_index, resp_time)
-  
-  # Map over the unlisted participants' ids to get the per-participant
-  # signals and responses, then return a combined dataframe of all participant
-  # including trial rows for signals, and if it exists, hit time and reaction time
-  map_dfr(unlist(participants), function(participant) {
-    participant_signals <- combined_signals_df %>%
-      filter(id == participant)
-    
-    participant_responses <- combined_responses_df %>%
-      filter(id == participant)
-      
-    participant_signals %>% mutate(
-        hit_time = get_hit_time(participant_signals$signal_time, participant_responses$resp_time),
-        reaction_time = hit_time - signal_time
-      )
-  })
-}
-```
+The function to get each participant's hit times vector is `itrackvalr::get_hit_times`, and the function to get a dataframe of all participants' hits is `itrackvalr::get_all_hits_with_reaction_times`, including each hit time and reaction time between that hit time and the nearest signal prompting a response within a fixed interval.
 
 ## Get the combined hits using the function
 
 ```{r}
-combined_hits_df <- get_combined_hits(participants, combined_df)
+combined_hits_df <- get_all_hits_with_reaction_times(participants, combined_df)
 ```
 
 ## Check out a quick preview of the table of hits
@@ -168,9 +105,9 @@ knitr::kable(head(
 ))
 ```
 
-## Show plot of reaction times by signal times
+## Gut-check plot of reaction times by signal times
 
-For hits, the reaction time does increase slightly across all participants.
+Note the 1.000s to 1.100s gap where hits were not recorded.
 
 ```{r}
 combined_hits_df %>%
@@ -181,9 +118,29 @@ combined_hits_df %>%
     theme_classic()
 ```
 
+## Reaction times per participant centered at the median
 
-
-
-
+```{r, fig.height = 10, fig.width = 10}
+combined_hits_df %>%
+  drop_na() %>%
+  arrange(reaction_time) %>%
+  ggplot(aes(x = reorder(id, reaction_time, FUN = median), y = reaction_time)) +
+    geom_point(alpha = 0.8, size = 0.7, color = 'darkblue', position = 'jitter') +
+    geom_boxplot(alpha = 0) +
+    geom_hline(yintercept = 1, color = 'coral') +
+  theme_minimal() +
+  theme(
+    aspect.ratio = 1,
+    text = element_text(size = 15),
+    plot.margin = margin(18, 18, 18, 18, 'pt')
+  ) +
+  labs(
+    title = 'Reaction times of participants by id',
+    subtitle = 'Boxplot per participant anchored at median of reaction time',
+    y = 'Reaction time for HIT after double-tick signal',
+    x = 'Participant'
+  ) +
+  coord_flip()
+```