Habit Task Data Parsing

Reward Learning and Habituation Task EEG, MR, and Behave data is coalesced by scripts in the task directory /Volumes/Hera/Projects/Habit/task/results but relies on task parsing scripts in /Volumes/Hera/Projects/Habit/mr/habit (MR directory, but parsing behave and eeg too). Age (when available) is pulled from yet another location, ../../redcap/data/ages.tsv (make -C /Volumes/Hera/Projects/Habit/redcap data/ages.tsv)

Task files of interest are lab.data.tsv and lab.summary.csv

trial_file=/Volumes/Hera/Projects/Habit/task/results/lab.data.tsv # also see lab.summary.csv
wc -l < $trial_file
# 80279

head -n 3 $trial_file
id      initials        age     sex     task    vdate   timepoint       run     ver     avatar  trial   block   left_pos        left_prob      up_pos   up_prob right_pos       right_prob      score   rt      n_keys  first_key_rt    first_key       picked  picked_prob     picked_step    avoided  avoided_prob    avoided_step    iti_onset       chose_onset     waiting_onset   feedback_onset  timeout_onset   trial_choices   survey_age      pick_dist       optimal_choice  blocknum        blocktype       picked_unified  avoid_unified
11885   NA      24      NA      habit_eeg       NA      20221012        1       eeg_1   shark   1       Lc20Uc50Rf100   c       20      c      50       f       100     true    557     1       510     39      right   100     2       up      50      1       211415.96       212424.14      212980.92        213243.78       NA      up-right        24      far     TRUE    1       init    first100        first50
11885   NA      24      NA      habit_eeg       NA      20221012        1       eeg_1   shark   2       Lc20Uc50Rf100   c       20      c      50       f       100     false   450     1       430     37      left    20      1       up      50      1       213536.44       214554.2       215004.04        215236.92       NA      left-up 24      close   FALSE   1       init    first20 first50

JSON files are exported from the browser at the end of the task. They're on bea_res like:

/Volumes/L/bea_res/Data/Tasks/Habit/MR/11734_20221111/sub-11734_task-mr_habit_ses-20221005_run-1_1668193517800.json

Age is pulled from redcap and qualtrics. Fallback db query fills for ages know via other studies. /Volumes/Hera/Projects/Habit/behave/redcap/02_visit_ages.R does all of this. Despite living in the REDCap scripts dir, it pulls to Qualtrics and DB).

Extracting and coalescing data happens with MR scripts. See make json2tsv (tsv per task file) and txt/all_habit.json (single large file).

cd /Volumes/Hera/Projects/Habit/mr/habit
make  json2tsv
# runs eg.
#  ./read_idv_task.bash "/Volumes/L/bea_res/Data/Tasks/Habit/MR/12010_20251124/12010_mr_habit_20251001_1_nover.json" "/Volumes/L/bea_res/Data/Tasks/Habit/MR/12010_20251124/12010_mr_habit_20251001_1_nover.tsv"

read.R adds picked_unified avoid_unified. All trials are put into one long file.

These transformation and scoring scripts are

• on rhea as /Volumes/Hera/Projects/Habit/task/results
• symlinked as /Volumes/Hera/Projects/Habit/mr/habit/task-results
• tracked with task presentation code: https://github.com/LabNeuroCogDevel/choice-landscape/tree/master/results

Scripts from age extraction to final lab.data.tsv should run unattended nightly using Cron and make.

crontab -l|grep habit -i # foranw
30 2 * * * make -C /Volumes/Hera/Projects/Habit/behave/redcap
00 4 * * * make -C /Volumes/Hera/Projects/Habit/mr
00 5 * * * make -C /Volumes/Hera/Projects/Habit/task results/lab.summary.csv