Data Dictionary: TRAINING.csv

This file defines the meaning and origin of each column in TRAINING.csv, the main dataset used by the Shiny app.

Column Descriptions

Column	Description
`file`	Filename of the original `.mat` file
`settings_file`	Settings file basename used in the session
`protocol`	Protocol name parsed from the filename
`data_source`	Source type: either `"bcontrol"` or `"bpod"`
`experimenter`	Name of the person who ran the session
`animal_id`	Unique identifier of the subject
`rig_id`	ID of the behavioral rig used
`date`	Session date (parsed from save timestamp)
`start_time`	Start time (converted from load timestamp)
`save_time`	Save time (timestamp of session end)
`stage`	Stage label (e.g., `"1_center_poke_on"`, `"3_NoReward"`)
`right_trials`	Number of right-sided choices
`left_trials`	Number of left-sided choices
`right_hit_frac`	Hit fraction for right trials (if available)
`left_hit_frac`	Hit fraction for left trials (if available)
`all_trials`	Total number of trials attempted
`completed_trials`	Trials with valid outcome (correct or error)
`correct_trials`	Number of correct (i.e., rewarded) trials
`error_trials`	Number of incorrect responses
`violation_trials`	Number of rule violations during the session
`timeout_trials`	Number of timeout trials
`init_CP`	Initial center poke duration
`total_CP`	Total center poke duration
`A1_time`, `A2_time`	Time spent in certain state/action windows
`reward_type`	Type of reward policy used
`session_length`	Derived from `save_time - start_time`
`choice_direction`	Either `"left_trials"` or `"right_trials"` (long format)
`No_pokes`	Number of pokes for the given choice direction

Some fields (like reward_type, stage, and A2_time) are re-coded during the load_data.R cleaning step.
The file is reshaped to long format before plotting: one row per choice direction (left_trials or right_trials).
Fields marked as "empty_field_in_mat_file" were missing in the original .mat.