Skip to content

Behavioral Data Format

John Pearson edited this page Aug 4, 2016 · 3 revisions

Formatting Behavioral Data

Behavioral data can be saved in any reasonable format, though human-readable formats like json and yaml are preferred.

However, each member of the lab is responsible for converting his or her behavioral data to the following json format for storage on AWS:

The final data file:

  1. Should be a single json object (i.e., not one object per trial)
  2. Should contain an object meta with fields
  3. subject: unique subject number
  4. start_time: wall (clock) start time for experiment - This time should be parseable as a time format (e.g., "9:35:01:898848" for 9:35 a.m.) - (TODO: other restrictions on wall time)
  5. stop_time: wall (clock) stop time for experiment
  6. experiment: name of experiment
  7. day: day within experiment (for multi-day experiments) - = 0 for single-day experiments or day of surgery - = days post surgery for surgical cases
  8. Other suggested meta fields: - PsychoPy start and stop times from global clock - time dilation factor if task is slowed/sped up - task version number
  9. Should contain an object data containing all other experimental data
  • It is strongly suggested that this be a list of objects, one per trial