Data is organized into pandas dataframes.
-
Subject Data: Information about each subject who took the test
-
Trial Data: Subject response data from all trials in the test across all days
-
Test Contexts: The stages of the song learning ladder to cross reference with the Trial Data
-
Stimulus Files: WAV files played as stimuli during the test, referenced by Trial Data
| Subject Name | Sex |
|---|---|
| GreBla5671F | F |
| GreBla7410M | M |
| WhiRed9510F | F |
| RedBla0907M | M |
| XXXOra0037F | F |
| HpiBlu6194F | F |
| YelPur7906M | M |
| WhiWhi2526M | M |
| BluYel2571F | F |
| YelRed3010F | F |
| GraWhi4040F | F |
| BlaGre1349M | M |
| XXXHpi0038M | M |
| GreBlu5039F | F |
| GreBla3404M | M |
| XXXRed0088M | M |
| XXXOra0039F | F |
| XXXBla0054F | F |
| XXXBla0055M | M |
| XXXBla0081M | M |
Trial data represents the stimulus playback and behavioral response data for all trials and subjects. Data has been cleaned to remove trials triggered erroneously due to hardware issues (e.g. peck double registering due to button sensitivity, button getting stuck in the down position).
| Column Name | Data Type | Description |
|---|---|---|
| Subject | String | |
| Trial | Integer | Trial number (within day) |
| Time | Datetime | System time of trial start |
| Date | Date | |
| Interrupt | Boolean | True if subject pecked to interrupt playback |
| RT | Float | Response time (in seconds) |
| Stimulus File | String | Name of stimulus wav file |
| Stimulus Vocalizer | String | Name of vocalizing subject |
| Stimulus Call Type | String | Call type (SO or DC) |
| Stimulus Class | String | "Rewarded" or "Unrewarded" |
| Rewarded | Boolean | True if subject received food reward (derived from Class and Interrupt) |
| Informative Trials Seen | Integer | Number of times that a stimulus from this vocalizer had previously been uninterrupted |
| Test Context | String | Test context (references Test Context table) |
| Condition | String | Normally NaN, but for some tests that occured after a month without reinforcement, indicated by "MonthLater" |
| Subject Sex | String | "M" or "F" |
Test contexts are referenced in the Test Context column of TrialData.csv and indicate what stage of the learning ladder the subject was on.
| Ladder | Test Name | # Rewarded Vocalizers | # Non-rewarded Vocalizers | Description |
|---|---|---|---|---|
| Week 1 / Week 2 | SovsSo_1v1 | 1 song | 1 song | |
| SovsSo_4v4 | 4 songs (3 new) | 4 songs (3 new) | ||
| SovsSo_8v8_d1 | 8 songs (4 new) | 8 songs (4 new) | New vocalizers played twice as frequently | |
| SovsSo_8v8_d2 | 8 songs | 8 songs | All vocalizers played at equal frequency | |
| DCvsDC_1v1 | 1 dc | 1 dc | ||
| DCvsDC_4v4 | 4 dcs (3 new) | 4 dcs (3 new) | ||
| DCvsDC_6v6_d1 | 6 dcs (2 new) | 6 dc (2 new) | New vocalizers played twice as frequently | |
| DCvsDC_6v6_d2 | 6 dcs | 6 dcs | All vocalizers played at equal frequency | |
| Week 3 / Week 4 | SovsSo_1v1_S2 | 1 song | 1 song | |
| SovsSo_4v4_S2 | 4 songs (3 new) | 4 songs (3 new) | ||
| SovsSo_8v8_d1_S2 | 8 songs (4 new) | 8 songs (4 new) | New vocalizers played twice as frequently | |
| SovsSo_8v8_d2_S2 | 8 songs | 8 songs | All vocalizers played at equal frequency | |
| DCvsDC_1v1_S2 | 1 dc | 1 dc | ||
| DCvsDC_4v4_S2 | 4 dcs (3 new) | 4 dcs (3 new) | ||
| DCvsDC_6v6_d1_S2 | 6 dcs (2 new) | 6 dc (2 new) | New vocalizers played twice as frequently | |
| DCvsDC_6v6_d2_S2 | 6 dcs | 6 dcs | All vocalizers played at equal frequency | |
| Week 5 | DCvsDC_12v12 | 12 dcs | 12 dcs | Combined stimuli from DCvsDC_6v6_d2 and DCvsDC_6v6_d2_S2 |
| SovsSo_16v16 | 16 songs | 16 songs | Combined stimuli from SovsSo_8v8_d2 and SovsSo_8v8_d2_S2 | |
| Week 6 | AllvsAll_4v4 | 2 songs + 2 dcs | 2 songs + 2 dcs | All vocalizers previously learned in earlier sets (refresher set) |
| AllvsAll_28v28 | 16 songs + 12 dcs | 16 songs + 12 dcs | Combined stimuli from DCvsDC_12v12 and SovsSo_16v16 |
Stimulus .wav files are saved in a zip folder, the download is available on Google Drive. Paths in the Trial Data reference directories relative to the top level of the stimuli folder.
Filenames are formatted with the following convention:
CALLTYPE_Stim_RENDITION_VOCALIZER_RENDITIONID_norm[_COPY].wav
Code is run and tested on Python3.6, using requirements listed in requirements.txt (install with pip install -r requirements.txt). A local installation of R is required for statistical functions and installation of rpy2 library.