Two-dimensional reward evaluation in mice
The folder "analysis/data/raw" contains csv files with raw data from full experimental runs of groups of mice inside a home cage with four water feeders.
The files were produced by control software PhenoSoft Control v.1.0.5163.31723 (PhenoSys, GmbH).
The folder "analysis/data/metadata" contains input csv files ("mastercolnames", "MasterTable", "Conditions", and "vol_measurements") necessary for the R scripts that analyze the raw data as well as the csv files produced as output from other scripts.
The folder "analysis/R" contains R scripts, including a script for reading and aggregating csv files from PhenSoft Control into a single data frame (load.R) and a script for performing the simulations (simulations.R). The output from these scripts are processed csv files, saved in the folder "analysis/data". The output from load.R is "ChoicesOutput.csv" and the remaining files are outputs from the simulations.R script (simulations.csv and files with sensitivity_xxx.csv naming scheme).
| Column label | Type | Information |
|---|---|---|
| DateTime | - | The astronomical date and time for each event of the experiment. Missing days due to experimental malfunctions, see below |
| IdRFID | - | RFID number of mouse |
| IdLabel | - | Short unique mouse identifier |
| unitLabel | - | Code of the unit that was activated, with numbers corresponding to the corners of the cage |
| LS | Detections of IR-beam interruptions without a transnponder number detection | |
| Reader | Detections of transponder numbers without IR-beam interruptions | |
| CondMod | Detections of both a transponder number and an IR-beam interruption (identified nose pokes) | |
| pumpBeh, Empty, Full | Events relating to states of the syringe and its refilling algorithm | |
| VersuchCS | Events related to the main program, clarified in SystemMsg | |
| exp | Events related to the programmed reward schedule, clarified in SystemMsg | |
| eventDuration | Duration of event in milliseconds | |
| sense1duration | Total duration of the IR-beam interruption | |
| sense1Events | Number of interruptions of IR beam. When such events happen fast enough (less than 200ms apart) these are registered as a single event, but the number of such short interruptions is given here | |
| senseRFIDrecords | Number of times the transponder number was detected; irrelevant for this experiment | |
| reinforce1value | Reward (in pump step units, delivered by a stepping motor syringe pump filled with water) | |
| reinforce1Total | Contents of this column are irrelevant for this experiment | |
| reinforce1Account | Contents of this column are irrelevant for this experiment | |
| outFuncLabel | Contents of this column are irrelevant for this experiment | |
| outLabel | Contents of this column are irrelevant for this experiment | |
| SystemMsg | System state messages, further clarified in the unitLabel column and indicating the beginning or end of, e.g. a preprogrammed reward schedule (no rewards delivered before start and after end of schedule) and beginning or end of syringe refill event (no rewards delivered while refilling) | |
| MsgValue1 | Name of parameter file started with the control software | |
| MsgValue2 | Contents of this column are irrelevant for this experiment | |
| MsgValue3 | Contents of this column are irrelevant for this experiment |
This file is the output of the load.R script which processes the folder of raw csv files, with further information supplied by "Conditions", "MasterTable", and "mastercolnames" csv files.
| Column label | Information |
|---|---|
| day | Experimental day (first true experimental day is 0, negative numbers for training and pre-tests) |
| IdLabel | Short unique mouse identifier |
| loc | Dispenser location, given as corner number clockwise from 1 to 4 |
| DateTime | Astronomical date and time for each event of the experiment. Missing days due to experimental malfunctions, see below |
| eventDuration | Duration of event in milliseconds |
| SystemMsg | Rewards that were supposed to be delivered but could not be delivered because the system was in a "busy" state are marked as "busy" |
| vol | Volume (in microliters) of delivered reward |
| rewardstatus | 1 if a reward was delivered, 0 otherwise |
| prob | long-term probability that a reward will be delivered for the particular dispenser location and mouse |
| volm | Maximal reward volume (rewards can be either this value or 0) for a particular dispenser location and mouse |
| rel | 1 if rewarding dispenser, 0 otherwise |
| prof | 1 if higher profitability dispenser, 0 for the lower profitability dispenser and NA for the nonrewarding dispensers. For pairs of options of equal profitability in the incongruent (I) condition, 1 denotes the higher volume option and 0, the higher probability option. In the "explore" and "training" conditions all |
| cohort | Cohort number |
| cond | Experimental condition. Before the main experiment, mice were in "explore" then "training" condition. In these conditions all four dispensers were rewarding with the same probabilities and volumes, but a prof of 1 was randomly assigned to two dispensers and a prof of 0 to the other two dispensers |
| experiment | Experiment number |
This file is generated by the simulations.R script and contains the main simulations in which different decision models are tested under the same experimental conditions from experiments 1-4.
| Column label | Information |
|---|---|
| ind | ID of the simulated individual |
| experiment | Experiment number |
| cond | Experimental condition |
| model | Abbreviated name of the model |
| performance | Discrimination performance over 1000 visits, relative choice for higher volume (in I condition) or higher profitability option (all other conditions) |
These files are generated by the simulations.R script and contain the sensitivity analyses data used to select the best free parameter values for each model.
| Column label | Information |
|---|---|
| par | value of other free parameter: threshold in pfirst and vfirst models and theta_v for rnonc model |
| cond | Experimental condition |
| gamma | Value of gamma parameter |
| performance | Discrimination performance over 1000 visits, relative choice for higher probability option |
| pred_perf | Predicted performance from loess model over all gamma values |
| obs_perf | Observed performance, median over all experimental mice for the baseline conditions "BPV1" and "BPV2" |
| deviance | (pred_perf - obs_perf)^2 |
This file is user-generated, providing relevant experimental information not present in the raw files.
| Column label | Information |
|---|---|
| day | Experimental day (first true experimental day is 0, negative numbers for training and pre-tests) |
| IdLabel | Short unique mouse identifier |
| loc | Dispenser location, given as corner number clockwise from 1 to 4 |
| prob | long-term probability that a reward will be delivered for the particular dispenser location and mouse |
| volm | Maximal reward volume (rewards can be either this value or 0) for a particular dispenser location and mouse |
| rel | 1 if rewarding dispenser, 0 otherwise |
| prof | 1 if higher profitability dispenser, 0 for the lower profitability dispenser and NA for the nonrewarding dispensers. For pairs of options of equal profitability in the incongruent (I) condition, 1 denotes the higher volume option and 0, the higher probability option |
| discard | 1 if data for particular day and mouse is to be discarded from analysis, 0 otherwise |
| cohort | Cohort number |
| cond | Experimental condition. Before the main experiment, mice were in "explore" then "training" condition. In these conditions all four dispensers were rewarding with the same probabilities and volumes, but a prof of 1 was randomly assigned to two dispensers and a prof of 0 to the other two dispensers |
| experiment | Experiment number |
This file is user-generated and allows mapping the raw csv files to the respective experimental days.
| Column label | Information |
|---|---|
| day | Experimental day (first true experimental day is 0, negative numbers for training and pre-tests) |
| path | Path of the raw csv file corresponding to the day |
| comments | When days are omitted a clarification is given here |
| Column label | Information |
|---|---|
| cage | Cage number (1 or 2) |
| loc | Dispenser location, given as corner number clockwise from 1 to 4 |
| steps | Number of steps that the stepping motor was set to |
| vol | Measured volume [microLiters] with a graduated pipette (resolution of 1 microLiter) |
| ulstep | vol/steps, microLiters per step |
For further information contact: [email protected]