Position and time coordinates, performance, and latencies of delayed non-match to place trials performed by rats with silenced hippocampal inputs to RSC
Data files
Jul 29, 2025 version files 53.29 MB
-
Data_from_Bárbara_P-C_et_al_2024_eLife.zip
53.28 MB
-
README.md
8.86 KB
Abstract
This dataset contains processed behavioral data obtained from Long-Evans rats whose hippocampal axon terminals in RSC were infected with a viral vector containing a transgene expressing either eArchT or only a fluorescent tag ("empty vector" CTRL), while animals perform a standard delayed non-match to place task, in which light was delivered via fiberoptic to such HIPP terminals at randomized individual trials. We then use extracted xyt position data to collect behavioral metrics and compare such metrics (performance levels, time delays) between these animals and CTRL animals under baseline, no-light, and light delivery conditions, as well as latencies at several points of individual trials. We also extracted specific trials according to specific conditions explained in the original eLife publication, and below. Files are either Excel spreadsheets with all the properties of each individual trial, or .omv files containing the same data as needed to perform statistical analysis in Jamovi. One folder named Position Data contains all xyt coordinates from all trials of each individual animal in each individual folder. We have included the Python scripts used to produce all the above, as .IPYNB files.
This dataset includes and the data, code and analyses used in the Manuscript: “Optogenetic silencing hippocampal inputs to the retrosplenial cortex causes a prolonged disruption of working memory” by Barbara Pinto-Correia, Patrícia Caldeira and Miguel Remondes (corresponding author).
This includes behavioral data (position and time coordinates, performance, latencies) of delayed non-match to place trials performed by rats in which hippocampal inputs to RSC have been silenced, as well as the appropriate control conditions and animals.
https://doi.org/10.5061/dryad.fxpnvx11z
In the folder “01-Position Data” there are individual folders from individual eArch and CTRL (empty vector) rats, named as such. In each of these folders we make available the pre-processed .csv files with xyt positions corresponding to individual trajectories and its specifications, and .csv files with ROI coordinates used for further analyses.
Finally, in most folders of this dataset data are stored as .csv Excel files and .omv Jamovi files used in statistical analysis.
Folder “02-Analysis_Performances\Individual Trial Data _ DNMP performance” contains the results of DNMP proficiency data as .csv files.
Folder “02-Analysis_Performances\Jamovi GLMM Files _ DNMP performance” contains the same data organized as Jamovi files, as well as GLMM Jamovi files (.omv) and statistical analysis results.
Folder “03-Analysis_Time_in_CP” contains equivalent data as before, but pertaining to time spent at the Choice Point (CP).
ANALYSIS WORKFLOW:
Pre-processed data: These are stored in individual animal dataset folders inside the main folder “01-Position_Data”, an example being “Banner_ARCH3” signifying the ARCH3 animal named “Banner”. Data was first collected as the xy coordinates of the rat’s centroid and corresponding timestamps. Data was then pre-processed to identify the runs, the choice point (CP), reward ports location (RW), and attribute to each run the specifications: number, run type (sample or test), stimulation condition (baseline, illumination, or non-illumination), outcome, rat code, session code, and group. Pre-processing generated the pre-processed data files necessary for further data analysis (scripts described below). These scripts also output the figures used in the publication, and the datasets used in statistical testing (described below).
Each individual animal folder is labeled with a “RAT CODE” which corresponds to the expressed vector (ARCH or CTRL) and the rat number (eg, ARCH3). Inside each of these folders 4 Excel files of pre-processed data can be found:
RATCODE_timestamped_position_df_pruned
Contains the session code (session, date and time), run number (run_nr), timestamp per data point collected in ms, position as xy coordinates in cm, the x coordinate differential needed to correctly split the runs (x_diff), run type (run_type, sample [s] or test [t]), stimulation condition (baseline [-1], illumination [1], no illumination [0]), and outcome (correct [1] or incorrect [0]).
And,
RATCODE_cp_rois_converted,
RATCODE_rw1_rois_converted,
RATCODE_rw2_rois_converted.
Containing the positions of the choice point and reward port locations in each session: xy coordinates in cm, width and heigh in cm, session code (session, date and time).
Analysis code and software
These are stored in folder “04-IPYNB scripts”.
All scripts were written in Python and the .IPYNB format can be opened using Jupyter Notebooks. The software is free and can be installed on the desktop or used from the browser (see https://jupyter.org). The scripts are heavily commented to allow comprehension of each step, and Jupiter notebooks optimize script organization which facilitates understanding.
The following scripts were used analyze the pre-processed data:
- Performance Assessment_yy axis change - Used to analyze the performance of baseline and protocol sessions, probability of correct following an illuminated trial, and probability of error following an error trial
- Probability of Correct Given Illumination-2 - Used to analyze the probability of correct when trial (t) t-1/t-2/t-3/t-4 was illuminated; and the probability of correct following 2 consecutively non-illuminated trials
- Time Spent in Choice Point Normalized-2 - Used to analyze the time spent on the choice point (CP) overall, and on illuminated and non-illuminated trials by outcome.
The above scripts generate the figures and data used for the statistical analyses.
Processed data (.csv files):
These are stored in the folder “05-Processed Data”.
baseprot_runs,
protocol_trials,
prob_correct_illum_binomial,
prob_error_given_error_binomial,
prob_error_given_error_baseline_sessions,
prob_correct_given illum_all,
CP_log_normalized
Some of these files may contain empty cells, which mean not applicable (eg, future outcome of the last trial of a session does not exist). Empty cells were not imputed as they were generated as NaN in python and are read as such.
- baseprot_runs - for performance assessment of baseline and baseline vs. illumination (“protocol”) sessions (corresponding to plot of figure 2C and 3B). Contains session code, run number, stimulus (“illumination”) condition, outcome, rat code, and group.
- protocol_trials - for performance assessment of illumination (protocol) sessions (corresponding to plot of figure 3A). Contains session code and number, run number, stimulus condition, outcomes, rat code, and group.
- prob_correct_illum_binomial - probability of the current trial being correct following an illuminated trial (corresponding to plot of figure 3C). Contains the variables mentioned above plus future outcome incorrect or correct (fut_outcome: 0,1).
-prob_error_given_error_binomial - probability of error following an error (corresponding to plot of figure 4A). Contains the variables mentioned in the above dataset plus “count2” which = 1 when the current and next trial is incorrect.
- prob_error_given_error_baseline_sessions - probability of error following an error trial in baseline session (corresponding to plot of figure 4B). Contains the variables above but only for baseline sessions (stim_condition == -1)
- prob_correct_given illum_all - probability of correct when trial t-n was illuminated (corresponding to plot of figure 5A). Contains trial number, stimulus condition (always illuminated == 1), outcome, session number, rat code, group, and the outcomes of future trials 1 to 5 (t+1, t+2, t+3, t+4, t+5).
- CP_log_normalized - log(time) spent at the choice point in Arch vs. CTRL(corresponding to plot of figure 6A and 6B). Contains session code, session number, rat code, run types (always test), run number, stimulus conditions, group, and time spent at the CP (time_spent), the log of that time (log10), the mean time spent at baseline for each rat (baseline), and the log of the time spent at the choice point normalized to the baseline time for each rat (normalized log).
Statistical analysis, datasets, and software
These are store in folder “06-Stats Analysis Jamovi Files”.
Statistical testing, including generalized linear mixed models, was conducted in JAMOVI using the datasets outputted from the above scripts. Jamovi files are in .omv format. Jamovi is a free software that can be installed on the desktop or used directly from the browser (see https://www.jamovi.org/)
- baseline_analysis_binomial - performance eArch+ vs. CTRL in baseline sessions (corresponding to plot of figure 2C). Note that the baseline sessions stim_condition was changed from -1 to 2, because Jamovi did not read -1.
- performance_binomial - performance of illuminated vs. non-illuminated trials (corresponding to plot of figure 3A)
- base_prot_analysis - performance of baseline vs. illuminated vs. non-illuminated trials (note that the baseline sessions stim_condition was changed from -1 to 2, because Jamovi did not read -1); (corresponding to plot of figure 3B)
- prob_correct_illum_binomial - probability of the current trial being correct following an illuminated trial (corresponding to plot of figure 3C)
- prob_error_given_error_binomial - probability of correct following an error (corresponding to plot of figure 4A)
- prob_error_given_error_baseline_sessions - probability of error following an error trials in baseline session (corresponding to plot of figure 4B)
- prob_correct_given illum_all - probability of correct when trial (t) t-1/t-2/t-3/t-4 was illuminated (corresponding to plot of figure 5A)
- CP_log_normalized - log(time) spent at the choice point in Arch vs. CTRL overall (corresponding to plot of figure 6A); and by illumination and outcome (corresponding to plot of figure 6B)
Data was collected from video recordings using Bonsai software (bonsai-rx.org). Error/correct was extracted directly, and XYT coordinates were used to determine individual trial latencies. Please refer to the original publication for detailed methods.