editor_options	markdown wrap sentence

Shrew Foraging

This repo contains scripts and code associated with a behavioural experiment conducted in common shrews (Sorex araneus) between 20xx-20-xx.

The set-up consists of a squared arena (110x110 cm), base covered in sand and closed on top by a transparent plastic layer. Inside the arena, 4 hexagonal feeders are disposed symmetrically at the same distance to each corner of the apparatus. Each of the 4 feeders (named A, B, C, D) features 6 openings covered by a little sliding door.

During the experiment, two of the feeders (A and D) are filled with one mealworm for each opening (6 in total), while the other two (feeders B & C) are empty. The experiment consists of 4 sessions over two consecutive days. Each trial was composed by two session (S1 and S2), so in total every shrew performed the experiment in four sessions among two different days (T1_S1, T1_S2 - first day and T2_S1, T2_S2 second day Each session lasts 30 minutes, but the shrew is free to roam inside the arena, interact with the feeders and open the feeders' doors.

During the first session of the first day (T1S1), the feeders A and D are filled with food (12 mealworms). The shrew can explore all the feeders and get any of the food items. After 30 minutes, the shrew can leave the arena. During the second session of the first day (T1S2), the feeders are not replenished and the arena is in the same "conditions" the shrew left it in (e.g. trail marks, urine marks). The shrew must remember the location they already got food from, or keep exploring to find more. The sessions on the second day are identical: in T2S1, the arena is clean, the feeders are filled and the shrew can explore/find food in the same spots as the first time. In T2S2, the only food available is the one not consumed during the previous trial.

All shrews were recorded in the arena with a security camera. The recordings have been tracked with the tracking software Trex. The results of the first tracking gave the coordinates of the position of the animal at each frame of the recording. Afterwards, each video has been visually analysed to assess the effective visit of the islands: every visit was manually registered with the letter of the island visited, the number corresponding to the door, the presence or absence of food and if the door was opened towards left or right.

Scripts in this repo

1. foraging_results.R enable to clean and tidy the data. Afterwards it merges the tracking data from Trex and the island visit data manually recorded. Finally, some plots to visually assert the result and some checks on the structure of the df.
2. foraging_master.R enable to calculate the variables that will be used for the modelling. Three csv are produced by this code:

   • interactions_count contains the number of interactions with the door of every islands for each shrew. Moreover, a success rate is calculated based on the number of doors visited.

   • foraging_master contains all the information that will be implemented in the models for exploration tendencies and path patterns similarities.

   • foraging_similarities contains the path similarities for each individual among the 4 trials.

3. foraging_edges.R enable to get data about the areas occupied by the animal during the experiment. More specifically it gives the time spent on the edge of the arena.
4. foraging_modelling.R enable to use the data produced in the previous codes and create statistical valuable results. FREQUENTIST STATISTIC ONLY
5. bayesian_models.Rmd enable to use the data produced in the previous codes and create statistical valuable results. BAYESIAN STATISTIC ONLY

Steps to make everything work

1. Run foraging_results.R in order to get the file: foraging_result.csv.

   • foraging_result.csv contains tracking data, including the data regarding the island visited. foraging_result.csv is often indicated as result in the scripts.

2. Run foraging_master.R in order to get the files: foraging_master.csv, foraging_similarities.csv, interactions_counts.csv.

   • foraging_master.csv contains data about explorative features.
   • foraging_similarities.csv contains data about the space usage similarities among trials.
   • interactions_counts.csv contains data about efficiency rate (no models on this).

3. Run foraging_edges in order to get the file: foraging_edges_new.csv. To run this script you need additional foraging_corners.csv, islands.csv.

   • foraging_edges_new.csv contains data about time spent on the edges of the arena.

4. Run bayesian_models.Rmd to perform the models and to investigate the statistical results.