Directional bias in movement while searching for a food reward during spatial cognitive tasks in mountain chickadees (Poecile gambeli)
Data analyzed in 2022. All data were collected as part of a long-term study (2014 - ongoing in 2022) of mountain chickadees at Sagehen Experimental Forest in the Sierra Nevada mountains (Sagehen Creek Field Station, University of California Berkeley), located 10 km north of Truckee, CA, USA. To band birds, birds were trapped annually at nestboxes during summer or at established bird feeders using mistnets in the fall and winter. Trapped birds were banded with color bands including a colored passive integrated transponder (PIT) tag with a unique alphanumeric ID (IB Technology, Leicestershire, U.K.). Sex (male or female) was determined from previous summer breeding survey records using physiological evidence (e.g., brood patch or cloacal protuberance), if possible. See published works for better characterization of banding efforts and cognition experiments: Branch et al. 2016, 2017; Kozlovsky et al. 2018, Tello-Ramos et al. 2018, Sonnenberg et al. 2019, Sonnenberg et al. 2020, Benedict et al. 2021. Directional bias was estimated in two was based on the direction that chickadees moved around an experimental apparatus during foraging-based search in several spatial cognitive tasks. Only trials with at least two location errors were analyzed (i.e. non-rewarding feeders), ensuring that chickadees had not yet learned the location of the correct rewarding feeder for each task. Initial bias was estimated from the relative direction birds moved between the first two location errors. Consistent bias was estimated when birds again moved in the same direction between the second and third visits as during the initial bias. Visits to feeders across the feeder array were not analyzed. The tasks included spatial learning and memory (winter 2019-2020, winter 2020-2021), single spatial reversal learning (winter 2019-2020, winter 2020-2021), and serial spatial reversal learning (winter 2020-2021). Data were collected automatically from 'smart' feeders equipped with a radio frequency identification (RFID) data logger connected to a perch-mounted antenna that detected and logged the PIT tags of birds that landed at the feeders. Each feeder had a mechanized door to control access to a supply of black oil sunflower seeds. Feeders were arranged in 8-feeder arrays on a square frame, 2 per elevation, ca. 1.2 km apart (based on established protocols, see Croston et al. 2016, 2017). This file contains the season of data collection, type of bias, individual bird ID, elevation (high or low), sex, total number of left-scored trials, total number of right-scored trials, total number of scored trials, bias index (calculated using a standard lateralization index), and several statistics for significant bias. These included a p-value from binomial tests (per bird), whether the resulting binomial p-value was significant using an alpha = .05 and bonferroni corrected alpha (.05/n birds) the Holm-adjusted p-value, the z-score, and z-score based bias (to estimate relative bias).
Steps to reproduce
Calculated metrics: (1) directional bias index = (left - right) / (left + right ) (2) z-score = (leftTrials - (totalTrials/2)) / (sqrt(totalTrials/4) ] (3) Z.bias is "relatively biased" to the left (L) or right (R) if z-score is less than -1.96 or greater than +1.96. Otherwise indicated as "A" (no relative bias). All analyses performed using R in 2022. Please contact Lauren Benedict for further data or R scripts: firstname.lastname@example.org