Inferring social structure from temporal data

Psorakis, Ioannis; Voelkl, Bernhard; Garroway, Colin J.; Radersma, Reinder; Aplin, Lucy M.; Crates, Ross; Culina, Antica; Farine, Dan; Firth, Josh A.; Hinde, Camilla A.; Kidd, Lisa; Milligan, Nicole D.; Roberts, Stephen; Verhelst, Brecht; Sheldon, Ben C.

doi:10.1007/s00265-015-1906-0

Cited by 94 publications

(117 citation statements)

References 24 publications

Supporting

Mentioning

117

Contrasting

Order By: Relevance

“…We recorded data from 1 November 2014 to 5 January 2015 and 29 January to 31 March 2015, and all sites had a minimum of 13 weeks of collection (range: 13–18). Social associations were determined from temporal data collected from marked individuals using Gaussian mixture models (GMMs; Psorakis, Roberts, Rezek, & Sheldon, ; Psorakis et al, ; Evans, Jones, & Morand‐Ferron, ). GMM detects “bursts” of increased activity, to which individual events are assigned (Farine et al, ; Psorakis et al, , ).…”

Section: Methodsmentioning

confidence: 99%

Urbanization and the temporal patterns of social networks and group foraging behaviors

Jones

Evans

Morand‐Ferron

2019

Ecology and Evolution

View full text Add to dashboard Cite

Urbanization causes dramatic and rapid changes to natural environments, which can lead the animals inhabiting these habitats to adjust their behavioral responses. For social animals, urbanized environments may alter group social dynamics through modification of the external environment (e.g., resource distribution). This might lead to changes in how individuals associate or engage in group behaviors, which could alter the stability and characteristics of social groups. However, the potential impacts of urban habitat use, and of habitat characteristics in general, on the nature and stability of social associations remain poorly understood. Here, we quantify social networks and dynamics of group foraging behaviors of black‐capped chickadees ( N = 82, Poecile atricapillus ), at four urban and four rural sites weekly throughout the nonbreeding season using feeders with radio frequency identification of individual birds. Because anthropogenic food sources in urban habitats (e.g., bird feeders) provide abundant and reliable resources, we predicted that social foraging associations may be of less value in urban groups, and thus would be less consistent than in rural groups. Additionally, decreased variability of food resources in urban habitats could lead to more predictable foraging patterns (group size, foraging duration, and the distribution of foraging events) in contrast to rural habitats. Networks were found to be highly consistent through time in both urban and rural habitats. No significant difference was found in the temporal clumping of foraging events between habitats. However, as predicted, the repeatability of the clumping of foraging events in time was significantly higher in urban than rural habitats. Our results suggest that individuals living in urban areas have more consistent foraging behaviors throughout the nonbreeding season, whereas rural individuals adjust their tactics due to less predictable foraging conditions. This first examination of habitat‐related differences in the characteristics and consistency of social networks along an urbanization gradient suggests that anthropic habitat use results in subtle modifications in social foraging patterns. Future studies should examine potential implications of these differences for variation in predation risk, energy intake, and information flow.

show abstract

Section: Methodsmentioning

confidence: 99%

Urbanization and the temporal patterns of social networks and group foraging behaviors

Jones

Evans

Morand‐Ferron

2019

Ecology and Evolution

View full text Add to dashboard Cite

show abstract

“…The methodology behind GMM inference of animal social structure and how this mitigates the potential bias associated with choosing a sampling period is already discussed in depth by Psorakis and colleagues [5,6]. Equally, there are a growing number of studies using this same model system (i.e.…”

Section: Introductionmentioning

confidence: 99%

Inferring animal social networks and leadership: applications for passive monitoring arrays

Jacoby

Papastamatiou

Freeman

2016

J. R. Soc. Interface.

View full text Add to dashboard Cite

Analyses of animal social networks have frequently benefited from techniques derived from other disciplines. Recently, machine learning algorithms have been adopted to infer social associations from time-series data gathered using remote, telemetry systems situated at provisioning sites. We adapt and modify existing inference methods to reveal the underlying social structure of wide-ranging marine predators moving through spatial arrays of passive acoustic receivers. From six months of tracking data for grey reef sharks (Carcharhinus amblyrhynchos) at Palmyra atoll in the Pacific Ocean, we demonstrate that some individuals emerge as leaders within the population and that this behavioural coordination is predicted by both sex and the duration of co-occurrences between conspecifics. In doing so, we provide the first evidence of long-term, spatially extensive social processes in wild sharks. To achieve these results, we interrogate simulated and real tracking data with the explicit purpose of drawing attention to the key considerations in the use and interpretation of inference methods and their impact on resultant social structure. We provide a modified translation of the GMMEvents method for R, including new analyses quantifying the directionality and duration of social events with the aim of encouraging the careful use of these methods more widely in less tractable social animal systems but where passive telemetry is already widespread.

show abstract

“…The Gaussian mixture models identify instances of individuals visiting the same feeding station close in time, which are defined as “gathering events” (Psorakis, Roberts, Rezek, & Sheldon, ). Those gathering events serve as snapshots of the composition of foraging flocks and have been demonstrated to outperform other flock‐detection methods (Psorakis et al., ). We used the movements of individuals between gathering events to quantify fission–fusion dynamics across the three winters see Fig.…”

Section: Methodsmentioning

confidence: 99%

Social and spatial effects on genetic variation between foraging flocks in a wild bird population

et al. 2017

Self Cite

View full text Add to dashboard Cite

Social interactions are rarely random. In some instances, animals exhibit homophily or heterophily, the tendency to interact with similar or dissimilar conspecifics, respectively. Genetic homophily and heterophily influence the evolutionary dynamics of populations, because they potentially affect sexual and social selection. Here, we investigate the link between social interactions and allele frequencies in foraging flocks of great tits (Parus major) over three consecutive years. We constructed co-occurrence networks which explicitly described the splitting and merging of 85,602 flocks through time (fission-fusion dynamics), at 60 feeding sites. Of the 1,711 birds in those flocks, we genotyped 962 individuals at 4,701 autosomal single nucleotide polymorphisms (SNPs). By combining genomewide genotyping with repeated field observations of the same individuals, we were able to investigate links between social structure and allele frequencies at a much finer scale than was previously possible. We explicitly accounted for potential spatial effects underlying genetic structure at the population level. We modelled social structure and spatial configuration of great tit fission-fusion dynamics with eigenvector maps. Variance partitioning revealed that allele frequencies were strongly affected by group fidelity (explaining 27%-45% of variance) as individuals tended to maintain associations with the same conspecifics. These conspecifics were genetically more dissimilar than expected, shown by genomewide heterophily for pure social (i.e., space-independent) grouping preferences. Genomewide homophily was linked to spatial configuration, indicating spatial segregation of genotypes. We did not find evidence for homophily or heterophily for putative socially relevant candidate genes or any other SNP markers. Together, these results demonstrate the importance of distinguishing social and spatial processes in determining population structure.

show abstract

Inferring social structure from temporal data

Cited by 94 publications

References 24 publications

Urbanization and the temporal patterns of social networks and group foraging behaviors

Urbanization and the temporal patterns of social networks and group foraging behaviors

Inferring animal social networks and leadership: applications for passive monitoring arrays

Social and spatial effects on genetic variation between foraging flocks in a wild bird population

Contact Info

Product

Resources

About