We investigate how the complex social dynamics between individuals change over time utilizing several behavioral sequential methods including Markov Chain and Lag Sequential Analysis.
For example, the image on the right illustrates the variability in how the behavior of pairs of mice changes over a 60 second period in a social interaction test. Pairs of mice are on rows 1-2, 3-4, 5-6, etc. A key objective of our research is to develop statistical methods to identify how the structure of behavior within relationships changes over time.
By coding behavioral event data over time we are able to identify patterns of related behaviors and which behaviors are most likely to follow others. By identifying such patterns we can determine variability in and perturbations from typical social behavior. For instance, the image on the right illustrates which behaviors are more likely to occur than chance after specific behaviors (yellow/orange/red) versus those that are less likely than chance to occur after specific behaviors (blue).
For further information on these projects please get in touch.
Prof Curley has developed an R package behavseq that can be used in the analysis of sequential behavioral patterns.