23 November 2013
In October, the National Institute for Mathematical and Biological Synthesis (NIMBioS) held an investigative workshop on Analyzing Animal Vocal Communication Sequences. The workshop was organized by Arik Kershenbaum (NIMBioS), Marie Roch (SDSU), and Dan Blumstein (UCLA) with the aim "to bridge the gap between mathematical and biological researchers with an interest in the quantitative analysis of animal vocal sequences." Through a series of talks and informal breakout sessions, the workshop was focused on how to define "the state of the art in this field, explore new horizons for collaboration, and provide new techniques through a synthesis of the mathematical and biological approaches to communication analysis."
The talk series at the workshop encompassed a broad range of taxonomic groups. While many of the species highlighted at the workshop have intricate and complex vocalization systems that can give us insights into their own intelligence and behavior, these animals can also potentially provide information about our own communication system. Some of the featured animals show vocal learning (songbird, dolphin), have grammatical structure (songbird, whale, bat), have similar functional neural pathways (songbird, bat), are related genetically (primates), and live complex social networks (mongoose, meerkat, dolphin, primates, elephant).
The talk series also addressed the wealth of methods that exist to analyze the vocal patterns of these animals, and consequently addressed many of the key issues in the field. First, in order to determine the content and ordering of the sequences, researchers must determine the units that comprise their analysis. These units may be a set of sounds that are perceptually distinguishable to the animals, are producible by the animals, or have functional significance when altered. These units have in the past been obtained through observation, playback studies, or clustering methods. Second, given the pieces, researchers want to understand how the pieces are recombined into different patterns. Typically this has been done through the use of graphical models that show the types of sounds made and the probabilities of moving from one sound to the other. Finally, researchers want to know the meanings, or the functional use, of these communication sequences. One method is to collect a large corpus of data including both the environmental context and the affect of the animals in order to look for statistical patterns that predict certain vocalizations. Human observation also typically provides a lot of the analysis concerning the context and researchers often have hints as to the significance of a vocal pattern. They can test their hypotheses through playback experiments and see if manipulations in the pieces or the structure give rise to predictable behavioral patterns. If there is a hypothesis about the possible meaning of a sequence, researchers can also quantify this by using information theory to determine how reliable the messages are and how much information is being transmitted.
Here is a quick overview of the speakers, their animals of interest, and the techniques being used (if applicable):
Manser, Marta (Univ. Zurich) - Mongoose and Meerkat
Bohn, Kirsten (Florida International Univ.) - Bats
Janik, Vincent (Univ. St. Andrews) - Dolphins
Sayigh, Laela (Woods Hole Oceanographic Inst.) - Dolphins, Clustering
Vallejo, Edgar (Monterrey Tech., Mexico) - Birds, Hierarchical Clustering
Johnson, Michael (Marquette) - Birds and Elephants, HMM Clustering
McCowan, Brenda (UC Davis) - Dolphins
Doyle, Laurance (SETI Institute) - Dolphins, Information Theory
Ferrer-i-Cancho, Ramon (Tech. Univ. Catalunya) - Information Theory
Bee, Mark (Univ. Minnesota) - Frogs
Casar, Cris (Univ. St. Andrews) - Primates
Coen, Michael (Univ. Wisconsin) - Gibbons
Harley, Heidi (New University, FL) - Dolphins
Edelman, Shimon (Cornell Univ.) - Graphical models of syntax
Jin, Dezhe (Pennsylvania Stat Univ.) - Birds
For more information on NIMBioS and other workshops being offered, please visit http://nimbios.org.
Related tags: animal communication, NIMBioS