The results highlight the potential for a “bottom-up” approach to drive discovery in voice communication
Researchers have developed a new algorithm capable of identifying the characteristics of male zebra songs that may underlie the distinction between a short phrase sung during courtship and the same phrase sung in a non-flirting context. Sarah Woolley of McGill University in Montreal, Canada, and colleagues presented these findings in the journal Open Access PLOS Computational Biology.
Like many animals, male zebrafish tunes in to their audiences. They may sing the same sequence of syllables during courtship interactions with females as they do singing alone, albeit with slight modifications. However, humans cannot detect these differences, and it was not clear if female zebra finches could either.
For the new study, Woolley and colleagues conducted behavioral experiments for the first time showing that female zebras are indeed adept at distinguishing short passages of male songs recorded in conditions of courtship versus non-courtship.
Next, they sought to expand on previous studies that focused on a few of the song’s specific features that might underlie the distinction between courtship and non-courtship song. Using a “bottom-up” approach, the researchers extracted more than 5,000 song features from the recordings and trained an algorithm to use these features to distinguish between expressions of courtship and non-flirtation.
The trained algorithm revealed features that may be fundamental to song perception, some of which were not previously identified. It also made predictions about the discrimination abilities of female zebra finches that align well with the results of the behavioral experiments.
These results highlight the potential of bottom-up approaches to uncover acoustic features important to communication and social distinction.
“As voice communicators ourselves, we tend to focus on the aspects of communication signals that are salient to us,” Woolley says. “Using our bottom-up approach, we identified features that might not have been on our radar.”
Next, the researchers plan to test whether tampering with the vocal features they discovered alters female birds’ view of those songs. They also hope to evaluate the generalizability of their findings to flirting and non-flirting songs in other genres.
Peer review simulation / modeling; Pilot study; the animals
In your coverage, please use this URL to provide access to the article freely available in PLOS Computational Biology:
Citation: Paul A, McLendon H, Rally V, Sakata JT, Woolley SC (2021) Behavioral discrimination and phenotypic modeling of sparrows’ performance time series. Plus Compute Biol 17 (4): e1008820. https: /
Funding: This work was funded through grants from the National Council of Science and Engineering of Canada (RGPIN 2018-05267 to SCW; RGPIN2015016 to JTS; https: /
Competing interests: The authors state that there are no competing interests.
PLOS Computational Biology
[email protected]http: // dx.