Scientists at UC San Francisco brought a new understanding of how experience and genetics interact to affect complex learning.
Published in the Jan. 9, 2018 issue of PNAS, the report showed that male songbirds largely inherit the song features of their genetic fathers, but learning from a foster father can override that genetic predisposition.
Song learning in birds is a useful model system that mirrors many components of human language learning. We already know that both experience and genetics affect learning. But how they interact to shape complex learned behaviors remains a mystery.
David Mets, a research specialist, and Michael Brainard, a Professor of Physiology and Psychiatry, compared the song tempos of the genetically diverse Bengalese finches (Lonchura striata domestica). “We chose to measure song tempo because it is easily quantified, behaviorally relevant, and varies across individuals, so likely to be centrally controlled,” said Brainard.
To understand the genetic contribution to individual variations, Mets collected eggs within 36 hours of laying. After hatching, the young birds grew up in a sound-isolated environment to control their sound exposure. Female birds, who do not sing, raised young finches.
Male finches normally learn their songs from their fathers. In the first set of experiments, computer-generated synthetic songs of various tempos did the song tutoring. These finches developed a broad range of tempos that strongly correlated with their genetic father’s songs. Using Multiple Linear Regression to access the contributions of genetics and experiences, Mets and Brainard found strong genetic influence (55%) and weak experiential influence (21%) in shaping the song tempo of the young finches from computer tutoring. So, this group of finches developed songs largely like their fathers’, which they had never heard.
Like birds, humans learn vocal patterns during early development.
In a second set of experiments, the researchers replaced the computer-generated synthetic tutor with finch foster fathers. “Then we wanted to see what happened if the instruction was more enriched,” said Mets. “Unlike songs played by a computer, songs of a daddy bird give a much more diverse acoustic experience in a more natural setting.”
In these set of experiments, Mets again removed the eggs from their original nests. Instead of computers, unrelated live adult foster parents taught the young bird to sing. This time, Mets and Brainard saw a complete reversal of the relative influences of genetics and experience with live tutoring. Multiple regression, strikingly, showed a dramatic shift in the influence that heavily favors experiential (53%) over genetics (16%).
“We were really excited to see that nurture plays a strong hand in molding the heritable complexity of learned behaviors,” said Mets. “What we found also supports higher quality in live tautology over sterile in silico instructions.”
Like birds, humans learn vocal patterns during early development. The result of this study has broader implications in understanding human speech learning, where heritable traits may similarly depend on the attribute of experience.