Experiment predicts syllables birds will sing based on brain signals

Researchers from the University of California San Diego have been conducting experiments involving birds and their brains. The team conducted experiments that predicted syllables a bird would sing and when they would sing them solely by reading electrical signals from its brain. The team believes gaining the ability to predict a bird's behavior based solely on brain activity is the next step towards creating a vocal prosthetic for humans.

Creating that vocal prosthesis is the main goal in the research, according to the first author of the study Daril Brown. According to Brown, the team is studying birdsong as a way to take them closer to engineering a human brain-machine interface allowing vocalization and communication. Researchers used a songbird called the zebra finch in their experiment.

In the study, researchers explored how brain activity in songbirds might be used to forecast what the bird would sing. As for why the team is focusing on songbirds, they say the vocalizations are of particular interest because they are similar to human speech in that it's a complex and learned behavior. For the study, male adult zebra finches had silicon electrodes implanted into their brains, allowing the researchers to study neural activity during singing.

Signals the researchers were investigating are called local field potentials and are recorded in sections of the bird's brain used for learning and singing. Similar local field potentials are also used to predict vocal behavior in humans.

In the longer term, researchers want to use the knowledge gained during their experiments on the brains of songbirds to improve the quality of life for humans who suffer from various disorders and illnesses impacting speech. Researchers identified different features of local field potentials that translate into specific syllables of the song the bird was singing. The team is also able to determine when the syllables would occur during a song. One of the more interesting discoveries was the ability to predict variations in song sequence on a syllable level.