Listening to bird dreams

In Christopher Nolan's iconic 2010 film Inception, the protagonist infiltrates people's dreams to extract vital information. A team of researchers has similarly infiltrated the dreams of birds and extracted the songs they silently replay in their sleep (bit.ly/Avian-dreams). The researchers achieved this by establishing a relationship among vocal muscle movement, neural activity and song production. About 40% of the bird species learn to vocalise, says Gabriel Mindlin, a physicist at the University of Buenos Aires, Argentina, who has been studying the physics of birdsong for two decades. "So, it's a good animal model to study how the brain reconfigures itself during learning."

Earlier, Mindlin's team had noticed that a bird's vocal muscles moved in a similar way during daytime singing and while the birds slept. Based on that, the researchers created a mathematical model. In the current study, they used a technique called electromyography (EMG) to record the electrical signals from vocal muscles of the Great Kiskadee (Pitangus sulphuratus) by implanting microelectrodes. Then, they fed this pattern information into their mathematical model, which transformed the patterns into synthetic songs.

Mindlin's mathematical model is complex and includes parameters such as air pressure and muscle movement. The model can predict vocal learning in a bird species based on these parameters. In birds, the vocal organ that controls airflow is called the syrinx. The study shows that the EMG activity recorded in syringeal muscles obeys the orders of the central nervous system, meaning much of the translation – from neural signal to vocal muscle movement – is already done in the body. Further, it reveals how the muscles change these signals into specific actions and how this affects bird behaviour.

Previous studies have shown that the brain circuits of birds involved in song learning are similar to those of speech or language learning in humans. However, it is still unknown why birds replay songs in their sleep.

The recreated song patterns show inherent similarity with the original vocalisations, "providing a unique window to the neural networks of the avian brain", says Shankha Sanyal, a senior researcher at the Sir C.V. Raman Centre for Physics and Music, Jadavpur University, Kolkata, who was not involved in the study.