Hidden neural states underlie canary song syntax

Coordinated skills such as speech or dance involve sequences of actions that follow syntactic rules in which transitions between elements depend on past actions. Canary songs are comprised of repeated syllables, called phrases, and the ordering of these phrases follows long-range rules, where the choice of what to sing depends on song structure many seconds prior. The neural substrates that support these long-range correlations are unknown. Using miniature head-mounted microscopes and cell-type-specific genetic tools, we observed neural activity in the premotor nucleus HVC as canaries explore various phrase sequences in their repertoire. We find neurons that encode past transitions, extending over 4 phrases and spanning up to 3 seconds and 40 syllables. These neurons preferentially encode past actions rather than future actions, can reflect more than a single song history, and occur mostly during the rare phrases that involve history-dependent transitions in song. These findings demonstrate that network dynamics in HVC reflect preceding behavior context relevant to flexible transitions.

[1]  J. F. Prather,et al.  Precise auditory–vocal mirroring in neurons for learned vocal communication , 2008, Nature.

[2]  Richard Hans Robert Hahnloser,et al.  Spike-Time-Dependent Plasticity and Heterosynaptic Competition Organize Networks to Produce Long Scale-Free Sequences of Neural Activity , 2010, Neuron.

[3]  Richard Hans Robert Hahnloser,et al.  Neural Mechanisms of Vocal Sequence Generation in the Songbird , 2004, Annals of the New York Academy of Sciences.

[4]  Jürgen Schmidhuber,et al.  Learning to Forget: Continual Prediction with LSTM , 2000, Neural Computation.

[5]  Timothy J Gardner,et al.  An open source, wireless capable miniature microscope system , 2017, Journal of neural engineering.

[6]  Yoshua Bengio,et al.  Attention-Based Models for Speech Recognition , 2015, NIPS.

[7]  Kristofer E. Bouchard,et al.  An Adapting Auditory-motor Feedback Loop Can Contribute to Generating Vocal Repetition , 2015, PLoS Comput. Biol..

[8]  Dai Watanabe,et al.  Neural Coding of Syntactic Structure in Learned Vocalizations in the Songbird , 2011, The Journal of Neuroscience.

[9]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[10]  Dezhe Z. Jin,et al.  Support for a synaptic chain model of neuronal sequence generation , 2010, Nature.

[12]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[13]  L. Nathan Perkins,et al.  Unstable neurons underlie a stable learned behavior , 2016, Nature Neuroscience.

[14]  Nancy Kopell,et al.  Neural Sequence Generation Using Spatiotemporal Patterns of Inhibition , 2015, PLoS Comput. Biol..

[15]  Michale S Fee,et al.  Miniature motorized microdrive and commutator system for chronic neural recording in small animals , 2001, Journal of Neuroscience Methods.

[16]  D. Jin Generating variable birdsong syllable sequences with branching chain networks in avian premotor nucleus HVC. , 2009, Physical review. E, Statistical, nonlinear, and soft matter physics.

[17]  K. D. Punta,et al.  An ultra-sparse code underlies the generation of neural sequences in a songbird , 2002 .

[18]  Liam Paninski,et al.  Efficient and accurate extraction of in vivo calcium signals from microendoscopic video data , 2016, eLife.

[19]  Dezhe Z. Jin,et al.  A Compact Statistical Model of the Song Syntax in Bengalese Finch , 2010, PLoS Comput. Biol..

[20]  Robin C. Ashmore,et al.  Brainstem and Forebrain Contributions to the Generation of Learned Motor Behaviors for Song , 2005, The Journal of Neuroscience.

[21]  Yonatan Sanz Perl,et al.  Elemental gesture dynamics are encoded by song premotor cortical neurons , 2013, Nature.

[22]  Timothy Q. Gentner,et al.  Working memory for patterned sequences of auditory objects in a songbird , 2010, Cognition.

[23]  Mengru Li,et al.  Stable propagation of a burst through a one-dimensional homogeneous excitatory chain model of songbird nucleus HVC. , 2006, Physical review. E, Statistical, nonlinear, and soft matter physics.

[24]  Tara N. Sainath,et al.  Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups , 2012, IEEE Signal Processing Magazine.

[25]  R. Mooney,et al.  Calcium‐binding proteins define interneurons in HVC of the zebra finch (Taeniopygia guttata) , 2005, The Journal of comparative neurology.

[26]  Georg B. Keller,et al.  Rapid Interhemispheric Switching during Vocal Production in a Songbird , 2008, PLoS biology.

[27]  Todd W Troyer,et al.  Development of temporal structure in zebra finch song. , 2013, Journal of neurophysiology.

[28]  A. Georgopoulos,et al.  Motor cortical encoding of serial order in a context-recall task. , 1999, Science.

[29]  Marc F. Schmidt,et al.  What birdsong can teach us about the central noradrenergic system , 2010, Journal of Chemical Neuroanatomy.

[30]  Jun Tani,et al.  Emergence of Functional Hierarchy in a Multiple Timescale Neural Network Model: A Humanoid Robot Experiment , 2008, PLoS Comput. Biol..

[31]  Kenichi Ohki,et al.  Conversion of Working Memory to Motor Sequence in the Monkey Premotor Cortex , 2003, Science.

[32]  Jessica A. Cardin,et al.  Noradrenergic Inputs Mediate State Dependence of Auditory Responses in the Avian Song System , 2022 .

[33]  M. Fee,et al.  Using temperature to analyze temporal dynamics in the songbird motor pathway , 2008, Nature.

[34]  Robert E. Lemon,et al.  Bird Song: An Acoustic Flag , 1977 .

[35]  Katharina Riebel,et al.  Light Flash Stimulation Alters the Nightingale's Singing Style: Implications for Song Control Mechanisms , 1997 .

[36]  F. Nottebohm,et al.  Central control of song in the canary, Serinus canarius , 1976, The Journal of comparative neurology.

[37]  F. Nottebohm,et al.  The telencephalon, diencephalon, and mesencephalon of the canary, Serinus canaria, in stereotaxic coordinates , 1974, The Journal of comparative neurology.

[38]  David Pfau,et al.  Simultaneous Denoising, Deconvolution, and Demixing of Calcium Imaging Data , 2016, Neuron.

[39]  Jeremy Hyman Countersinging as a signal of aggression in a territorial songbird , 2003, Animal Behaviour.

[40]  Razvan Pascanu,et al.  Relational recurrent neural networks , 2018, NeurIPS.

[41]  Michael S Brainard,et al.  Auditory-induced neural dynamics in sensory-motor circuitry predict learned temporal and sequential statistics of birdsong , 2016, Proceedings of the National Academy of Sciences.

[42]  D Margoliash,et al.  Behavioral state modulation of auditory activity in a vocal motor system. , 1998, Science.

[43]  D. Todt,et al.  Song acquisition and acquisition constraints in the nightingale, Luscinia megarhynchos , 1989, Naturwissenschaften.

[44]  Moshe Abeles,et al.  Corticonics: Neural Circuits of Cerebral Cortex , 1991 .

[45]  Michale S Fee,et al.  Singing-related neural activity distinguishes four classes of putative striatal neurons in the songbird basal ganglia. , 2010, Journal of neurophysiology.

[46]  Timothy J. Gardner,et al.  Long-range Order in Canary Song , 2013, PLoS Comput. Biol..

[47]  Franz Goller,et al.  A circular model for song motor control in Serinus canaria , 2015, Front. Comput. Neurosci..

[48]  Kosuke Hamaguchi,et al.  A Distributed Recurrent Network Contributes to Temporally Precise Vocalizations , 2016, Neuron.

[49]  M. Fee,et al.  A hypothesis for basal ganglia-dependent reinforcement learning in the songbird , 2011, Neuroscience.

[50]  Stefan R. Pulver,et al.  Ultra-sensitive fluorescent proteins for imaging neuronal activity , 2013, Nature.

[51]  Ann M. Graybiel,et al.  Neurobiology of rodent self-grooming and its value for translational neuroscience , 2016, Nature Reviews Neuroscience.

[52]  K. Lorenz,et al.  Taxis und Instinkthandlung in der Eirollbewegung der Graugans. I1 , 2010 .

[53]  Annemarie van der Linden,et al.  A customizable 3-dimensional digital atlas of the canary brain in multiple modalities , 2011, NeuroImage.