Hidden neural states underlie canary song syntax

Area of Science:

• Neuroscience
• Animal Behavior
• Bioacoustics

Background:

• Complex vocalizations like human speech and birdsong exhibit syntactic rules.
• Canary songs demonstrate long-range dependencies, where phrase selection depends on past song structure.
• The neural mechanisms underlying these long-range syntactic rules remain largely unknown.

Purpose of the Study:

• To investigate the neural substrates responsible for encoding long-range sequential dependencies in canary song.
• To identify specific neurons and their activity patterns related to song syntax.

Main Methods:

• Utilized miniature head-mounted microscopes to record neural activity in the premotor nucleus HVC of canaries.
• Employed cell-type-specific genetic tools for targeted neural observation.
• Analyzed neural activity during the exploration of various phrase sequences in the canary's repertoire.

Main Results:

• Identified neurons in HVC that encode past transitions in canary song, spanning up to four phrases and four seconds.
• These neurons preferentially encode past actions over future actions.
• Observed that these neurons reflect multiple song histories and are active during history-dependent transitions.

Conclusions:

• The dynamics within HVC involve 'hidden states' that carry information about prior actions, not just current behavior.
• These findings suggest a neural substrate for the control of syntax transitions governed by long-range rules in vocal learning.