Abstract
Accurate goal-directed behavior requires the sense of touch to be integrated with information about body position and ongoing motion1,2,3. Behaviors like chewing, swallowing and speech critically depend on precise tactile events on a rapidly moving tongue4,5, but neural circuits for dynamic touch-guided tongue control are unknown. Using high speed videography, we examined 3D lingual kinematics as mice drank from a water spout that unexpectedly changed position during licking, requiring re-aiming in response to subtle contact events on the left, center or right surface of the tongue. Mice integrated information about both precise touch events and tongue position to re-aim ensuing licks. Surprisingly, touch-guided re-aiming was unaffected by photoinactivation of tongue sensory, premotor and motor cortices, but was impaired by photoinactivation of the lateral superior colliculus (latSC). Electrophysiological recordings identified latSC neurons with mechanosensory receptive fields for precise touch events that were anchored in tongue-centered, head-centered or conjunctive reference frames. Notably, latSC neurons also encoded tongue position before contact, information important for tongue-to-head based coordinate transformations underlying accurate touch-guided aiming. Viral tracing revealed tongue sensory inputs to the latSC from the lingual trigeminal nucleus, and optical microstimulation in the latSC revealed a topographic map for aiming licks. These findings demonstrate for the first time that touch-guided tongue control relies on a collicular mechanosensorimotor map, analogous to collicular visuomotor maps associated with visually-guided orienting across many species.
Competing Interest Statement
The authors have declared no competing interest.
Footnotes
Extended Data Fig. 1 added, showing that water dispensation was not required for touch-guided re-aiming. Extended Data Fig. 3 added, showing electrophysiological validation of methods used for cortical and latSC photoinhibition. Extended Data Fig. 4 added, showing that touch-guided re-aiming was intact with lesions of all three lingual cortical areas at once. Extended Data Fig. 6d added, showing that many single latSC neurons diverge within 10 - 15 ms relative to contact. Extended Data Fig. 8h added, showing that latSC neurons with contact responses in a head-centric reference frame have longer divergence latencies than neurons with tongue-centric or conjunctive reference frames.
