Self-timed movement initiation requires rapid sequential coordination of circuits in prefrontal cortex and cerebellum

Technical Summary: Self-timed movement initiation requires rapid sequential coordination of circuits in prefrontal cortex and cerebellum

Problem and Motivation
The neural mechanisms governing the precise timing of self-initiated (endogenous) movements remain poorly understood compared to externally triggered actions. While externally driven movements rely on a known cascade involving reticular and pedunculopontine neurons projecting to the motor cortex, the millisecond-scale coordination required for internally driven actions—where no external "go" cue is present—is less defined. Although cortico-basal ganglia networks are established as key players in self-paced action, the specific causal role of cortico-cerebellar circuits in generating the internal "go" signal and executing the movement with high temporal precision is unclear. This study aims to dissect the causal contributions of the dorsomedial prefrontal cortex (dmPFC) and the lateral cerebellum (specifically the dentate nucleus) to the initiation of self-timed movements.

Methodology
The authors developed a novel behavioral task for mice called the "Omitted Oddball" (OO) task. In this paradigm, mice are exposed to a periodic train of LED flashes (e.g., every 300 ms) followed by an aversive airpuff. Unlike classical conditioning, the "go" signal is not an external cue but an internal prediction: mice must learn the tempo of the flashes, predict the next occurrence, and initiate a protective eyeblink when the expected flash is omitted. The airpuff is delivered 200 ms after the predicted time of the omitted stimulus.

To probe the causal roles of specific brain regions, the authors employed high-temporal-resolution optogenetic inhibition ("opto-blinks"). They used:

1. L7-ChETA mice: Expressing channelrhodopsin in inhibitory Purkinje cells to suppress activity in the dentate nucleus (DN) of the cerebellum via axon terminal stimulation.
2. PV-IRES-Cre x Ai32 mice: Expressing channelrhodopsin in parvalbumin interneurons to suppress output neurons in the dmPFC.

Brief laser pulses (35 ms) were delivered at various time points relative to the stimulus omission. To prevent the laser light itself from acting as a confounding sensory cue, mice were trained with a randomly intermittent background blue light to mask the optogenetic laser pulses. Control tasks included standard trace conditioning and "trace offset" conditioning to ensure mice were not simply reacting to stimulus offsets but were detecting the omission.

Key Results

1. Internal Timing Mechanism: Mice successfully learned the OO task, initiating eyeblinks with high temporal precision relative to the omission of the stimulus, rather than the offset of the last flash. Movement latencies were consistent across different stimulus intervals, confirming that mice rely on internal predictions of the periodic rhythm.
2. Role of the Dentate Nucleus (DN): Brief optogenetic inhibition of the ipsilateral DN caused a delay in movement initiation, but did not abolish the movement. The delay was maximal when inhibition occurred 70 ms after the stimulus omission. Crucially, the inhibition window (65–75 ms) revealed that DN activity is necessary for the execution of the movement trigger but that the upstream drive persists during inhibition, allowing the movement to occur once the inhibition ceases.
3. Role of the dmPFC: In contrast to the DN, brief inhibition of the dmPFC did not delay movement latency; instead, it profoundly reduced the probability and amplitude of the movement. This effect was time-locked to a narrow window (approx. 100 ms) surrounding the moment of stimulus omission. Inhibition outside this window had no effect. This indicates that dmPFC activity is causally essential for generating the internal "go" signal but is not continuously required for movement preparation or execution.
4. Sequential Circuit Model: The data support a model where an internal "go" signal is generated in the dmPFC at the time of stimulus omission. This signal is transmitted to the lateral cerebellum, where it triggers a ramping of activity in the dentate nucleus. When this activity reaches a threshold, the movement is initiated.

Significance and Claims
The paper claims to provide a causal, circuit-level model for self-timed movement initiation. The authors argue that the OO task reveals a distinct mechanism from externally cued movements:

• Sequential Coordination: Movement initiation involves two sequential, computationally distinct steps: the dmPFC generates the internal "go" signal, and the lateral cerebellum converts this signal into a precisely timed motor trigger.
• Cortico-Cerebellar Specificity: While basal ganglia circuits are known to contribute to self-timing, this study highlights that cortico-cerebellar pathways are critical when the brain must transform a temporal prediction error (the omission) into a precisely timed, self-initiated action.
• Mechanistic Distinction: The study distinguishes between the generation of the decision to move (dmPFC) and the precise timing of the motor command (cerebellum). The dmPFC signal is transient and necessary for initiation, whereas the cerebellum acts as the "starting gun" that executes the movement, capable of ramping toward a threshold even if briefly interrupted.

The authors conclude that this rapid sequential recruitment of the dmPFC and lateral cerebellum is the fundamental mechanism allowing for the millisecond-precision initiation of endogenous actions in the absence of external cues.