SummaryNeuronal activity in primary motor cortex (M1) correlates with behavioral state, but the cellular mechanisms underpinning behavioral state-dependent modulation of M1 output remain largely unresolved. Here, we performed in vivo patch-clamp recordings from layer 5B (L5B) pyramidal neurons in awake mice during quiet wakefulness and self-paced, voluntary movement. We show that L5B output neurons display bidirectional (i.e., enhanced or suppressed) firing rate changes during movement, mediated via two opposing subthreshold mechanisms: (1) a global decrease in membrane potential variability that reduced L5B firing rates (L5Bsuppressed neurons), and (2) a coincident noradrenaline-mediated increase in excitatory drive to a subpopulation of L5B neurons (L5Benhanced neurons) that elevated firing rates. Blocking noradrenergic receptors in forelimb M1 abolished the bidirectional modulation of M1 output during movement and selectively impaired contralateral forelimb motor coordination. Together, our results provide a mechanism for how noradrenergic neuromodulation and network-driven input changes bidirectionally modulate M1 output during motor behavior.

Highlights d DN/IPN thalamocortical activity conveys a reliable feedforward motor timing signal d Silencing DN/IPN or recipient regions of thalamus blocks movement initiation d Photostimulation of the DN/IPN thalamocortical pathway triggers movement d Thalamocortical activation drives behavioral contextdependent movement initiation

Tel. +44 131 650 3113 53 Email: Ian.Duguid@ed.ac.uk 54 65 Stimulating the MThDN/IPN thalamocortical pathway in the absence of the cue recapitulated cue-66 evoked M1 membrane potential dynamics and forelimb behavior in the learned behavioral 67 context, but generated semi-random movements in an altered behavioral context. Thus, 68 cerebellar-recipient motor thalamocortical input to M1 is indispensable for the generation of 69 motor commands that initiate goal-directed movement, refining our understanding of how the 70 cerebellar-thalamocortical pathway contributes to movement timing. 71 72 129 ( Figure 1c and Video S1). Mice rapidly learned to execute the task (mean = 7.5 days, 95% CI 130 [6.3, 8.6], N = 16 mice, all data, unless otherwise stated, are presented as mean, 131 [bootstrapped 95% confidence interval]; last session task success, mean = 0.64 rewards per 132 cue presentation, 95% CI [0.56, 0.72]), displaying relatively fast reaction times (last session, 133 median = 0.32s [0.30, 0.34]) and reproducible forelimb kinematic trajectories (Figures 1d-1f, 134 Video S1).135 136

Understanding cortical function requires studying its multiple scales: molecular, cellular, circuit and behavior. We developed a biophysically detailed multiscale model of mouse primary motor cortex (M1) with over 10,000 neurons, 30 million synapses. Neuron types, densities, spatial distributions, morphologies, biophysics, connectivity and dendritic synapse locations were derived from experimental data. The model includes long-range inputs from 7 thalamic and cortical regions, and noradrenergic inputs from locus coeruleus. Connectivity depended on cell class and cortical depth at sublaminar resolution. The model reproduced and predicted in vivo layer- and cell type-specific responses (firing rates and LFP) associated with behavioral states (quiet and movement) and experimental manipulations (noradrenaline receptor blocking and thalamus inactivation), and enabled us to evaluate different hypotheses of the circuitry and mechanisms involved. This quantitative theoretical framework can be used to integrate and interpret M1 experimental data and sheds light on the M1 cell type-specific multiscale dynamics associated with a range of experimental conditions and behaviors.