Fig. 2: The duration of velocity modulation of individual units depends on the temporal lag to behavior and on the cortical area.

A Two hypotheses of sequential neuronal activity relative to movement. The duration of the neuronal activity can be constant across lags (upper panel) or different across lags (lower panel). B Units sorted according to the lag of their maximum velocity modulation in the locomotor task (top), and for the joystick task (bottom). C Relationship between average modulation duration and temporal lag (black line), across the different animals (colored lines) for the locomotor (left) and joystick task (right). D Top panel: Duration of the velocity modulation for each cortical area for the locomotor task (left) and the joystick task (right). Two-sided non-paired t-test corrected for dependence (see methods). Locomotor task: S1 (n = 1766), M1 (n = 2450), M2 (n = 1197), M1-M2: p = 1.7e-35, S1-M2: p = 3.7e-53, S1-M1: p = 1.3e-9. Joystick task: S1 (n = 837), M1 (n = 1015), M2 (n = 759), M1-M2: p = 3.9e-9, S1-M2: p = 2.1e-31, S1-M1: p = 2.8e-17. The data for each box-plot is bootstrapped and corrected for dependence across sessions (see methods). Box-plots: central mark indicates the median, bottom and top edges refer to the 25th and 75th percentiles. Primary sensory cortex (S1), primary motor cortex (M1), and secondary motor cortex (M2). Bottom panel: Histograms of the modulation duration for each cortical area for the locomotor task (left) and the joystick task (right). A mixed-effect model was used to calculate significances between the cortical areas. Mixed-effect model (see methods): Locomotor task: M1-M2: p = 2.2e-6, S1-M2: p = 4.0e-5, S1-M1: p = 0.23. Joystick task: M1-M2: p = 0.0142, S1-M2: p = 6.2e-5, S1-M1: p = 0.016. Significances are indicated according to *p < 0.05, **p < 0.01, and ***p < 0.001. Source data are provided as a Source Data file.