Changes of Mind after Movement Onset Depend on the State of the Motor System

Abstract

Decision-making is traditionally described as a cognitive process of deliberation followed by commitment to an action choice, preceding the planning and execution of the chosen action. However, this is challenged by recent data suggesting that during situated decisions, multiple options are specified simultaneously and compete in premotor cortical areas for selection and execution. Previous studies focused on the competition during planning and left unaddressed the dynamics of decisions during movement. Does deliberation extend into the execution phase? Are nonselected options still considered? Here we studied a decision-making task in which human participants were instructed to select a reaching path trajectory from an origin to a rectangular target, where reward was distributed nonuniformly at the target. Critically, we applied mechanical perturbations to the arm during movement to study under which conditions such perturbations produce changes of mind. Our results show that participants initially selected the direction of movement toward the highest reward region and changed their mind most frequently when the two choices offered the same reward, showing that deliberation continues and follows cost-benefit considerations during movement. Furthermore, changes of mind were dependent on the intensity of the perturbation and the current state of the motor system, including velocity and distance to targets. Although reward remains most relevant, our results indicate that the state of the motor system when the perturbation occurs is a crucial determinant of changes of mind.

Keywords: KINARM; change of mind; decision-making; motor system; reward.

Figure 1.

A, Geometrical arrangements of the stimuli on different trials consist of a circular origin cue (1 cm) and a rectangular target (width, 10 cm; depth, 1 cm), placed ∼15 cm away from the origin at an orientation of 135°. We show the three distributions of reward, from left to right: 3-3, 1-5, and 5-1. respectively. B, KINARM setup. C, Time course of a baseline trial. The trial starts with an empty screen for 1000 ms. After this interval, the origin cue (1 cm blue circular cue) is shown on the bottom right part of the screen. When the end point (right-hand fingertip) enters the cue, the cue turns green, and the rectangular target and the distribution of reward are shown on the top left part of the screen. One second later, the origin cue turns white to indicate the GO signal. If the subject’s end point leaves before the GO is given, the screen turns blank and the trial is invalidated. When the end point enters the target, the rectangular cue turns green. The screen turns blank 500 ms after that. D, Time course of a perturbed trial. The trial follows the same baseline trial pattern described in C. However, the arm is perturbed perpendicularly to the straight line from the origin to the center of the target, 1–3 cm after the end point leaves the origin cue. Three separate factors are considered for perturbed trials: early/late, weak/strong, and right/left.

Figure 2.

A, Typical baseline trajectories for subjects 5, 11, 12, and 15 during baseline trials. Please note that the figure shows trials from all three reward distributions: 3-3, 5-1, and 1-5. B, Perturbed trajectories for subjects 5, 11, 30, and 33, classified as a function of (1) the initial choice direction (rightward/leftward) to the direction of motion and (2) the direction of the perturbation (rightward/leftward). Blue trajectories indicate CoM trials, while red trajectories indicate non-CoM trials (Extended Data Fig. 2-1).

Figure 3.

A, Case of ICR/PL. Perturbed radial velocities, tangential velocities, and tangential accelerations, aligned at the time of perturbation, indicated by a vertical black line, for four kinds of trials, using data from S11. The thin black arrow indicates the time at which the perturbation is applied. B, Same as A, but for initial choice right, perturbed left. The thick black arrow indicates the time at which the CoM comes into effect on the trajectory. C, Same as A, but for the case of initial choice left, perturbed right. D, Same as A, but for the case of initial choice left, perturbed left.

Figure 4.

A, Group average and SEM regression coefficients on the PCoM as a function of reward gain (−4, 0, 4), perturbation direction (inward/outward), and their interaction. B, Group average and SEM PCoM as a function of reward, perturbation direction, and initial choice. C, Baseline and probe trial trajectories (Fig 3) for two subjects. D, Scatter plots of the average PCoM per subject when aiming at R = 5 versus R = 3 in four cases: ICR/PR 3-3 versus 1-5; ICR/PL 3-3 versus 1-5; ICL/PR 3-3 versus 5-1; ICL/PL 3-3 versus 5-1.

Figure 5.

A, Group average regression coefficients a GLM on the p binomial distribution parameter, calculated on an individual subject basis, as a function of the following four factors: reward to be gained from PCoM, time of perturbation, intensity of perturbation, tangential velocity, and their interactions.Group significance is reported by a * symbol (p<0.05). B, Group average effect of reward gain on PCoM. C–E, Group average effects for the perturbation time, perturbation intensity and velocity at the time of perturbation. Error bars indicate the SEM.

Figure 6.

A, Probe trajectories for the cases of study: ICR/PR, ICR/PL, ICL/PR, and ICL/PL for subject 4. B, Distance right and left in each of the four cases shown in A. Blue, CoM trajectories; red, non-CoM trajectories. C, State of the motor apparatus, characterized by DR and DL at the time of peak deviation, and at the time the difference between CoM and non-CoM trajectories became significant. D, Geometrical definitions for the distance-based analysis of the state of the motor apparatus: DL/DR for a sample perturbed trajectory. Illustration of the trajectory peak deviation. E, State of the motor apparatus, characterized by DR and DL at the time they became statistically significant with p = 0.05 (Extended Data Figs. 6-1, 6-2, 6-3, 6-4).

Figure 7.

Effects of reward on DR and DR at peak deviation on non-CoM probe trials in four cases listed next. A, ICR/PR, 3-3 versus 5-1; ICR/PL, 3-3 versus 5-1; ICL/PR, 3-3 versus 1-5; ICL/PL, 3-3 versus 1-5. B, Sample trajectories for the cases described in A for subjects 4 and 11. C, Group average and SE of the peak DR for the cases described in A. D, Scatter plot of each subject’s average DR, evaluated at R = 5 versus R = 3 trials. E, Group distributions of peak DR, evaluated in the cases described in A. F–H, Same as C–E, but for DL.