Visuo-motor interference is modulated by task interactivity: A kinematic study

Matilde Rocca, Lucia Maria Sacheli, Luca Romeo, Andrea CavalloPsychonomic 2023 · DOI

TL;DR

When we watch someone else move, our own motor system often "leaks" and tries to imitate them—a glitch known as visuo-motor interference. This study reveals that this interference significantly drops when we are working with a partner toward a shared goal, suggesting that our brains switch from automatic imitation to active, predictive "dyadic motor planning."

What Problem Does This Paper Address?

Psychologists have long known that observing an action can interfere with performing one. If you see someone use a whole-hand power grip while you are trying to perform a delicate precision pinch, your movement will likely be slower or less accurate. This is called visuo-motor interference (VMI) or automatic imitation.

However, the literature is conflicted on how social context affects this. Does working together make us more sensitive to a partner's movements (increasing interference), or does it allow us to integrate their actions into a shared plan (decreasing interference)? The authors set out to resolve this by analyzing the fine-grained "kinematic unfolding" of movements—the literal path and speed of a hand in flight—rather than just looking at simple reaction times.

Methods

The researchers paired 16 participants with a confederate (an actor working for the researchers) to perform reach-to-grasp tasks.

The Setup: * Actions: Participants performed either a Precision Grip (PG) or a Whole-Hand Prehension (WHP). * Conditions: 1. Noninteractive: Participants focused only on their own speed. 2. Joint-Movement: The pair aimed for a specific joint "configuration." 3. Joint-Outcome: The pair worked together to produce a sensory result (a specific colored circle on a screen).

The Technical Twist: Instead of traditional statistical averages, the team used a Machine Learning approach. They trained SVM-LASSO (Support Vector Machines with Least Absolute Shrinkage and Selection Operator) classifiers on the participants' movement data (wrist velocity, acceleration, grip aperture, etc.).

The logic was clever: if a machine learning model can accurately predict what the confederate did just by looking at the participant’s hand kinematics, it proves that the confederate’s action "interfered" with and left a signature on the participant's movement.

Datasets

The study utilized a primary dataset collected by the authors:

  • N = 3,840 individual reach-to-grasp movements.
  • Data was captured using a Vicon motion-capture system at 100Hz with 20 retro-reflective markers per agent.
  • The dataset is available via the Open Science Framework (OSF).

Key Insights and Ideas

  • The Dyadic Motor Plan: The most significant insight is that when we share a goal, we don't just simulate the other person's movement (imitation); we form a "dyadic motor plan." We represent the partner's action in terms of its predicted consequences for the team goal.
  • Prediction Over Simulation: In a non-social or individualistic setting, the motor system defaults to "automatic simulation"—essentially a mirror-neuron response that causes interference. In joint action, the brain shifts to "active prediction," which is less disruptive to our own motor execution.
  • Kinematic Leakage: Interference isn't just a delay in reaction time; it's a physical change in the movement. The study found that participants' velocity and acceleration profiles actually became more similar to the confederate's over time in individual tasks, but this "motor contagion" was buffered during cooperation.

Key Results

  • Classification Accuracy as a Proxy for Interference: The SVM could identify the confederate's action from the participant's hand movements significantly above chance in all conditions. However, accuracy was highest in the noninteractive condition (~62.5%), indicating the strongest interference.
  • Reduction in Joint Tasks: Interference dropped significantly in the joint-movement (~58.6%) and joint-outcome (~55.4%) conditions.
  • Movement Time (MT): In the noninteractive condition, participants were significantly slower when the confederate performed an "incongruent" action. In both joint action conditions, this "slowing down" effect vanished.
  • Feature Importance: Wrist velocity, acceleration, grip aperture, and wrist height were the most "informative" features for the classifier, meaning these are the specific dimensions where motor interference manifests most clearly.

Limitations and Future Work

  • Indirect Evidence: While the results strongly support the "dyadic motor plan" hypothesis, the study did not measure brain activity (e.g., fMRI or EEG) directly to confirm the shift in neural processing.
  • Cooperation vs. Competition: The study focused entirely on cooperative joint actions. The authors note that future research should investigate if competitive settings (where you might want to avoid being predicted) change the modulation of visuo-motor interference.
  • Task Complexity: Participants were slower overall in joint conditions, suggesting these tasks were cognitively more demanding. While the authors argue this doesn't explain the reduction in interference, the interplay between cognitive load and motor imitation remains a complex area for further study.

Summary updated

Abstract

Extensive evidence shows that action observation can influence action execution, a phenomenon often referred to as visuo-motor interference. Little is known about whether this effect can be modulated by the type of interaction agents are involved in, as different studies show conflicting results. In the present study, we aimed at shedding light on this question by recording and analyzing the kinematic unfolding of reach-to-grasp movements performed in interactive and noninteractive settings. Using a machine learning approach, we investigated whether the extent of visuo-motor interference would be enhanced or reduced in two different joint action settings compared with a noninteractive one. Our results reveal that the detrimental effect of visuo-motor interference is reduced when the action performed by the partner is relevant to achieve a common goal, regardless of whether this goal requires to produce a concrete sensory outcome in the environment (joint outcome condition) or only a joint movement configuration (joint movement condition). These findings support the idea that during joint actions we form dyadic motor plans, in which both our own and our partner's actions are represented in predictive terms and in light of the common goal to be achieved. The formation of a dyadic motor plan might allow agents to shift from the automatic simulation of an observed action to the active prediction of the consequences of a partner's action. Overall, our results demonstrate the unavoidable impact of others' action on our motor behavior in social contexts, and how strongly this effect can be modulated by task interactivity.
BibTeX
@inproceedings{rocca2023visuo,
  author = {Matilde Rocca and Lucia Maria Sacheli and Luca Romeo and Andrea Cavallo},
  title = {{Visuo-motor interference is modulated by task interactivity: A kinematic study}},
  booktitle = {Psychonomic},
  year = {2023},
  doi = {10.3758/s13423-023-02297-z},
}
Figure