🧠 neuroscience

Involuntary feedback responses reflect a representation of partner actions

This study demonstrates that high-level representations of a partner's actions and movement costs rapidly modulate involuntary visuomotor feedback responses within 180–230 ms, revealing that the sensorimotor system integrates social considerations into fast, reflexive motor control to facilitate seamless joint coordination.

Original authors: Sullivan, S. R., Buggeln, J. H., Calalo, J. A., Ngo, T. T., Semrau, J. A., Carter, M. J., Cashaback, J. G. A.

Published 2026-02-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sullivan, S. R., Buggeln, J. H., Calalo, J. A., Ngo, T. T., Semrau, J. A., Carter, M. J., Cashaback, J. G. A.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The Big Idea: Your Brain's "Auto-Pilot" Knows Your Friend's Goals

Imagine you are playing a video game with a friend where you both have to steer a single spaceship together to hit a target. Usually, we think of "reflexes" (like pulling your hand away from a hot stove) as dumb, automatic reactions that only care about you.

This paper asks a fascinating question: Does your brain's "auto-pilot" reflex actually care about your friend's goals, even before you consciously think about it?

The answer is yes. The researchers found that within a split second (faster than you can blink), your brain automatically adjusts your movements to help your partner, even if it costs you a little extra energy.

The Experiment: The "Dance of Two Hands"

To test this, the researchers set up a clever game:

The Setup: Two people sat across from each other, each holding a robotic handle. They couldn't see their own hands, only a screen showing three dots:
- A Green Dot (The "Center"): This was the average of both their hands.
- A Grey Dot (Their own hand).
- A White Dot (Their partner's hand).
The Goal: They had to work together to move the Green Dot into their own specific target zone on the screen.
The Twist: The researchers changed the size of the targets.
- Narrow Target (The "Hard Mode"): You have to be very precise. If you drift even a little, you miss.
- Wide Target (The "Easy Mode"): You can be sloppy. Even if you drift, you still hit the target.

The Test:
Suddenly, the Green Dot would jump sideways (a "glitch" in the simulation). The researchers measured how fast and how hard the participants' hands pushed back to correct the dot.

They tested four scenarios:

Scenario A: Both targets are wide (Easy/Easy).
Scenario B: Your target is wide, but your partner's is narrow (Easy/Hard).
Scenario C: Your target is narrow, but your partner's is wide (Hard/Easy).
Scenario D: Both targets are narrow (Hard/Hard).

The Discovery: The "Altruistic Reflex"

Here is the magic part. When the Green Dot jumped, the participants' hands reacted involuntarily (in about 180–230 milliseconds). This is too fast for them to be thinking, "Oh, my partner is struggling, I should help." It's a pure reflex.

What they found:

If your partner had a "Hard Mode" (Narrow) target: Your hand automatically pushed harder to correct the dot, even if your target was "Easy Mode" (Wide). You were willing to spend extra energy to save your partner from missing.
If your partner had an "Easy Mode" (Wide) target: Your hand pushed less, even if your own target was "Hard." You didn't need to help them as much.

The Metaphor:
Imagine you and a friend are pushing a heavy shopping cart.

If your friend is trying to park the cart in a tiny, tight spot (Narrow Target), and the cart suddenly rolls sideways, you instinctively push harder to correct it, even if you are just trying to park in a huge, empty lot (Wide Target).
Your brain realizes, "Hey, my friend is in trouble," and your muscles react to help them before your conscious mind even says, "I should help."

The "Cost" Calculation

The researchers used a computer model (like a video game AI) to figure out how the brain was doing this. They tested four theories:

The Selfish Robot: "I only care about my own target." (This didn't match the data).
The Mind-Reader: "I know what my friend is doing, but I only care about my own target." (This didn't match either).
The Perfect Altruist: "I care about my friend's goal exactly as much as my own." (Close, but not quite).
The Weighted Partner: "I care about my friend, but I care about myself a little bit more."

The Winner: The brain acts like a Weighted Partner. It has a mental map of what your friend is trying to do. It calculates the "cost" (how much effort it takes) for both of you. It decides to sacrifice a tiny bit of its own energy to help the friend, but it doesn't go into "hero mode" if it's not necessary. It's a smart, balanced cooperation.

Why This Matters

This study changes how we think about human connection.

Old View: Reflexes are dumb, automatic, and only about survival (e.g., don't touch fire).
New View: Our "dumb" reflexes are actually social. They are tuned by our high-level understanding of other people.

It's like your brain has a "Social Auto-Pilot." Even when you aren't thinking about it, your nervous system is constantly asking, "What does my partner need right now?" and adjusting your movements to help them. This suggests that being a team player isn't just a conscious choice; it's built into our very muscle reflexes.

Summary in One Sentence

Your brain is so good at teamwork that even your involuntary muscle reflexes automatically adjust to help a partner succeed, proving that cooperation is hardwired into our nervous system.

1. Problem Statement

Humans possess a remarkable ability to rapidly coordinate movements with others (e.g., dancing, sports). While it is well-established that humans use high-level cognitive representations of a partner's goals and movement costs (accuracy and energy) to plan voluntary actions, it remains unknown whether these high-level representations influence low-level, involuntary sensorimotor circuits.

Specifically, the authors investigate:

Can a high-level representation of a partner modulate rapid, involuntary visuomotor feedback responses (occurring within 180–230 ms)?
Does the sensorimotor system consider a partner's movement costs (e.g., energy expenditure to correct errors) when generating these involuntary reflexes, or does it only optimize for self-cost?

2. Methodology

Experimental Design

The study employed two experiments involving human pairs performing a cooperative reaching task using robotic manipulanda (KINARM robots).

Task: Two participants jointly controlled a "center cursor" (the midpoint of their two hands) to reach a target. Each participant saw their own target and their partner's target.
Manipulation: The width of the targets was manipulated to be either narrow (task-relevant) or wide (task-irrelevant).
- Relevant: Participants must correct lateral deviations to hit the target.
- Irrelevant: Participants do not need to correct lateral deviations.
Conditions: A $2 \times 2$ $2 \times 2$ repeated-measures design:
1. Partner-Irrelevant / Self-Irrelevant
2. Partner-Relevant / Self-Irrelevant
3. Partner-Irrelevant / Self-Relevant
4. Partner-Relevant / Self-Relevant
Perturbations:
- Experiment 1: A lateral jump of the center cursor occurred during the reach.
- Experiment 2: A lateral jump of both targets occurred.
Probe Trials: To isolate involuntary responses, participants were constrained by a stiff force channel (preventing lateral movement). When a cursor or target jumped, the lateral force exerted by the participant against the channel was measured. This force serves as a proxy for the involuntary visuomotor feedback response.

Computational Modeling

The authors developed a Dynamic Game Theory model (Linear-Quadratic Game) to generate a priori predictions. The model treated the pair as two controllers ( $u_1$ and $u_2$ ) minimizing cost functions. Four distinct hypotheses were tested:

No Partner Representation & Self Cost: The controller ignores the partner and only minimizes its own cost.
Partner Representation & Self Cost: The controller knows the partner's policy but only minimizes its own cost.
Partner Representation & Equal Joint Cost: The controller knows the partner's policy and minimizes the sum of self and partner costs equally.
Partner Representation & Weighted Joint Cost: The controller knows the partner's policy and minimizes a weighted sum where self-cost is prioritized over partner-cost ( $\alpha \in [0, 1]$ ).

3. Key Contributions

Novel Paradigm: The first study to demonstrate that high-level social representations (partner goals/costs) modulate involuntary sensorimotor feedback loops (180–230 ms), distinct from voluntary control.
Computational Framework: Application of Dynamic Game Theory to human-human sensorimotor interaction, specifically modeling how controllers incorporate a partner's cost function into their own feedback gains.
Evidence of Altruism in Reflexes: Demonstrating that the nervous system is willing to "sacrifice" energy (increase feedback gain) to assist a partner, even when the partner's goal is irrelevant to the self.

4. Results

Voluntary Behavior (Trajectories)

In conditions where the partner's target was relevant but the self's was irrelevant, participants still made lateral corrections to help the partner.
The group average trajectories aligned most closely with the Partner Representation & Weighted Joint Cost model, indicating that participants considered the partner's goals but prioritized their own.

Involuntary Visuomotor Feedback Responses (Probe Trials)

Partner Representation: In the Partner-Relevant/Self-Irrelevant condition, participants exhibited a significantly larger involuntary feedback response compared to the Partner-Irrelevant/Self-Irrelevant condition ( $p < 0.001$ $p < 0.001$ ).
- Interpretation: Even though the self did not need to correct, the sensorimotor system generated a stronger reflex to help the partner because the partner's target was relevant.
Weighted Joint Cost: In the Partner-Relevant/Self-Relevant condition, the feedback response was smaller than in the Partner-Irrelevant/Self-Relevant condition ( $p = 0.002$ $p = 0.002$ in Exp 1).
- Interpretation: When both targets were relevant, the self reduced its own corrective effort because it anticipated the partner would also contribute (a form of shared load). This matches the Weighted Joint Cost model, where the self optimizes by expecting the partner to share the burden.
Consistency: These patterns were observed in both Experiment 1 (cursor jump) and Experiment 2 (target jump), as well as in semi-involuntary (230–300 ms) and voluntary (300–400 ms) epochs.

Model Fit

The empirical data across both experiments closely matched the predictions of the Partner Representation & Weighted Joint Cost model. Models assuming no partner representation or equal joint costs failed to predict the specific modulation of feedback gains observed in the data.

5. Significance

Top-Down Modulation: The findings provide strong evidence for top-down modulation from high-level cognitive circuits (Theory of Mind, partner representation) to low-level sensorimotor pathways (long-latency reflexes). This challenges the view that involuntary reflexes are purely reactive to physical perturbations.
Social Affordances: The study extends the concept of "affordances" (opportunities for action) to include "social affordances." The sensorimotor system evaluates not just physical constraints but also the social context and the partner's needs.
Altruistic Control: The results suggest that the sensorimotor system is inherently "altruistic" in a cost-minimization framework; it is willing to expend extra energy to ensure a partner succeeds, even at an involuntary level.
Evolutionary Implication: This rapid integration of social goals into motor control may be an evolutionary adaptation for complex cooperative behaviors, allowing for seamless coordination without the delay of conscious deliberation.

In conclusion, the paper establishes that human sensorimotor control is not merely a self-optimizing process but a socially integrated system where involuntary reflexes are tuned by a representation of the partner's goals and costs.