Two Ways to Change Your Mind: Effects of Intentional Strength and Motor Costs on Changes of Intention

Anne Löffler, Anastasia Sylaidi, Zafeirios Fountas, Patrick HaggardbioRxiv 2019 · DOI

TL;DR

This paper explores how we change our minds during voluntary actions, not just when reacting to external stimuli. They found that the strength of our initial intention and the effort required to carry it out both influence whether we stick with our plan or switch gears. This work proposes a computational model that integrates these factors, offering a unified view of dynamic decision-making.

What Problem Does This Paper Address?

We often think of "changing our minds" as a reaction to new information – like seeing a sign for a better restaurant after already deciding where to eat. However, our decisions are also influenced by internal factors like our own intentions and the effort involved in executing them. Previous research on "Changes of Mind" (CoM) primarily focused on how we update our actions based on external sensory evidence. This paper tackles the question of how our internal intentions and the motor costs associated with pursuing them influence our decisions, particularly in voluntary actions where the intention originates internally. They aim to understand the interplay between our internal goals and the physical demands of achieving them, and how this dynamic influences our flexibility in action.

Methods

The researchers developed a novel version of the random-dot motion (RDM) task. In this task, participants first made an endogenous choice of a target color (e.g., blue or green) and then had to move a cursor to a target that matched both their chosen color and the direction of the moving dots. By analyzing the movement trajectories, they could distinguish between:

  • Changes of Movement (CoMov): Updating the motor command to a different target that still matched the original color intention. This reflects a change in response to the external dot-motion stimulus.
  • Changes of Movement + Intention (CoMov+Int): Updating the motor command to a different target that also involved a change in the original color intention. This indicates a shift in the internal goal.

They conducted two experiments:

  • Experiment 1 investigated the effect of intentional strength. They used performance costs in conflict trials (where perceptual information clashed with the intended color) as a proxy for how strongly participants committed to their initial color choice.
  • Experiment 2 manipulated motor costs by varying the distance between targets of different colors. This allowed them to see how the effort required to switch intentions affected the likelihood of doing so.

Finally, they developed an Attractor Network Model (ANM) to simulate these processes. This model integrates voluntary intentions, sensory evidence, and motor costs, proposing a biologically plausible mechanism for how these factors dynamically interact to guide action selection and allow for changes of mind.

Datasets

Not discussed in the paper. The study used data collected from human participants performing the described experimental tasks.

Key Insights and Ideas

  • Intentions aren't static: Our internal goals aren't fixed once formed; they are continuously evaluated and can be changed, even during action execution.
  • Effort matters: The physical cost of pursuing an intention significantly influences our willingness to stick with it. If switching to a new goal is easier, we're more likely to do so.
  • Intentional strength is quantifiable: By measuring performance costs in conflict situations, researchers can get a handle on how strongly individuals commit to their initial intentions.
  • A unified framework: The attractor network model suggests that changes of mind, whether driven by external stimuli or internal shifts in intention, can be explained by a common underlying mechanism of dynamic information integration.
  • Hierarchy in decision-making: The model proposes a hierarchical structure where higher-level intentions (what goal) influence lower-level sensorimotor processes (how to achieve it), including regulating "noise" in the system.

Key Results

  • Experiment 1 (Intentional Strength): Participants with stronger initial color intentions (indicated by higher performance costs in conflict trials) were less likely to change their intention (CoMov+Int).
  • Experiment 2 (Motor Costs): When the motor cost of switching to a different color target was high (i.e., targets were far apart), participants were less likely to change their intention (CoMov+Int). Conversely, when the cost was low (targets close), CoMov+Int occurred more frequently.
  • CoM Frequency: Changes of Mind (CoM) were more frequent when perceptual information was noisy (test trials) compared to easy trials.
  • Sense of Agency (SoA): Changes of Movement (CoMov) led to a reduced sense of agency, but this effect was largely explained by longer movement times. Changes of Intention (CoMov+Int) did not significantly affect SoA.
  • Model Performance: The proposed attractor network model successfully replicated the key behavioral findings from both experiments, including the effects of intentional strength and motor costs on CoMov+Int.

Limitations and Future Work

  • Trial Numbers for SoA: The authors note that the absence of a significant effect of Changes of Intention on SoA should be interpreted with caution due to low trial numbers for these specific conditions. They suggest that strong and sustained intentions might still contribute to SoA.
  • Inter-individual Differences: While the study captured inter-individual differences in intentional strength, they acknowledge that their design didn't allow for capturing trial-by-trial variability in intentional strength within individuals. Future work could explore this dynamic.
  • Model Predictions: The model makes further predictions, such as the direct link between action node firing rates and fine-grained motor policies (e.g., movement speed), which can be tested in future studies.
  • Distinguishing Errors from Intentional Shifts: The model allows for analyzing how initial color errors might contribute to CoMov+Int, which could be a fruitful avenue for future research.

Summary updated

Abstract

Abstract Changes of Mind are a striking example of the human ability to flexibly reverse decisions after commitment to an initial choice, and to change actions according to circumstances. Previous studies of Changes of Mind largely focused on perceptual choices. Here we investigate reversals of voluntary, endogenous action decisions. In a novel version of the random-dot motion task, participants moved to a target that matched both the perceived dot-motion direction and an internally-generated intention (which colour target to move to). Movement trajectories revealed whether and when participants 1) perceived a change in dot-motion direction, or additionally 2) changed the colour that they chose to move to (‘Change of Intention’). Changes of Intention were less frequent in participants with strong colour intentions, as indicated by high performance costs in trials where perceptual information conflicted with the endogenous intention (Exp. 1). Additionally, Changes of Intention were more frequent when motor costs of intention pursuit were high (Exp. 2). These findings were simulated using an attractor network model that continuously integrates voluntary intentions, sensory evidence, and motor costs. This argues in favour of a unifying framework for dynamic decision-making processes, in which voluntary actions emerge from a dynamic combination of internal action tendencies and external environmental factors.
BibTeX
@misc{loffler2019two,
  author = {Anne Löffler and Anastasia Sylaidi and Zafeirios Fountas and Patrick Haggard},
  title = {{Two Ways to Change Your Mind: Effects of Intentional Strength and Motor Costs on Changes of Intention}},
  year = {2019},
  doi = {10.1101/841882},
  howpublished = {bioRxiv},
}

Loading related content...

Figure