Voluntary action—our ability to decide what to do and when to do it— is a complex challenge in cognitive science. New research published in Cognition explored this concept.
Volition involves more than just reacting to external cues or following ingrained habits; it encompasses our ability to initiate and control actions on our own terms. Traditional studies on this topic often struggle with experimental designs that don’t fully capture the spontaneity and complexity of real-world decision-making.
To address these limitations, researchers Keiji Ota and colleagues developed a novel approach using a competitive gaming framework, challenging participants to adapt their decision-making strategies in response to dynamic, unpredictable virtual competitors.
A total of 152 participants completed this study. The experimental setup required participants to press a key to deliver food in a digital game, aiming to strategically avoid interceptions by virtual birds. These birds were programmed to predict and react to participants’ actions based on the timing of their previous key presses. The game was structured into four distinct blocks, each escalating in complexity and designed to penalize specific decision-making biases.
In the baseline block (immediate response bias), the game penalized participants for reacting immediately to stimuli. This block tested participants’ ability to resist the impulse to respond immediately, promoting a more deliberate and controlled approach to the task.
In Block 1 (choice bias), the focus shifted to penalizing a preference for selecting certain actions over others without rational justification. The task encouraged participants to diversify their choices across different timing intervals, thereby challenging any preference for habitual selections.
Block 2 (transition bias) addressed the propensity to follow predictable sequences of actions. Participants were encouraged to make each choice independently of the previous ones, disrupting any sequence-based patterns in their decision-making.
Block 3 (reinforcement bias) targeted the influence of past outcomes on current decisions. This phase intended to detach participants’ choices from the outcomes of previous trials, promoting decisions that were less predictable and independent of prior rewards or penalties.
Throughout these blocks, the participants were not provided with explicit instructions on how the virtual competitors (birds) would adapt their strategies. Instead, they had to infer the rules and adjust their strategies based solely on the feedback from trial outcomes.
Results showed that participants were successful in the baseline block, avoiding immediate responses with a success rate of 96.6%. However, as the tasks progressed to penalize more complex biases, success rates dropped significantly—to 64.3% in Block 1 and even lower in subsequent blocks (59.0% and 56.9% respectively). This pattern indicated increasing difficulty in adapting to more complex constraints on decision-making.
Despite attempts to adapt, the study revealed that while participants could adjust to simpler task demands, they struggled significantly with complex biases like reinforcement bias, highlighting the intrinsic challenges in achieving true behavioral autonomy.
A notable limitation is the potential influence of learning or fatigue over the course of the game, given the fixed sequence of competitors.
The study, “Autonomous behaviour and the limits of human volition” was authored by Keiji Ota, Lucie Charles, and Patrick Haggard.
Discussion about this post