A computational account of how positive performance bias supports cognitive effort

This study provides a computational account demonstrating that a positive bias in self-evaluation, where individuals overestimate their performance especially under high cognitive demands, helps attenuate the subjective cost of effort and sustain engagement in effortful tasks.

The Gist

Imagine your brain is like a car engine. Sometimes, you need to drive up a steep, rocky hill (a difficult mental task). Your brain knows this will burn a lot of fuel (cognitive effort), so it naturally wants to stay on the flat, easy road instead. Usually, the fear of running out of gas keeps us from tackling those hard hills.

But this paper suggests there's a special kind of "optimism filter" in our minds that helps us keep driving up those steep hills.

The Experiment: Guessing Your Own Score
The researchers asked people to play a game where they had to choose between easy tasks and hard tasks. Before each round, the players had to guess how well they thought they would do.

Here is the twist: The players were almost always too confident. They thought they would do better than they actually did. This gap between what they thought they would achieve and what they actually achieved is called a "positive performance bias."

The study found something interesting: The harder the task was, the more the players overestimated their success. It's as if the steeper the hill, the more the driver convinces themselves, "I've got this! I'm a pro at this!" even when the road is actually quite treacherous.

The "Learning" Engine
To understand why people kept choosing the hard tasks, the researchers built a computer model to simulate the brain's decision-making process. They found that the brain learns in a specific way:

1. When you fail: If you try a hard task and fail, your brain gets a little more sensitive to the "cost" of effort next time. It's like saying, "Ouch, that was too expensive; I need to be more careful."
2. When you succeed (with a boost): If you try a hard task and succeed, but you also thought you were going to do even better than you actually did (that positive bias), your brain actually becomes less sensitive to the cost of effort. The optimism acts like a discount coupon, making the hard task feel cheaper and easier to tackle next time.

Confidence vs. Bias
The researchers also checked if this was just about "confidence" (feeling sure of yourself). They found that simple confidence wasn't enough to explain the behavior. It was specifically the mismatch—the fact that people were overestimating their performance—that kept them motivated. It's not just about feeling good; it's about feeling better than reality suggests.

The Big Picture
In simple terms, this paper shows that our tendency to be overly optimistic about our own abilities isn't just a harmless quirk. It acts as a protective shield. By convincing ourselves we are doing better than we actually are, we lower the "price tag" of mental effort in our minds. This positive bias helps us ignore the feeling of exhaustion and keeps us engaged in difficult mental work that we might otherwise avoid.

Technical Summary

Technical Summary: A Computational Account of How Positive Performance Bias Supports Cognitive Effort

Problem Statement
The willingness to exert cognitive effort is a fundamental component of adaptive behavior, yet it is inherently constrained by the subjective cost of effort. While individuals often avoid effortful tasks, the literature suggests that positive biases regarding one's own performance may serve as a mechanism to sustain engagement with cognitive demands. However, the specific computational mechanisms by which such biases influence effort-based decision-making remain unclear.

Methodology
To investigate this, the authors employed an effort-based decision-making task where participants were required to choose between low-effort and high-effort options. Crucially, on each trial, participants reported their predictions of their own performance prior to task execution. This design allowed the researchers to quantify Performance Prediction Error (PPE), defined as the discrepancy between subjective performance predictions and objective accuracy.

The study utilized a novel computational modeling approach to analyze these choices. The authors compared a learning model, which posits that PPE updates the sensitivity to effort costs, against a confidence-based control model. The learning model specifically tested how unsuccessful outcomes and rewarded trials accompanied by positive PPE influenced subsequent effort sensitivity.

Key Results
The empirical data revealed two primary findings regarding PPE:

1. Predominantly Positive Bias: PPE was predominantly positive across participants, indicating a general tendency to overestimate performance.
2. Effort-Dependent Overestimation: The magnitude of this positive bias increased with the level of effort required, suggesting that individuals overestimate their performance more significantly under higher cognitive demands.

In terms of computational modeling, the results demonstrated that participants' choices were best explained by the learning model. Specifically:

• Unsuccessful outcomes led to an increase in subsequent effort sensitivity (making individuals more sensitive to the cost of effort after failure).
• Rewarded trials accompanied by positive PPE led to a decrease in subsequent effort sensitivity (making individuals less sensitive to the cost of effort).

The confidence-based control model failed to provide a superior account of the observed choices, indicating that the effect is better captured by the specific mechanism of positive performance bias rather than by confidence alone.

Significance and Claims
The paper claims to provide a computational account of how biased self-evaluation functions to attenuate the subjective cost of cognitive effort. By demonstrating that positive performance bias specifically modulates effort sensitivity following rewarded trials, the study extends the existing literature on positive bias to the domain of the need for cognitive effort. The findings suggest that overestimating one's performance is not merely a cognitive error but may serve a functional role in sustaining engagement with difficult tasks by reducing the perceived cost of exerting effort.