A computational approach to understanding effort-based decision-making in depression

Vincent Valton, Anahit Mkrtchian, Madeleine Moses-Payne, Alan Gray, Karel Kieslich, Samantha VanUrk, Veronika Samborska, Don Chamith Halahakoon, Sanjay Manohar, Peter Dayan, Masud Husain, Jonathan P. RoiserbioRxiv 2024 · DOI

Abstract

Motivational dysfunction is a core feature of depression that significantly impairs daily functioning. However, the specific cognitive processes underlying this impairment—and whether they persist after clinical remission—remain poorly understood. This study utilizes the Apple Gathering Task (AGT) and computational modeling to decompose effort-based decision-making into distinct parameters: reward sensitivity, effort sensitivity, and a general acceptance bias. By comparing unmedicated individuals who are currently depressed, remitted, or at high familial risk against healthy controls, we identify that motivational deficits are primarily driven by a lower baseline "acceptance bias" rather than an aversion to effort or a reduced appreciation for rewards. This bias persists in remission, suggesting it may serve as a trait-like marker or a psychological "scar" of the disorder.

Introduction

Anhedonia (the loss of pleasure) and apathy (the loss of motivation) are hallmark symptoms of Major Depressive Disorder (MDD). While it is well-established that depressed individuals are less willing to exert effort for rewards, the underlying mechanism is often debated. Is the "cost" of physical effort perceived as more taxing? Is the "value" of the reward diminished? Or is there a more fundamental breakdown in the initial decision to engage in a challenge?

Previous research has been confounded by two major factors: the use of antidepressants, which can alter reward processing, and a lack of individual calibration for physical ability. This study addresses these gaps by using a calibrated physical effort task in an unmedicated sample to determine if motivational shifts are a transient "state" of an active episode or a stable "trait" that precedes or follows the illness.

Methods

The Apple Gathering Task (AGT)

Participants performed a computerized task where they decided whether to accept or reject offers to gather "apples" (rewards) in exchange for physical effort. Effort was exerted using a hand-grip dynamometer. Crucially, the task was calibrated to each participant’s Maximum Voluntary Contraction (MVC), ensuring that a 60% effort requirement felt subjectively similar for all participants regardless of their raw physical strength.

Computational Modeling

The researchers evaluated 70 mathematical models to identify the one that best captured participant behavior. The winning model utilized four key parameters to describe the decision-making process:

  1. Linear Reward Sensitivity ($LinR$): The degree to which increasing rewards makes an offer more attractive.
  2. Linear Effort Sensitivity ($LinE$): The baseline perceived cost of exerting effort.
  3. Quadratic Effort Sensitivity ($E^2$): The rate at which effort becomes disproportionately more difficult as it approaches the participant's physical limit.
  4. Acceptance Bias ($K$): A constant intercept representing a general tendency to engage in a task, independent of the specific reward or effort levels.

Participant Groups

The study included 180 unmedicated participants across four groups:

  • MDD: Currently depressed individuals ($N=41$).
  • REM: Individuals remitted from depression ($N=46$).
  • REL: Healthy individuals with a first-degree relative with depression (high-risk group, $N=36$).
  • CTR: Healthy controls with no personal or family history ($N=57$).

Results

The Dominance of Acceptance Bias

The most significant finding was that the difference between depressed individuals and controls was not found in the "slopes" of reward or effort sensitivity. Instead, the difference lay in the Acceptance Bias ($K$). Both currently depressed (MDD) and remitted (REM) individuals showed a significantly lower baseline tendency to accept challenges ($p = 0.025$). Once an offer was accepted, however, all groups showed high success rates (>80%), indicating that the results were not due to physical incapacity.

State vs. Trait

Because the remitted group (REM) exhibited the same lower acceptance bias as the currently depressed group (MDD), the findings suggest that this motivational deficit is not merely a symptom of an active depressive episode. It may represent a stable trait—either a pre-existing vulnerability factor or a "scar" left by previous depressive episodes. Interestingly, the high-risk (REL) group did not show this bias, though this may be due to the older age of the cohort, potentially representing a "resilient" population.

Discussion

These results suggest that motivational impairment in depression is driven by a general reluctance to engage—a "no" by default—rather than a specific overestimation of effort or underestimation of reward. The authors propose that this low acceptance bias may reflect a lack of metacognitive confidence. Even when participants are physically capable of completing a task, their internal expectation of success may be lower, leading them to reject the challenge entirely.

Clinical Implications

This research highlights Behavioral Activation (BA) as a particularly relevant therapeutic approach. Since the primary barrier is the initial decision to engage (the "intercept") rather than the experience of the work itself (the "slope"), interventions that focus on scheduling and initiating activities regardless of current mood may directly counteract this low acceptance bias. Future work should focus on longitudinal studies in younger, high-risk populations to determine if this bias can predict the future onset of depressive episodes.

Summary updated

Abstract

Abstract Background Motivational dysfunction is a core feature of depression, and can have debilitating effects on everyday function. However, it is unclear which disrupted cognitive processes underlie impaired motivation, and whether impairments persist following remission. Decision-making concerning exerting effort to obtain rewards offers a promising framework for understanding motivation, especially when examined with computational tools which can offer precise quantification of latent processes. Methods Effort-based decision-making was assessed using the Apple Gathering Task, where participants decide whether to exert effort via a grip-force device to obtain varying levels of reward; effort levels were individually calibrated and varied parametrically. We present a comprehensive computational analysis of decision-making, initially validating our model in healthy volunteers (N=67), before applying it in a case-control study including current (N=41) and remitted (N=46) unmedicated depressed individuals, and healthy volunteers with (N=36) and without (N=57) a family history of depression. Results Four fundamental computational mechanisms that drive patterns of effort-based decisions, which replicated across samples, were identified: overall bias to accept effort challenges; reward sensitivity; and linear and quadratic effort sensitivity. Traditional model-agnostic analyses showed that both depressed groups showed lower willingness to exert effort. In contrast with previous findings, computational analysis revealed that this difference was primarily driven by lower effort acceptance bias, but not altered effort or reward sensitivity. Conclusions This work provides insight into the computational mechanisms underlying motivational dysfunction in depression. Lower willingness to exert effort could represent a trait-like factor contributing to symptoms, and might represent a fruitful target for treatment and prevention.
BibTeX
@misc{valton2024computational,
  author = {Vincent Valton and Anahit Mkrtchian and Madeleine Moses-Payne and Alan Gray and Karel Kieslich and Samantha VanUrk and Veronika Samborska and Don Chamith Halahakoon and Sanjay Manohar and Peter Dayan and Masud Husain and Jonathan P. Roiser},
  title = {{A computational approach to understanding effort-based decision-making in depression}},
  year = {2024},
  doi = {10.1101/2024.06.17.599286},
  howpublished = {bioRxiv},
}
Figure