In everyday life, people encounter different situations where decisions regarding the course of actions to obtain a reward need to be made. These decisions often incur a cost, such as patience or effort to obtain a reward. Sayings like “no pain no gain” are common, and as a result, people often deliberate if a reward is worth working for. Such decisions can be simple, like debating if one should make the physical effort to walk to the refrigerator to get a favorite snack. In more difficult cases, the process of achieving a goal may be tiring, for instance, working out daily to get your body toned. It is easy to give up when a great amount of effort is required. Moreover, it is hard to even initiate actions when anticipating the work needed over time.

When facing effortful actions, one often estimates the costs and benefits of available options and selects the most beneficial response in a given circumstance (Rangel et al., 2008). In decision-making, effort is regularly regarded as a principle cost, whereby effort discounting devalues rewards by decreasing the utility of related outcomes (Botvinick et al., 2009, Kool et al., 2010, Kurniawan et al., 2010). Thus, when available options include choices with high effort costs, the choice can become less appealing (Iyengar and Lepper, 2000, Kool et al., 2010).

Apart from physical effort, the willingness to engage in cognitively demanding tasks can also be characterized by effort cost. The concept of cognitive effort is intuitive as it is accompanied by a phenomenological experience in that engaging in a cognitively demanding task “feels” different compared to daydreaming. The study of cognitive effort is important because of its impact in various scenarios, from arithmetic, problem solving, and rational reasoning (Shah and Oppenheimer, 2008, Toplak et al., 2011), to economic decision-making (Garbarino and Edell, 1997, Payne et al., 1988, Shah and Oppenheimer, 2008, Smith and Walker, 1993, Westbrook and Braver, 2015), and much more. Cognitive effort expenditure may also act as a predictor of academic achievements (Arafa et al., 2018, von Stumm et al., 2011) as it can be defined as the amount of cognitive capacity allocated to learning and task performance (Arafa et al., 2018, De Jong, 2010). The willingness to expand cognitive effort in discounting motivation is a goal-directed behavior with its influence in managing cognitive control to reach a valuable goal (Shenhav et al., 2013). Moreover, Kool et al. (2010) showed the relevance of the law of less work or the tendency to minimize cognitive effort exertion in order to conserve limited cognitive resources. The study showed participant’s avoidance of cognitive demand in behavioral experiments where participants chose freely between courses of action that involved various demand levels of controlled information processing. Participants’ bias to select the less demanding choice changes according to the task incentives but were not completely accounted for by the strategic avoidance of errors, minimization of time on task, or maximization of rate of goal achievement.

Effort-based decision-making paradigms are typically carried out to determine how physical or cognitive effort affects the value of a given outcome in action choice selection (Treadway & Zald, 2011). The Effort Expenditure for Reward Task (EEfRT) has often been used in previous studies to examine effort-based decision-making (Treadway et al., 2009). The EEfRT has been used widely to study choice selections that involve varying degrees of physical effort allocation for a monetary reward (Gill et al., 2020, Treadway et al., 2009). The EEfRT can be presented as a key pressing game, where participants decide on the level of physical effort that they are willing to engage in to achieve varying monetary rewards. The reward magnitudes are usually presented with differing probability levels for reward receipt. This combination allows for examining how reward magnitude, probability of reward receipt, and expected value modulate effort-based decision-making. Studies using the EEfRT show evidence for the avoidance of high effort tasks with fixed magnitudes of reward (Kool et al., 2010). In addition, people with schizophrenia and major depression often show decreasing tendencies to select high effort options to maximize reward in the EEfRT (Barch et al., 2014, Hammar, 2009, Hartlage et al., 1993). Effort aversion is often seen in motivational disorders and particularly in neurobiological illnesses with dopaminergic dysfunction (Chong and Husain, 2016, Salamone and Correa, 2018), where effort deficiency leads to poor performance on cognitively demanding tasks without affecting performance in tasks with lower cognitive demands (Cohen et al., 2001, Hammar, 2009, Hartlage et al., 1993, Zakzanis et al., 1998).

Many studies have been carried out to identify the relationship between cognitive and physical effort. Białaszek et al. (2017) showed both physical and cognitive effort devalues rewards less in lower effort tasks and more in higher effort tasks. Moreover, the ventral striatum has been suggested by Schmidt et al. (2012) to be a common motivational node in representing the expected reward after effort exertion driving both the cognitive and the physical domains. However, not all studies have shown equivalent scaling of cognitive and physical effort. While physical effort measured through the EEfRT has been used to compare willingness to expend effort for rewards in patients with major depressive disorders and healthy controls (Treadway et al., 2012), adapted cognitive effort tasks relying on working memory have failed to discriminate patients with major depressive disorders from control groups (Tran et al., 2021). In a study comparing the EEfRT to a cognitive set-switching task, participants tend to choose the hard task more often in the cognitive task as compared to the EEfRT, even though they perceived the cognitive counterpart to be more difficult (Lopez-Gamundi & Wardle, 2018).

The issue with most of the available studies comparing decisions about physical and cognitive effort for reward is the failure in direct comparison between the physical and cognitive measures. Studies on physical effort typically measure either single sustained or multiple accumulated muscle contractions. However, such measures of physical effort is often contrasted with cognitive paradigms such as numerical Stroop tasks (Schmidt et al., 2012), memory search tasks (Ennis et al., 2013), working memory N-back tasks (Westbrook et al., 2013), set-switching tasks (Lopez-Gamundi & Wardle, 2018), and more. These tasks measure avoidance of cognitive effort as a free-choice, often measuring the ability to suppress and switch. In contrast, most physical effort measures are either repeated or single sustained muscle contractions. As an analogy, think about the difference between lifting a heavy weight one time (similar to task switching) vs. repeatedly lifting a lighter weight (similar to repeatedly pressing a key). Both tasks are effortful, but they are very different. As a result, the lack of translational and parallel measures of physical and cognitive effort costs in previous studies may prevent appropriate comparison of effort discounting between the two domains. A better alternative for tasks examining cognitive effort discounting may be related to attention maintenance, where the effort intensity is a measure of maximum voluntary effort sustained over a set temporal duration. Attention maintenance over a shorter period requires less sustained effort as compared to the same task over a longer duration, similar to the physical counterpart in EEfRT, whereby the smaller number of rapid key presses requires less sustained effort than the case when more rapid key presses are required.

In this article, we propose using a new task called the “shell game task” (SGT) as an alternative to the EEfRT. The task requires target trailing by following the movement and position of a target. The effort required in the SGT can be adjusted by changing the speed of movement, duration of movement, and the number of objects in motion. Similar to the key-pressing task in the EEfRT, participants can select a hard or easy choice as a function of the reward presented. The choice selection reflects the cognitive effort expenditure as a measure of the voluntary mobilization of cognitive resources. In the EEfRT, since the task demands varies by changing the intensity and duration of sustained effort in the form of the physical work done, the comparison of the choice offered is rather straightforward. Similarly, the task demands in the SGT varies in terms of the intensity and duration of sustained attention. Thus, the SGT may be more translational to the EEfRT.

The main goal of this experiment was to construct a cognitive effort-based reward-motivated decision-making task that could be appropriately compared to performance in the EEfRT. After creating the task (SGT), our next goal was to demonstrate that willingness to exert effort in the SGT would parallel willingness to exert physical effort in the EEfRT. To achieve this goal, we first demonstrated the presence of effort discounting in the cognitive paradigm. The effort comparison in the same domain is driven by the two levels of task demands. The difference in task demands in the same domain differed in terms of the “force” exerted over time. With consistent task demand difference throughout the paradigm, we hypothesized that participants would choose to exert more effort in both cognitive and physical paradigms when reward motivation was sufficiently high. With this, we hypothesized similar motivation discounting with effort when the willingness to exert effort was matched in the two domains. The present study focused solely on the impact of task demand differences on reward motivation, leaving out the assessment of the nature of cognitive and physical demands. In line with the idea that motivating effects of rewards can be offset by task demands or reduce the net utility of effort exertion (Apps et al., 2015, Kool et al., 2010, Shenhav et al., 2013), effort cost was computed as the turning point to choose the relatively demanding task as reward motivation increased. People choose to work less when the reward motivation was less than the effort cost, and choose to work more if the work was accompanied with reward motivation that is greater than the effort cost. Thus, the study compared the willingness to exert effort to obtain a certain reward in the physical and cognitive paradigms. If increasing task demands reduces the motivation to obtain a higher reward in both the proposed SGT and the existing EEfRT, the SGT may be deemed an appropriate alternative to the EEfRT.

Another goal of this research was to explore if the pattern of effort discounting in the two domains would hold when varying the reward-uncertainty combinations as seen in the original EEfRT. This was to show task stability with different incentive combinations. As a result, the possible changes to effort discounting in cognitive or physical paradigms were investigated according to the changes in reward-uncertainty combinations in different risk-reward conditions. However, it is also noted that subjective probabilities might not scale linearly; the subjective difference in the probability combinations in one risk condition might not be perceived as the same as in the other risk conditions, despite the same objective difference between two probabilities (Winman et al., 2014). Hence, we speculate a low possibility of changes in choice behavior with the adjusted probabilities in the risk conditions given the tendency to underestimate or overestimate probabilities or risk in the environment. The study aims can be achieved by (1) demonstrating that participants discount motivation with a similar estimated effort cost across the two tasks, (2) examining if the two measures are similarly sensitive to changes in reward motivation, and (3) testing task stability with different risk-incentive combinations. Moreover, through the comparison of the SGT and the EEfRT, this work compared the willingness to exert effort for reward in the cognitive and physical domains.