Introduction

Every day, we engage in cognitive and physical activities that challenge us to move beyond our habits and require greater effort, such as learning a new language, solving a complex mathematical equation, running a marathon, manipulating information in working memory, and managing multiple priorities. Such effortful tasks require higher-order executive functions, such as self-control, inhibitory control of our thoughts, emotions, impulses, and behaviors, to reach desired goals (Baumeister et al., 2007). When self-control is weakened by a first effortful task, a drop in a subsequent effortful task is generally observed. This phenomenon is known as the ”ego-depletion effect” (Baumeister et al., 1998).
However, over the past decades, the existence of the ego-depletion effect has been challenged by meta-analytic studies and faced a replication crisis (Carter et al., 2015; Dang, 2018). For example, a worldwide replication study from 12 different laboratories (N = 1,775) showed a small effect size (d = 0.16) for the ego-depletion phenomenon (Dang et al., 2021). Another recent multilaboratory project (k = 36; N = 3,531) reported no evidence for the ego-depletion effect (d = 0.06) (Vohs et al., 2021). The ego depletion effect was first evidenced in social psychology with the sequential task protocol (Hagger et al., 2010). Then, sport sciences used the same protocol to examine the effects of mental fatigue induced by an effortful task on a subsequent physical task (Brown et al., 2020; Giboin & Wolff, 2019; Hunte et al., 2021; Van Cutsem et al., 2017).
Three theoretical approaches mainly explained the subsequent task performance decrease observed in the sequential task protocol. Based on the strength model of self-control, the first approach assumes that the ego-depletion effect occurs as the result of resource depletion, such as brain glucose (Baumeister et al., 2007). According to the second model, the ego-depletion effect is caused by the switching of motivation and/or attention toward a more pleasant task (Inzlicht et al., 2014). The third approach refers to the cost-benefit model and suggests that, when the costs to achieve the task goal are higher than the benefits associated with the achievement of the task goal, participants decrease their engagement in effortful control or drop out (Kurzban et al., 2013; Shenhav et al., 2017). In this regard, mental or cognitive fatigue can be viewed as a cost (Boksem & Tops, 2008), which has also been considered the cause of the ego-depletion effect experienced in prolonged cognitive activities (Muraven & Baumeister, 2000).
In a more recent model, the main cause of the decrease in the functional capacity to exert effortful control observed after a long and effortful task is explained by a weakening of the connectivity within and between large-scale neuronal networks underpinning effortful control, such as the salience network and the executive control network, and an accumulation of metabolites in brain regions involved in effortful control, such as the anterior cingulate cortex (André et al., 2019). Three possible metabolites have been proposed to be byproducts of neuronal activity that could decrease the capacity to exert effortful control: (1) the adenosine that operates in a negative feedback loop on neuronal activity and then decreases it (André et al., 2019; Cunha, 2001; Martin et al., 2018; Smith et al., 2019) ; (2) the glutamate, the neurotransmitter released by pyramidal neurons in cortical columns involved in cognitive control (Wiehler et al., 2022); (3) the amyloid-β (Aβ) peptides continuously secreted into the interstitial fluid by active neurons (Holroyd & Umemoto, 2016).
The integrative model of effortful control reconciles the three models cited in a previous paragraph: (1) in accordance with the strength model of self-control (Baumeister et al., 2007), it predicts that exerting effortful control leads to a progressive weakening of the capacity to exert effortful control; (2) in accordance with the process model of self-control depletion (Inzlicht et al., 2014), it conceives that ego depletion and mental fatigue can be accompanied by a decrease in motivation to exert effortful control; (3) in accordance with the cost-benefit models of effort-based decision-making (e.g., Shenhav et al., 2017), it views effortful control as costly and mental fatigue as an intrinsic cost generated by the sustained deployment of mental effort.
However, the integrative model of effortful control differs from these three models in different ways: (1) contrary to the strength model of self-control, it assumes that the duration of the effortful task and the executive control required to perform this task are two critical parameters to observe ego depletion (i.e., only long and demanding tasks induce ego depletion and mental fatigue); (2) contrary to the process model of self-control depletion, it assumes that the shift in motivation to exert effortful control is not the cause but a possible consequence of ego depletion and mental fatigue; (3) contrary to the cost-benefit models, it assumes that exerting effortful control is not intrinsically aversive.
According to the integrative model of effortful control, we assume that the replication studies aiming to examine the ego depletion effect did not use adequate protocols and that the ego-depletion effect can be replicated by considering some methodological precautions. First, according to the previous literature (Mangin et al., 2021), the depleting task must be sufficiently effortful to observe the ego-depletion effect. In general, tasks that involve core executive functions, including cognitive flexibility, updating of working memory, and inhibitory control, are good candidates for generating effortful control costs (André et al., 2019; Hofmann et al., 2012) and inducing the ego-depletion effect (Dang et al., 2021). For instance, while the choice of the letter-crossing task as a depleting task was not successful (Etherton et al., 2018; Wimmer et al., 2019; Xu et al., 2014), a modified Stroop task tapping two executive functions, cognitive flexibility and inhibitory control, was effective at inducing the ego depletion effect (Dang, 2018; Hagger et al., 2010; Mangin et al., 2021, 2022).
Second, the duration of the depleting task also plays an important role in the occurrence of the ego-depletion effect. In most of the previous ego-depletion studies, the duration of the depleting tasks was short: an average of 6.27 min (SD = 3.22 min) (Hagger et al., 2010). However, healthy subjects must exert their self-control long enough to be depleted (Blain et al., 2016). Recently, Boat et al (2020) showed that the longer the duration of the depleting task (4, 8, and 16 min), the greater the effects of deterioration on a subsequent physical task. In the same way, sport science studies using the sequential task protocol have observed that the performance of a physical task is generally degraded after a long (> 30 min) and continuous effortful cognitive activity (Van Cutsem et al., 2017). In this regard, assessing the engagement of the participants in the depleting task as a function of time on task can be helpful to examine if the task is sufficiently depleting.
Third, a recent study showed that the choice of the control task is also an essential factor in replicating the ego-depletion effect (Mangin et al., 2021). This study emphasized that the use of a boring control task, such as a congruent Stroop task, can prevent the observation of the ego-depletion effect. Performing a boring task seems as costly as a task involving executive functions and requires a certain amount of effortful control to be completed.
According to the three aforementioned points, an adequate protocol to replicate the ego-depletion effect would be a sequential task protocol that uses a 30-min effortful modified Stroop task tapping inhibitory control and cognitive flexibility as the depleting task, a time-to-exhaustion handgrip task as the dependent task, and a not boring 30-min documentary video-watching task requiring little effort as the control task.
The integrative model of effortful control also assumes that an increase in mid-frontal theta wavebands (4–7 Hz), a control signal generated by the salience network, more specifically the anterior cingulate cortex (ACC), one of its central nodes, is an indicator of effort engagement and compensatory effort. André et al. (2019) also claimed that the decrease in the capacity to exert effortful control induced by ego-depletion and mental fatigue can be observed in changes in the density of mid-frontal theta waves and performance drop throughout the depleting task. In this respect, two patterns of results induced by ego-depletion and mental fatigue can be expected during a long effortful task: (1) a decrease of performance associated with a disengagement of effortful control; (2) a stability of performance associated with an increase of effortful control (i.e., compensatory effort). These changes in theta activity throughout the depleting task should be observed with electroencephalography (EEG).
In this respect, Umemoto et al (2019) assumed that the ACC is responsible for regulating control levels by balancing the reward-related benefits of control in contrast to its effort-related costs. These authors measured reward valuation and cognitive effort with two electrophysiological indices of ACC function during a 2-hour time estimation task (i.e., participants had to estimate 1 second on every trial), respectively the reward positivity (RewP) and frontal midline theta. Reward positivity (RewP) is an event-related potential (ERP) component produced by fast, phasic midbrain dopamine reward prediction and error signals that are modulated by ACC activity (Holroyd & Coles, 2002), and it is aroused by feedback stimuli with negative versus positive valence (Miltner et al., 1997). The authors observed the participants’ performance throughout the time estimation task, giving them error and reward feedback based on their performance while their neural activities were assessed with EEG. The results showed a decrease in the reward positivity (RewP) amplitude with time on task in contrast to an increase in frontal midline theta power throughout the effortful task. According to their findings, when a long cognitively-demanding task is performed, two different phases can be distinguished. During the first phase of the task, high control levels are associated with strong reward valuation, which both contribute to significant improvements in task performance. In the later phase of the task, high control levels counteract the decrease in reward valuation, thereby helping to maintain stable task performance.
In addition, other EEG studies have also shown an increase in theta power during tasks requiring mental effort (e.g., Smit et al., 2005). Cavanagh and Frank (2014) also argued that the mid-frontal theta signal reflects diverse cognitive operations, including those involved in cognitive control. In a later study, they proposed that frontal-midline theta reflects midcingulate cortex (MCC) activity, which is involved in cognitive control and anxiety (Cavanagh & Shackman, 2015). An association of mid-frontal theta activity with different cognitive or motor sustained attention tasks is assumed to originate from the ACC (Kao et al., 2013; Onton et al., 2005; Sauseng et al., 2007). Evidence from a recent systematic review with a meta-analysis of 21 EEG studies performed on healthy individuals reported that frontal-midline theta is a robust biomarker of mental fatigue, with significant increases in theta activity in frontal, central, and posterior areas mainly observed (Tran et al., 2020). Another literature review also showed an increase in frontal midline theta wavebands as the demand for the task, alertness, arousal and mental fatigue increase (Borghini et al., 2014). We can conclude from the two reviews mentioned above that the increase in theta during a long effortful task is an index of compensatory effort to maintain the level of performance instead of mental fatigue.
In addition to mid-frontal theta assessed by EEG, other psychophysiological indices, such as pupil diameter, contraction of specific muscles, and sympathetic arousal (e.g., cardiac reactivity), can also be considered different indices of effort engagement (Shenhav et al., 2017). Heart rate variability (HRV), or the analysis of beat-to-beat intervals, is one of the most common and reliable physiological measurements used to assess mental effort (Aasman et al., 1987; Mukherjee et al., 2011; Mulder & Mulder, 1981; Veltman & Gaillard, 1993). The sympathetic and parasympathetic nervous systems influence HRV, which can vary depending on the subject’s physiological and psychological state (Ernst, 2017). During the performance of a task that requires mental effort, the previous literature has mainly reported a decrease in heart rate variability (Capa et al., 2008; De Rivecourt et al., 2008; Wang et al., 2005; Weippert et al., 2009).
Accordingly, the principal aim of this study was to replicate the ego-depletion effect while controlling for the engagement of effort during the depleting and control task and using a similar sequential task protocol as in a previous study that successfully replicated this effect (Mangin et al., 2021). The electrophysiological changes in effort engagement throughout the depleting task will be assessed through mid-frontal theta density and cardiac reactivity. We expected higher mid-frontal theta density during the depleting task than during the control task. We also predicted a progressive increase in mid-frontal theta power as a function of time on task (TOT) during the depleting task. Finally, we assumed that we would observe a decrease in HRV throughout the depleting task compared to the control task. Along with these physiological measures, we hypothesized that the participants might report a higher level of subjective fatigue and a lower motivation to perform the handgrip task after performing the mentally demanding task.