Discussion

The principal aim of this study was to replicate the ego-depletion effect while controlling for effort engagement in depleting and control tasks with the use of psychophysiological measurements, such as electrophysiological changes in mid-frontal theta and cardiac reactivity. The ego-depletion effect did not escape the replication crisis (Alós-Ferrer et al., 2019; Hagger et al., 2016; Lurquin et al., 2016; Vohs et al., 2021; Xu et al., 2014), and testing one more time its reproducibility in more favorable conditions seems pertinent for the advancement of the discipline. Consequently, we used a protocol that already successfully obtained the ego-depletion effect (Mangin et al., 2021). This protocol included a long and effortful depleting task (i.e., a modified computerized incongruent Stroop task), a nonboring control task (i.e., a documentary viewing task) and an effortful physically dependent task (i.e., time-to-exhaustion handgrip task at 13% of MVC).
Overall, the participants in each group followed the instructions during the mental task. On the one hand, the participants who performed the Stroop task had a percentage of errors less than 12% for the reading trials and 20% for the ‘naming ink color’ trials throughout the cognitively demanding task. On the other hand, the participants who performed the documentary viewing task had a percentage of correct responses significantly greater than the chance level. In addition, the participants’ MVC significantly decreased after the time-to-exhaustion handgrip task compared to beforehand, regardless of the group and the moment of the session (before or after the mental task). This decrease in MVC reflects muscular fatigue and suggests that the participants truly contracted their forearm muscles until exhaustion. This inference is confirmed by the significant increase in the perception of pain and effort throughout the time-to-exhaustion handgrip task. It is also interesting to note that the motivation to perform the time-to-exhaustion handgrip task did not change regardless of the group and the moment of the session. Several theoretical models predict a decrease in motivation to perform an effortful task after the completion of a first effortful task (Inzlicht et al., 2014; Kurzban et al., 2013). Our experiment did not support this hypothesis.
The mentally demanding Stroop task, which requires repeated inhibitory control and cognitive flexibility, led to a faster performance drop in the subsequent dependent task. Furthermore, the subjective data showed that the Mental task performed by the experimental group was considered more difficult than that performed by the control group. In a previous study (Mangin et al., 2021), the ego-depletion effect was replicated in a within-subjects design with a large sample size (N = 55). In the current study, we observed the ego-depletion effect in a between-subjects design with a smaller sample size (N = 32; 16 in each condition). This outcome could be an argument to support the reproducibility of the ego-depletion effect when using the aforementioned experimental protocol, even with a small number of participants. In addition, a recent meta-analysis assessing the effect of mental fatigue on physical performance also indicated that between-subjects designs lead to larger effect sizes compared to within-subjects designs (Brown et al., 2020).
The results concerning mid-frontal theta confirmed that the Stroop task required more mental effort than the documentary viewing task. We observed higher theta waveband power mainly in the frontal, prefrontal and central areas, specifically at the AF7, AF3, FC5, FC3, FP2, AF4, AF8, F8, F6, FC6, C6, C4, C2 and P4 electrodes, during the Stroop task. These results were in accordance with the results that have been observed in previous studies as an index of mental effort investment during other mentally demanding tasks (Fairclough & Ewing, 2017; Puma et al., 2018).
The results of our source localization also showed that the sources of cerebral activation originated mainly from the prefrontal, frontal and central areas and, more precisely, regions close to the anterior cingulate cortex (ACC), thalamus and posterior cingulate cortex (PCC). Therefore, these results support one of the hypothesis of the integrative model of effortful control (André et al., 2019), according to which an increase in the mid-frontal theta waveband (4–7 Hz) generated by the ACC, one of the principal nodes of the salience network, is an indicator of effort engagement in a task requiring effortful control.
However, contrary to our expectations, but in agreement with the results observed by Arnau et al (2021), the results of time on task assessment of task-related theta power showed a decrease in theta power throughout the Stroop task, and regardless the type of trial (reading vs. naming ink color), suggesting a progressive disengagement of effort. In accordance with the EEG results, we also noted a gradual increase in HRV parameters (SDNN, HF & LF) toward the end of the task for the participants who performed the Stroop task. This observation suggested that parasympathetic activity was more dominant toward the end of the Stroop task, indicating that the participants were habituated to the process of the task and were performing the task more spontaneously. Surprisingly, we observed a negative correlation between the increase in reaction time for the trials that required more inhibitory control (i.e., the ‘naming ink color’ trials) and the increase in heart rate variability assessed with SDNN: the higher the increase in RT throughout the Stroop task, the lower the increase in SDNN (r = -.546; p < .05). This negative correlation suggests that participants who showed a higher decline in performance, also showed a lower decrease in parasympathetic activity. This result does not support the hypothesis of a progressive disengagement of effort throughout the Stroop task and is more compatible with the view that in parallel to the habituation to the Stroop task resulting in an increase in parasympathetic activity, participants who showed a higher decline in performance need to invest more effortful control to stay committed to the Stroop task.
The EEG results during the video condition showed a higher theta power mainly in the occipital areas, more precisely in the POz, O1, Iz, Oz, PO7 and PO8 electrodes. During this task, the participants were watching an emotionally neutral movie showing animals in their natural environment, and at the end of the task, they were asked to answer some simple questions about the content of the movie. Since this task was relatively easy and did not require much effort, the occipital areas were mainly activated to have a good perception of the visual scenes. Although other brain areas could also be activated to store pertinent information in long-term memory to answer the questions at the end of the movie, activation was not particularly observed in our analysis. The electrophysiological results of the video task mainly established less engagement of effort in the mid-frontal areas and more engagement in visual processing areas in the occipital regions. Our main focus was on the theta waveband as far as it is an index of effort, and we wanted to ensure that the control task required significantly less effort. On that account, we can confirm that the control task was less mentally demanding than the experimental task, as argued in our previous study (Mangin et al., 2021).
Overall, this study showed that it is crucial to control performance and effort engagement during depleting and control tasks to verify whether the depleting task requires more effort than the control task, and whether the participants stayed engaged throughout the whole depleting task. In the present study, although we observed a disengagement of effort during the Stroop task, we still observed the ego depletion effect.

Limits

As a limitation of this study, we first note the small sample size. Although, as mentioned earlier, we have successfully replicated the ego-depletion effect, it is difficult to generalize the results to the whole population with our limited sample size.
The dissimilarity between the experimental and control tasks can be mentioned as the second limitation of this study. In the documentary viewing task, the participants were more passive and confronted with a large variety of visual stimuli. In contrast, in the depleting Stroop task, they were more active (i.e., naming and reading aloud) and confronted with the repetition of a small number of different visual stimuli. However, it was shown in our previous study (Mangin et al., 2021) that the video task was an effective control task because it was not boring but was less effortful than the Stroop task. For instance, it has been shown that the congruent version of the Stroop task is highly boring and, in this way, taxing the self-control resources of the participants to continue the task. Finding a control task as similar as possible to the depleting task but requiring little effort and inducing little boredom is not easy.
According to the TOT results observed in the depleting task in the present study, we can view the disengagement of effort that occurred in the Stroop task as the third limitation of this study. The integrative model of effortful control (André et al., 2019) predicts that the higher and the longer the engagement in effort during the depleting task, the larger the size of the ego depletion effect. According to this prediction, if the participants of our experiments had maintained their effort throughout the depleting task, the ego-depletion effect would have been larger. Previous studies showed that knowledge of results (Sanders, 1983) and rewards (Herlambang et al., 2019) allows participants to maintain effort throughout long effortful task. Manipulating knowledge of results and/or rewards during the depleting task could be an interesting way to test this prediction.

Future perspectives

The increase in parasympathetic activity observed during the depleting task can also be interpreted as a habituation and/or automatization of the Mental task. For future studies aiming to induce an ego-depletion effect, it would be crucial to use a depleting task that is as little automatable as possible to observe a TOT effect throughout the depleting task. The beneficial effect of practice and automatization can reduce or nullify the detrimental effect of mental fatigue. During the computerized Stroop task, certain participants with short reaction times were able to rest between two trials for a few hundred milliseconds. Throughout the task, these participants could have learned to benefit from these microbreaks to invest less effort and save energy. A task constraining the participants to maintain information in working memory between two trials, such as a dual 2-back task requiring the encoding of two characteristics of each stimulus – for instance, its color and spatial location in a 4x4 matrix – could be a strategy to force participants to maintain a high level of concentration throughout the task and then reduce task automatization.
In the present study, we assessed the deployment of mental effort during a depleting task and a control task through three different indices of heart rate variability (HRV): one time-domain index (SDNN) and two frequency-domain indices (LF and HF). As a general rule, HRV indices decrease when the sympathetic nervous system is activated in response to stressors, such as performing a task that requires effortful control. In contrast, HRV indices increase when the parasympathetic nervous system is activated during resting periods (Appelhans & Luecken, 2006; Berntson et al., 1997). However, several authors have suggested that the low frequency component (LF), which highly correlates with SDNN and HF, is largely determined by the central autonomic outflow and, more particularly, the parasympathetic nervous system (Cooley et al., 1998; Reyes del Paso et al., 2013). Consequently, heart rate variability indices would be indirect, inversely related indices of sympathetic activity. More recent works have shown that another cardiac reactivity index would be a more direct index of sympathetic activation and therefore of mental effort engagement (Drost et al., 2022; Mallat et al., 2020): the pre-ejection period (PEP). This index corresponds to the time interval between the beginning of the depolarization of the ventricles (Q point on the ECG) and the ejection of blood into the aorta (B point on the impedance-cardiogram) (Brenner & Beauchaine, 2011). The shorter that the PEP is, the higher that the effort engaged in the Mental task is. Because Goedhart et al (2008) demonstrated that LF and HF did not show the expected negative correlation with PEP, it seems more appropriate to use PEP in future studies aiming to assess mental effort during the depleting, control and/or dependent task of the sequential task protocol.
Another point that can be addressed in future studies is the assessment of the event-related potential (ERP) components in addition to theta power density, such as the N2 component during a depleting task eliciting a stimulus-response conflict, which has been associated with dorsal ACC-related control processes (Cavanagh & Frank, 2014). From that perspective, variations in the amplitude of the N2 component could reflect variations in effortful control throughout the task.
In other respects, the integrative model of effortful control (André et al., 2019) assumes that the connectivity of the large-scale neuronal networks involved in effortful control, such as the salience network and the central executive network, can be weakened in long, depleting tasks. Therefore, in future studies, other brain imaging methods, such as resting-state functional magnetic resonance imaging (r-fMRI), can be used to assess the connectivity of brain networks involved in the ego-depletion effect. Several studies have used activation fMRI (e.g., Friese et al., 2013; Gergelyfi et al., 2021) to examine the pattern of activation in different brain regions during the dependent task of a sequential task protocol. However, to our knowledge, only one study has already used the sequential task protocol with the aim of examining the influence of mental fatigue on brain connectivity at rest (Esposito et al., 2014). This study reported a decrease in connectivity in the left and right frontoparietal executive attentional networks (bilateral middle frontal gyrus and right angular gyrus). These results must be confirmed by other studies varying the depleting and dependent tasks.
Finally, it could also be interesting to conduct studies with noninvasive brain stimulation methods, such as transcranial direct current stimulation (TDCs), that can assess the question of effort capacity amelioration by stimulating the ACC region of the brain. We can question whether, in the case of boosting the salience network capacity to generate the mid-frontal theta or the effort signal, it is possible to reduce the ego-depletion effect and mental fatigue. To this extent, since humans are frequently engaged in depleting Mental tasks, it would be interesting to find different countermeasures to reduce the cognitive fatigue caused by effortful tasks and then improve individuals’ performance in their daily lives.