In both animals and humans, there is support for a basic distinction between survival needs and needs to invest in future benefits (Schneider, Wise, Benton, Brozek, & Keen-Rhinehart, Reference Schneider, Wise, Benton, Brozek and Keen-Rhinehart2013; Tang & West, Reference Tang and West1997). Animal research suggests that neural systems are fundamentally organized to distinguish conditions of low resources and unmet energy need from conditions of high levels of resources and met energy needs, and to regulate behavior, effort, and homeostasis accordingly. Energy acquisition and storage is an important prerequisite for reproductive success. Thus, in most species, behavioral sequences are organized so that a period of eating and fattening typically precedes a period of mating and caring for offspring. This is particularly important in habitats where food availability fluctuates in an unpredictable manner (Schneider et al., Reference Schneider, Wise, Benton, Brozek and Keen-Rhinehart2013). Perceptions of predictability and having a surplus of resources and energy shift the regulatory focus from immediate, momentary concerns and harm prevention toward future-directed behavior and long-term investments. Human evolution has taken this shift from immediate survival toward mating and caring for offspring further, exploiting environmental predictability through the development of a large neocortex and extended parental investment, facilitating the development and learning of prospective abilities (Carter, Reference Carter2014).
The different systems for behavioral control, as referred to above, are the main focus of the predictive and reactive control systems (PARCS) theory (Tops et al. Reference Tops, Quirin, Boksem and Koole2017, Reference Tops, IJzerman, Quirin and Rauthmann2021). PARCS theory proposes that people are equipped with separate neural systems for dealing with different types of environments, organized in a ventral-to-dorsal direction in corticostriatal loops and associated large-scale networks.
The relatively right-lateralized reactive control system includes the salience network (Downar, Crawley, Mikulis, & Davis, Reference Downar, Crawley, Mikulis and Davis2002), and the ventral attentional system (Shulman et al., Reference Shulman, Astafiev, Franke, Pope, Snyder, McAvoy and Corbetta2009), for example, the anterior insula (AI) and inferior frontal gyrus (IFG). Reactive control systems are for dealing with unfavorable environments that are unpredictable, unstable and novel, and for times when resources are low. When behavior is under reactive control, autonomic, homeostatic, and motor control is guided by feedback from environmental stimuli.
By contrast, predictive control systems are for dealing with predictable, familiar, and stable environments. Predictive control is guided by internally organized, model-based predictions and expectancies that are based on people's prior experiences. This control includes, among other areas, the dorsal attentional network (Shulman et al., Reference Shulman, Astafiev, Franke, Pope, Snyder, McAvoy and Corbetta2009; e.g., FEF). Predictive control can be “proactive”: impulsive but rigid, still very much driven by predicted rewards. However, especially in humans there is further development of flexible predictive control. Flexible predictive control involves the default mode network (DMN). This control is detached from the immediate environment and takes place at rest, as well as simultaneously with habitual motor control (Vatansever, Menon, & Stamatakis, Reference Vatansever, Menon and Stamatakis2017). The DMN is thought to be implicated in prospection by simulating and comparing alternative actions and outcomes (Buckner & Carroll, Reference Buckner and Carroll2007) and in rapidly selecting appropriate responses and applying learned rules under predictable behavioral contexts (Vatansever et al., Reference Vatansever, Menon and Stamatakis2017).
Our theory suggests that different brain areas should control behavior in future- versus present moment-focused ways depending on the stability and predictability of the environment. There is support from human functional magnetic resonance imaging (fMRI) studies. One study showed graded maps of time scales within the right IFG–insula and the striatum: ventroanterior regions predicted immediate rewards and dorsoposterior striatal regions (and dorsolateral prefrontal cortex, posterior cingulate cortex) predicted future rewards (Tanaka et al., Reference Tanaka, Doya, Okada, Ueda, Okamoto and Yamawaki2004). A follow-up study showed that the different learning systems in corticostriatal loops are sensitive to the predictability of the environment: the IFG–ventral striatum loop is involved in action learning based on the present state, whereas the dorsolateral prefrontal cortex–dorsal striatum loop is involved in action learning based on predictable future states (Tanaka et al., Reference Tanaka, Samejima, Okada, Ueda, Okamoto, Yamawaki and Doya2006). When subjects chose small-immediate or large-delayed liquid rewards under dietary regulation of tryptophan, a precursor of serotonin, activity of the ventral part of the striatum correlated with reward prediction at shorter time scales, and this correlated activity was stronger at low serotonin levels (Tanaka et al., Reference Tanaka, Schweighofer, Asahi, Shishida, Okamoto and Yamawaki2007). By contrast, the activity of the dorsal part of the striatum was correlated with reward prediction at longer time scales, and was stronger at high serotonin levels.
The function of serotonin may lay in its relative promotion of dorsal systems and flexible predictive control (Carver, Johnson, & Joormann, Reference Carver, Johnson and Joormann2009; Tops, Russo, Boksem, & Tucker, Reference Tops, Russo, Boksem and Tucker2009). We proposed that serotonin facilitates predictive control that guides behavior that is best performed without interference from high levels of unpredictable environmental stimulation (Tops et al., Reference Tops, Russo, Boksem and Tucker2009, Reference Tops, Luu, Boksem and Tucker2010). Serotonin may function as a neuromodulator of a drive to withdraw: a phylogenetically conserved motive to reduce the present or anticipated environmental stimulation mentally or behaviorally, such as by moving into an environment of lower stimulation levels (Tops et al., Reference Tops, Russo, Boksem and Tucker2009; cf. Lowry, Lightman, & Nutt, Reference Lowry, Lightman and Nutt2009). Serotonin increases satiety and decreases responsiveness to motivational stimuli. By increasing restraint, it allows for responding to cues of longer-term outcomes and delay of gratification (Depue, Reference Depue1995). The associations described by Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) of 5-HT with sensitivity to the history of reward, longer harvest time, decreased foraging drive, reductions in locomotor activity, but sparing of habitual movements, are consistent with facilitation of flexible predictive control: as we have seen, flexible predictive control takes place at rest or during habitual control. The latter two effects were not observed in tasks with highly motivational or highly threatening components which would trigger reactive control.
Serotonin may facilitate flexible predictive control not only in stable, predictable environments according to history of reward and reward prediction at longer time scales. Serotonin may also be sensitive to the levels of environmental and physiological resources. Tryptophan, the precursor of serotonin, is the amino acid most sensitive to depletion. Russo et al. (Reference Russo, Kema, Fokkema, Boon, Willemse, de Vries and Korf2003) proposed that tryptophan has a signaling role in physiology. The state in which tryptophan becomes depleted is associated with both external and internal unfavorable circumstances such as inflammation, stress, and food shortage. Under such conditions, chances of survival may increase by more aggressiveness and vigor in an attempt to obtain food (Russo et al., Reference Russo, Kema, Fokkema, Boon, Willemse, de Vries and Korf2003). Decreases in serotonin may facilitate such behavior by disinhibition of catecholaminergic (e.g., dopamine) systems. Conversely, high serotonin indicates sufficiency of resources.
Interestingly, environmental energy conditions have also been proposed to be central in dopamine function. According to Beeler, Frazier, and Zhuang (Reference Beeler, Frazier and Zhuang2012), the primary role of dopamine in behavior is to modulate activity such that it matches energy expenditure to the prevailing environmental energy conditions. In other words, it couples energy sensing to regulated voluntary energy expenditure. Similarly, Berke (Reference Berke2018) proposed that dopamine provides a dynamic estimate of whether it is worth expending a limited internal resource, such as energy, attention, or time.
In terms of neural structures, the AI is involved in monitoring the conditions of peripheral resources and may influence actions by signaling the adequacy of these resources (Tops, Boksem, & Koole, Reference Tops, Boksem and Koole2013, Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015; Tops & de Jong, Reference Tops and de Jong2006). Already in 1964, Gellhorn argued that neural networks of interoception are involved in a peripheral feedback mechanism on implicit motivation. Evidence suggests that the right AI integrates interoceptive information (e.g., energy levels and muscle condition) to connect motor control with feelings and motivation (Craig, Reference Craig2002; Damasio, Reference Damasio1999). The AI has been described as a critical relay between interoceptive and motor cortices, limbic motivational areas, and the orbital frontal cortex, which is thought to be involved in valuation (Carr, Iacoboni, Dubeau, Mazziotta, & Lenzi, Reference Carr, Iacoboni, Dubeau, Mazziotta and Lenzi2003). Similarly, Tinaz et al. (Reference Tinaz, Para, Vives-Rodriguez, Martinez-Kaigi, Nalamada, Sezgin and Constable2018) highlighted the interaction between the AI and dACC in generating intentional movements. Integration by the insula of sensory signals from the body and the emotional and motivational context provides the impetus to the dACC to initiate and sustain movement (cf. Tops & Boksem Reference Tops and Boksem2011; Tops et al., Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015). Neuroimaging studies using such paradigms as the sustained static handgrip exercise and extension/flexion wrist movements have shown that muscle force sense and effort sense relate to insula activity (e.g., de Graaf et al., Reference de Graaf, Gallea, Pailhous, Anton, Roth and Bonnard2004). Connecting such monitoring functions to the regulation of effort mobilization, good heartbeat perceivers showed a more finely tuned behavioral feedback-regulation of physical load than poor heartbeat perceivers (Herbert, Ulbrich, & Schandry, Reference Herbert, Ulbrich and Schandry2007).
A phenomenon often observed in sports could reflect the insula in action: clenching a fist for self-encouragement or the encouragement of others and to invigorate performance. This “encouragement gesture” may be part of a physiological feedback mechanism that functions to increase or sustain levels of vigor in challenging situations by signaling the sufficiency of resources such as muscle strength (Tops & de Jong, Reference Tops and de Jong2006). It involves the contraction of forearm flexors, similar to the grasping action that is part of acquisitive actions. Notice the paradox in this effect: in a challenging and taxing situation, effort is spent on an instrumentally useless action. The use of the action is rewarding the overcoming of challenge and facilitating persistence. This mechanism may be implicated in other examples of paradoxes of effort (Inzlicht, Shenhav, & Olivola, Reference Inzlicht, Shenhav and Olivola2018). For instance, Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) described their finding (Yoon, Geary, Ahmed, & Shadmehr, Reference Yoon, Geary, Ahmed and Shadmehr2018) that exertion of effort, apparently associated with increased dopamine, increased harvest time, and subsequent vigor.
Persistence is adaptive in stable environments with sufficient resources. In reactive control, on the contrary, exertion of effort should be subjectively effortful (i.e., aversive) to prevent loss of resources and because persistence is not adaptive in unpredictable situations. However, when the environment is stable and resources are sufficient, exertion of effort and persistence can be investments in long-term rewards and long-term resources such as skills and self-efficacy (Ainslie, Reference Ainslie2020). For instance, overcoming the challenge of putting together a piece of furniture creates an internal model of how to do this, and increases self-efficacy in this domain. The piece of furniture itself has become proof of the skill. As we have seen in the example of the encouragement gesture, exertion of effort is not always subjectively effortful, but can be rewarding. Possibly, dopamine functioning to overcome response costs (Salamone et al., Reference Salamone, Steinpreis, McCullough, Smith, Grebel and Mahan1991) could be a mechanism behind why exertion of effort is less subjectively effortful and more rewarding in predictive control than reactive control, if the former involves higher dopamine function.
Alternatively, there may be a difference between hemispheres. Support for laterality in motivation to expend effort to obtain reward was found in Parkinson's patients with asymmetric dopamine loss (Porat, Hassin-Baer, Cohen, Markus, & Tomer, Reference Porat, Hassin-Baer, Cohen, Markus and Tomer2014). Predominant left-sided loss impaired the expending of effort to increase gains whereas right-sided loss impaired the expending of effort to minimize losses. This is consistent with laterality in PARCS, given that reactive control is more concerned with preventing losses in low resource conditions whereas predictability and resources allow predictive control to invest in gains (Tops et al., Reference Tops, Quirin, Boksem and Koole2017). The reactive system is relatively right-lateralized. By contrast, the corresponding ventral system in the left hemisphere is implicated in language functions that are important in the construction of internal models in dorsal networks. This organizational pattern is also reflected in human handedness (Sainburg, Reference Sainburg2014). That is, the left hemisphere (in right handers) relies on feedforward use of vision and proprioception in control that is most effective under predictable and stable mechanical conditions, whereas the right hemisphere is specialized for impedance (i.e., online feedback-guided) control, which imparts stability when mechanical conditions are unpredictable. The left hemisphere exploits predictive processes that assure mechanical efficiency and minimize costs, such as energy and smoothness, when environmental conditions are predictable (Sainburg, Reference Sainburg2014).
We think the stable reward-acquisition (foraging) paradigms described in Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) trigger and measure mostly rigid predictive (i.e., proactive) control. They also describe shifts toward flexible predictive control in terms of increased serotonergic function. Reactive control seems to be missing. In unpredictable and urgent situations there is often over-mobilization of effort because efficient responses cannot be predicted. Moreover, specifically in reactive control, sufficiency and loss of resources are important. We probably all know from experience that external and internal conditions and resources (i.e., temperature, dimness, hunger, and sickness) impact on our subjective and motor vigor alike. Therefore, we would like to challenge Shadmehr and Ahmed to consider the role of resources in their model of vigor (cf. Boksem and Tops, Reference Boksem and Tops2008). Could resource be the common currency between reward (resource obtainment) and effort (resource mobilization/expenditure/cost)? (Tops et al., Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015).
In both animals and humans, there is support for a basic distinction between survival needs and needs to invest in future benefits (Schneider, Wise, Benton, Brozek, & Keen-Rhinehart, Reference Schneider, Wise, Benton, Brozek and Keen-Rhinehart2013; Tang & West, Reference Tang and West1997). Animal research suggests that neural systems are fundamentally organized to distinguish conditions of low resources and unmet energy need from conditions of high levels of resources and met energy needs, and to regulate behavior, effort, and homeostasis accordingly. Energy acquisition and storage is an important prerequisite for reproductive success. Thus, in most species, behavioral sequences are organized so that a period of eating and fattening typically precedes a period of mating and caring for offspring. This is particularly important in habitats where food availability fluctuates in an unpredictable manner (Schneider et al., Reference Schneider, Wise, Benton, Brozek and Keen-Rhinehart2013). Perceptions of predictability and having a surplus of resources and energy shift the regulatory focus from immediate, momentary concerns and harm prevention toward future-directed behavior and long-term investments. Human evolution has taken this shift from immediate survival toward mating and caring for offspring further, exploiting environmental predictability through the development of a large neocortex and extended parental investment, facilitating the development and learning of prospective abilities (Carter, Reference Carter2014).
The different systems for behavioral control, as referred to above, are the main focus of the predictive and reactive control systems (PARCS) theory (Tops et al. Reference Tops, Quirin, Boksem and Koole2017, Reference Tops, IJzerman, Quirin and Rauthmann2021). PARCS theory proposes that people are equipped with separate neural systems for dealing with different types of environments, organized in a ventral-to-dorsal direction in corticostriatal loops and associated large-scale networks.
The relatively right-lateralized reactive control system includes the salience network (Downar, Crawley, Mikulis, & Davis, Reference Downar, Crawley, Mikulis and Davis2002), and the ventral attentional system (Shulman et al., Reference Shulman, Astafiev, Franke, Pope, Snyder, McAvoy and Corbetta2009), for example, the anterior insula (AI) and inferior frontal gyrus (IFG). Reactive control systems are for dealing with unfavorable environments that are unpredictable, unstable and novel, and for times when resources are low. When behavior is under reactive control, autonomic, homeostatic, and motor control is guided by feedback from environmental stimuli.
By contrast, predictive control systems are for dealing with predictable, familiar, and stable environments. Predictive control is guided by internally organized, model-based predictions and expectancies that are based on people's prior experiences. This control includes, among other areas, the dorsal attentional network (Shulman et al., Reference Shulman, Astafiev, Franke, Pope, Snyder, McAvoy and Corbetta2009; e.g., FEF). Predictive control can be “proactive”: impulsive but rigid, still very much driven by predicted rewards. However, especially in humans there is further development of flexible predictive control. Flexible predictive control involves the default mode network (DMN). This control is detached from the immediate environment and takes place at rest, as well as simultaneously with habitual motor control (Vatansever, Menon, & Stamatakis, Reference Vatansever, Menon and Stamatakis2017). The DMN is thought to be implicated in prospection by simulating and comparing alternative actions and outcomes (Buckner & Carroll, Reference Buckner and Carroll2007) and in rapidly selecting appropriate responses and applying learned rules under predictable behavioral contexts (Vatansever et al., Reference Vatansever, Menon and Stamatakis2017).
Our theory suggests that different brain areas should control behavior in future- versus present moment-focused ways depending on the stability and predictability of the environment. There is support from human functional magnetic resonance imaging (fMRI) studies. One study showed graded maps of time scales within the right IFG–insula and the striatum: ventroanterior regions predicted immediate rewards and dorsoposterior striatal regions (and dorsolateral prefrontal cortex, posterior cingulate cortex) predicted future rewards (Tanaka et al., Reference Tanaka, Doya, Okada, Ueda, Okamoto and Yamawaki2004). A follow-up study showed that the different learning systems in corticostriatal loops are sensitive to the predictability of the environment: the IFG–ventral striatum loop is involved in action learning based on the present state, whereas the dorsolateral prefrontal cortex–dorsal striatum loop is involved in action learning based on predictable future states (Tanaka et al., Reference Tanaka, Samejima, Okada, Ueda, Okamoto, Yamawaki and Doya2006). When subjects chose small-immediate or large-delayed liquid rewards under dietary regulation of tryptophan, a precursor of serotonin, activity of the ventral part of the striatum correlated with reward prediction at shorter time scales, and this correlated activity was stronger at low serotonin levels (Tanaka et al., Reference Tanaka, Schweighofer, Asahi, Shishida, Okamoto and Yamawaki2007). By contrast, the activity of the dorsal part of the striatum was correlated with reward prediction at longer time scales, and was stronger at high serotonin levels.
The function of serotonin may lay in its relative promotion of dorsal systems and flexible predictive control (Carver, Johnson, & Joormann, Reference Carver, Johnson and Joormann2009; Tops, Russo, Boksem, & Tucker, Reference Tops, Russo, Boksem and Tucker2009). We proposed that serotonin facilitates predictive control that guides behavior that is best performed without interference from high levels of unpredictable environmental stimulation (Tops et al., Reference Tops, Russo, Boksem and Tucker2009, Reference Tops, Luu, Boksem and Tucker2010). Serotonin may function as a neuromodulator of a drive to withdraw: a phylogenetically conserved motive to reduce the present or anticipated environmental stimulation mentally or behaviorally, such as by moving into an environment of lower stimulation levels (Tops et al., Reference Tops, Russo, Boksem and Tucker2009; cf. Lowry, Lightman, & Nutt, Reference Lowry, Lightman and Nutt2009). Serotonin increases satiety and decreases responsiveness to motivational stimuli. By increasing restraint, it allows for responding to cues of longer-term outcomes and delay of gratification (Depue, Reference Depue1995). The associations described by Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) of 5-HT with sensitivity to the history of reward, longer harvest time, decreased foraging drive, reductions in locomotor activity, but sparing of habitual movements, are consistent with facilitation of flexible predictive control: as we have seen, flexible predictive control takes place at rest or during habitual control. The latter two effects were not observed in tasks with highly motivational or highly threatening components which would trigger reactive control.
Serotonin may facilitate flexible predictive control not only in stable, predictable environments according to history of reward and reward prediction at longer time scales. Serotonin may also be sensitive to the levels of environmental and physiological resources. Tryptophan, the precursor of serotonin, is the amino acid most sensitive to depletion. Russo et al. (Reference Russo, Kema, Fokkema, Boon, Willemse, de Vries and Korf2003) proposed that tryptophan has a signaling role in physiology. The state in which tryptophan becomes depleted is associated with both external and internal unfavorable circumstances such as inflammation, stress, and food shortage. Under such conditions, chances of survival may increase by more aggressiveness and vigor in an attempt to obtain food (Russo et al., Reference Russo, Kema, Fokkema, Boon, Willemse, de Vries and Korf2003). Decreases in serotonin may facilitate such behavior by disinhibition of catecholaminergic (e.g., dopamine) systems. Conversely, high serotonin indicates sufficiency of resources.
Interestingly, environmental energy conditions have also been proposed to be central in dopamine function. According to Beeler, Frazier, and Zhuang (Reference Beeler, Frazier and Zhuang2012), the primary role of dopamine in behavior is to modulate activity such that it matches energy expenditure to the prevailing environmental energy conditions. In other words, it couples energy sensing to regulated voluntary energy expenditure. Similarly, Berke (Reference Berke2018) proposed that dopamine provides a dynamic estimate of whether it is worth expending a limited internal resource, such as energy, attention, or time.
In terms of neural structures, the AI is involved in monitoring the conditions of peripheral resources and may influence actions by signaling the adequacy of these resources (Tops, Boksem, & Koole, Reference Tops, Boksem and Koole2013, Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015; Tops & de Jong, Reference Tops and de Jong2006). Already in 1964, Gellhorn argued that neural networks of interoception are involved in a peripheral feedback mechanism on implicit motivation. Evidence suggests that the right AI integrates interoceptive information (e.g., energy levels and muscle condition) to connect motor control with feelings and motivation (Craig, Reference Craig2002; Damasio, Reference Damasio1999). The AI has been described as a critical relay between interoceptive and motor cortices, limbic motivational areas, and the orbital frontal cortex, which is thought to be involved in valuation (Carr, Iacoboni, Dubeau, Mazziotta, & Lenzi, Reference Carr, Iacoboni, Dubeau, Mazziotta and Lenzi2003). Similarly, Tinaz et al. (Reference Tinaz, Para, Vives-Rodriguez, Martinez-Kaigi, Nalamada, Sezgin and Constable2018) highlighted the interaction between the AI and dACC in generating intentional movements. Integration by the insula of sensory signals from the body and the emotional and motivational context provides the impetus to the dACC to initiate and sustain movement (cf. Tops & Boksem Reference Tops and Boksem2011; Tops et al., Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015). Neuroimaging studies using such paradigms as the sustained static handgrip exercise and extension/flexion wrist movements have shown that muscle force sense and effort sense relate to insula activity (e.g., de Graaf et al., Reference de Graaf, Gallea, Pailhous, Anton, Roth and Bonnard2004). Connecting such monitoring functions to the regulation of effort mobilization, good heartbeat perceivers showed a more finely tuned behavioral feedback-regulation of physical load than poor heartbeat perceivers (Herbert, Ulbrich, & Schandry, Reference Herbert, Ulbrich and Schandry2007).
A phenomenon often observed in sports could reflect the insula in action: clenching a fist for self-encouragement or the encouragement of others and to invigorate performance. This “encouragement gesture” may be part of a physiological feedback mechanism that functions to increase or sustain levels of vigor in challenging situations by signaling the sufficiency of resources such as muscle strength (Tops & de Jong, Reference Tops and de Jong2006). It involves the contraction of forearm flexors, similar to the grasping action that is part of acquisitive actions. Notice the paradox in this effect: in a challenging and taxing situation, effort is spent on an instrumentally useless action. The use of the action is rewarding the overcoming of challenge and facilitating persistence. This mechanism may be implicated in other examples of paradoxes of effort (Inzlicht, Shenhav, & Olivola, Reference Inzlicht, Shenhav and Olivola2018). For instance, Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) described their finding (Yoon, Geary, Ahmed, & Shadmehr, Reference Yoon, Geary, Ahmed and Shadmehr2018) that exertion of effort, apparently associated with increased dopamine, increased harvest time, and subsequent vigor.
Persistence is adaptive in stable environments with sufficient resources. In reactive control, on the contrary, exertion of effort should be subjectively effortful (i.e., aversive) to prevent loss of resources and because persistence is not adaptive in unpredictable situations. However, when the environment is stable and resources are sufficient, exertion of effort and persistence can be investments in long-term rewards and long-term resources such as skills and self-efficacy (Ainslie, Reference Ainslie2020). For instance, overcoming the challenge of putting together a piece of furniture creates an internal model of how to do this, and increases self-efficacy in this domain. The piece of furniture itself has become proof of the skill. As we have seen in the example of the encouragement gesture, exertion of effort is not always subjectively effortful, but can be rewarding. Possibly, dopamine functioning to overcome response costs (Salamone et al., Reference Salamone, Steinpreis, McCullough, Smith, Grebel and Mahan1991) could be a mechanism behind why exertion of effort is less subjectively effortful and more rewarding in predictive control than reactive control, if the former involves higher dopamine function.
Alternatively, there may be a difference between hemispheres. Support for laterality in motivation to expend effort to obtain reward was found in Parkinson's patients with asymmetric dopamine loss (Porat, Hassin-Baer, Cohen, Markus, & Tomer, Reference Porat, Hassin-Baer, Cohen, Markus and Tomer2014). Predominant left-sided loss impaired the expending of effort to increase gains whereas right-sided loss impaired the expending of effort to minimize losses. This is consistent with laterality in PARCS, given that reactive control is more concerned with preventing losses in low resource conditions whereas predictability and resources allow predictive control to invest in gains (Tops et al., Reference Tops, Quirin, Boksem and Koole2017). The reactive system is relatively right-lateralized. By contrast, the corresponding ventral system in the left hemisphere is implicated in language functions that are important in the construction of internal models in dorsal networks. This organizational pattern is also reflected in human handedness (Sainburg, Reference Sainburg2014). That is, the left hemisphere (in right handers) relies on feedforward use of vision and proprioception in control that is most effective under predictable and stable mechanical conditions, whereas the right hemisphere is specialized for impedance (i.e., online feedback-guided) control, which imparts stability when mechanical conditions are unpredictable. The left hemisphere exploits predictive processes that assure mechanical efficiency and minimize costs, such as energy and smoothness, when environmental conditions are predictable (Sainburg, Reference Sainburg2014).
We think the stable reward-acquisition (foraging) paradigms described in Shadmehr and Ahmed (Reference Shadmehr and Ahmed2020) trigger and measure mostly rigid predictive (i.e., proactive) control. They also describe shifts toward flexible predictive control in terms of increased serotonergic function. Reactive control seems to be missing. In unpredictable and urgent situations there is often over-mobilization of effort because efficient responses cannot be predicted. Moreover, specifically in reactive control, sufficiency and loss of resources are important. We probably all know from experience that external and internal conditions and resources (i.e., temperature, dimness, hunger, and sickness) impact on our subjective and motor vigor alike. Therefore, we would like to challenge Shadmehr and Ahmed to consider the role of resources in their model of vigor (cf. Boksem and Tops, Reference Boksem and Tops2008). Could resource be the common currency between reward (resource obtainment) and effort (resource mobilization/expenditure/cost)? (Tops et al., Reference Tops, Schlinkert, Tjew, Sin, Samur, Koole, Gendolla, Tops and Koole2015).
Financial support
JMM is supported by the Wellcome Trust Grant (104908/Z/14/Z/). This research received no further specific grant from any funding agency, commercial, or not-for-profit sectors.
Conflict of interest
None.