Medial frontal cortex: from self-generated action to reflection on one's own performance
It was suggested over 20 years ago that the supplementary motor cortex is involved in self-generated behaviour. Since then, there have been many studies using electrophysiology and brain imaging of the role of the supplementary motor cortex and anterior cingulate cortex. In light of the findings, the proposal that these regions are crucial for self-generated action has recently been challenged. Here, we review the recent literature and argue that the proposal survives the findings. We further argue that it can be generalised to cover reflection on mental states. Finally, we suggest that the pattern of anatomical connections is consistent with the proposal that the medial frontal cortex is crucially involved in self-generated action and self-reflection. It was suggested over 20 years ago that the supplementary motor cortex is involved in self-generated behaviour. Since then, there have been many studies using electrophysiology and brain imaging of the role of the supplementary motor cortex and anterior cingulate cortex. In light of the findings, the proposal that these regions are crucial for self-generated action has recently been challenged. Here, we review the recent literature and argue that the proposal survives the findings. We further argue that it can be generalised to cover reflection on mental states. Finally, we suggest that the pattern of anatomical connections is consistent with the proposal that the medial frontal cortex is crucially involved in self-generated action and self-reflection. The psychologist B.F. Skinner was the first to distinguish between respondent and operant behaviour. In the former, the animal responds to an external stimulus, such as a light. In the latter, the animal operates on the environment and is or is not rewarded. For example, food is presented to a dog and it comes up and eats it (respondent behaviour). Alternately, the dog comes up of its own accord and is given a scratch on the back (operant behaviour). In one case, there is a change in the environment and, in the other, the initial change is in the dog. Spontaneous actions such as this are ‘self-initiated’ or ‘self-generated’. These terms are used loosely here to include decisions both as to when to act and as to which action to make when there are no external cues to specify the appropriate action [1Passingham R.E. Lau H.C. Free choice and the human brain.in: Pockett S. Does Consciousness Cause Behavior? MIT Press, 2006: 53-72Google Scholar]. In an early study, Romo and Schultz [2Romo R. Schultz W. Neuronal activity preceding self-initiated or externally timed arm movements in area 6 of monkey cortex.Exp. Brain Res. 1987; 67: 656-662Crossref PubMed Scopus (133) Google Scholar] compared self-initiated and externally triggered movements, and reported that, on average, cells in both the supplementary motor cortex (SMA) and lateral pre-motor cortex fired earlier on self-initiated movements. Okano and Tanji [3Okano K. Tanji J. Neuronal activity in the primate motor fields of the agranular frontal cortex preceding visually triggered and self-paced movements.Exp. Brain Res. 1987; 66: 155-166Crossref PubMed Scopus (163) Google Scholar] confirmed the fact that such cells could be found in both areas. However, the authors reported that more cells in the SMA (89%) than in the pre-motor cortex (39%) fired mainly or exclusively in the self-initiated task. In the same task, ∼50% of the cells in the SMA fired well in advance of movement compared with only 12% of the cells in the pre-motor cortex. The pattern in the anterior cingulate sulcus is similar to that in the SMA [4Shima K. et al.Two movement-related foci in the primate cingulate cortex observed in signal-triggered and self-paced forelimb movements.J. Neurophysiol. 1991; 65: 188-202PubMed Google Scholar]. If this activity is essential for self-initiated movements, interfering with it should reduce or abolish such movements. Thus, monkeys whose pre-SMA and SMA were removed bilaterally, made few self-initiated movements [5Thaler D. et al.The functions of the medial premotor cortex (SMA). I. Simple learned movements.Exp. Brain Res. 1995; 102: 445-460Crossref PubMed Scopus (145) Google Scholar]. Notably, the animals were tested in the dark and, hence, there were no visual cues. Lesions to the same regions did not affect externally triggered movement: animals initiated movements relatively normally when they were required to respond to a tone [5Thaler D. et al.The functions of the medial premotor cortex (SMA). I. Simple learned movements.Exp. Brain Res. 1995; 102: 445-460Crossref PubMed Scopus (145) Google Scholar]. In the same study, lesions of the cingulate motor areas [6Dum R.P. Strick P.L. Cingulate motor areas.in: Vogt B.A. Gabriel M. Neurobiology of Cingulate Cortex and Limbic Thalamus. Birkhauser, 1993: 415-441Crossref Google Scholar] in the cingulate sulcus also led to a drastic reduction of self-initiated movements [5Thaler D. et al.The functions of the medial premotor cortex (SMA). I. Simple learned movements.Exp. Brain Res. 1995; 102: 445-460Crossref PubMed Scopus (145) Google Scholar]. These effects were selective for the medial frontal cortex: lesions of the lateral pre-motor cortex had little effect on the number of self-initiated movements [7Passingham R.E. The Frontal Lobes and Voluntary Action. Oxford University Press, 1993Google Scholar], although the movements were less accurate (reflecting perhaps a kinematic impairment) [8Kurata K. Hoffman D.S. Differential effects of muscimol microinjection into dorsal and ventral aspects of the premotor cortex of monkeys.J. Neurophysiol. 1994; 64: 1151-1164Google Scholar]. The same dissociation was shown for sequences of movement when there were no external cues to specify the order of the moves. Halsband and Passingham [9Halsband U. Higher disturbances of movement in monkeys (Macaca mulatta).in: Gantchev G.N. Motor Control. Plenum Press, 1987: 79-85Crossref Google Scholar, 10Passingham R.E. Two cortical systems for directing movements.in: Porter R. Motor Areas of the Cerebral Cortex. Wiley, 1987: 151-164Google Scholar] found that monkeys with lesions of the SMA and pre-SMA failed to learn a new sequence of three movements (pull, turn and lift) from memory. By contrast, animals with lateral pre-motor lesions could learn the task without difficulty. These and other experiments led us to conclude that the SMA and pre-SMA are particularly involved in self-generated movements [10Passingham R.E. Two cortical systems for directing movements.in: Porter R. Motor Areas of the Cerebral Cortex. Wiley, 1987: 151-164Google Scholar]. This proposal was first made by Eccles [11Eccles J.C. The initiation of voluntary movements by the supplementary motor cortex.Arch. Psychiatry. 1982; 231: 423-441Crossref Scopus (191) Google Scholar], and developed by Goldberg [12Goldberg G. Supplementary motor area structure and function: review and hypotheses.Behav. Brain Res. 1985; 8: 567-588Google Scholar, 13Goldberg G. From intent to action: evolution and function of the premotor systems of the frontal lobe.in: Perecman E. The Frontal Lobes Revisited. IRBN Press, 1987: 273-306Google Scholar] on the basis of the effects of lesions in patients. However, this proposal has recently been challenged in a commentary by Nachev et al. [14Nachev P. et al.Functional role of the supplementary and pre-supplementary motor areas.Nat. Rev. Neurosci. 2008; 9: 856-869Crossref PubMed Scopus (1255) Google Scholar]. Several more recent studies using electrophysiology and brain imaging have addressed the role of the SMA and anterior cingulate cortex. Here, we review this literature and argue that the proposal survives the findings and can be generalised to cover reflection on mental states. We also argue that the evidence for anatomical connections is consistent with the proposal. It is first necessary to clarify the proposal. The issue is not whether there are external cues that might trigger a movement, but whether these uniquely specify the response. For example, in one study by Mushiake et al. [15Mushiake H. et al.Neuronal activity in the primate premotor, supplementary, and motor cortex visually and movements.J. Neurophysiol. 1991; 66: Google Scholar] on and externally sequences of movement, there were for the monkeys to respond However, in the the did not specify the order in which the monkeys should In the externally up in The was that more cells in the SMA were in the than in the visually and more cells in the pre-motor cortex in the visually than in the in the experiments using sequences by et al. K. et al.Neuronal activity in medial frontal cortex of Neurophysiol. Google Scholar], there were for the monkeys to respond but the did not specify the order in which the animal should were cells in the pre-SMA that activity the of such this the proposal can be challenged. In the study by Mushiake et al. [15Mushiake H. et al.Neuronal activity in the primate premotor, supplementary, and motor cortex visually and movements.J. Neurophysiol. 1991; 66: Google Scholar], there were cells in both the SMA and pre-motor cortex that fired in both the and visually It be to these cells as of no with medial lesions that include the SMA an in to a perhaps as the of a of for the et and in with lesions of a medial frontal the supplementary motor 1995; PubMed Scopus Google Scholar]. However, monkeys with lesions in the SMA and pre-SMA not make more than on a visual motor task in which cues specify which action should be [7Passingham R.E. The Frontal Lobes and Voluntary Action. Oxford University Press, 1993Google Scholar]. This is in to the in monkeys with pre-motor lesions [7Passingham R.E. The Frontal Lobes and Voluntary Action. Oxford University Press, 1993Google Scholar]. the SMA or pre-SMA are with there is no effect on sequences that are by visual but a effect on sequences that are from K. Tanji J. supplementary and motor areas are crucial for the of movements.J. Neurophysiol. PubMed Scopus Google Scholar]. It is to the and findings it is that the cells in the pre-motor cortex that on sequences from the SMA between the areas G. et connections of arm in the agranular frontal cortex.Exp. Brain Res. PubMed Scopus Google Scholar]. If the pre-SMA and SMA are this abolish such activity in the pre-motor cortex and, this area could not This could be tested by in the pre-motor cortex in monkeys with a in the pre-SMA and experiments also that there is in the SMA and as well as in the cingulate motor externally triggered movements. In other of these areas is not to self-initiated movements et motor areas in self-initiated externally triggered movements with effect of movement and Neurophysiol. PubMed Scopus Google Scholar, R. et al.The and of self-initiated and movement: a study of PubMed Scopus Google Scholar]. However, it can be shown with that the in the pre-SMA and SMA earlier when the movements are self-initiated than externally triggered R. et al.The and of self-initiated and movement: a study of PubMed Scopus Google Scholar], and that this is also earlier than in the motor cortex R. et al.The and for voluntary movement: a study of the PubMed Scopus Google Scholar]. In other activity in the SMA and pre-SMA for the These experiments made no between the pre-SMA SMA and the cingulate motor areas in the anterior cingulate there is a was shown by Lau et al. H.C. et the of Neurosci. PubMed Scopus Google Scholar], used the task et of to act in to of activity The initiation of a voluntary Scopus Google Scholar], and required to the which they or the which they were first of to was an in a cingulate motor area when to movements, and an in the pre-SMA when they to H.C. et to PubMed Scopus Google Scholar]. These are consistent with the proposal that the pre-SMA SMA and cingulate motor areas are involved in self-generated In this the actions were in both However, the is between actions that are with or without external cues. The is that there are in this a self-generated task, human are to or between actions in the and the is that they This that the might as review moves. This is a in the of imaging studies of the role of the cortex in actions [1Passingham R.E. Lau H.C. Free choice and the human brain.in: Pockett S. Does Consciousness Cause Behavior? MIT Press, 2006: 53-72Google Scholar]. However, in the experiments by et al. [5Thaler D. et al.The functions of the medial premotor cortex (SMA). I. Simple learned movements.Exp. Brain Res. 1995; 102: 445-460Crossref PubMed Scopus (145) Google Scholar] on self-initiated action in there was no such on the which the animals to an external can for is not [14Nachev P. et al.Functional role of the supplementary and pre-supplementary motor areas.Nat. Rev. Neurosci. 2008; 9: 856-869Crossref PubMed Scopus (1255) Google Scholar]. This is also a in the of imaging studies on cortex. This was shown by Lau et al. H.C. et action and to the of PubMed Scopus Google Scholar], found no in in the cortex for self-generated and externally actions the were However, in the same there was in the pre-SMA for the self-generated task, the externally task was also could be that the fact that there can be actions in a self-generated task that there is a between the et al. et and anterior cingulate cortex: an 8: PubMed Scopus Google Scholar] have suggested that, in self-generated the are and that the of medial frontal cortex could be as the of Thus, there is in the anterior cingulate cortex in on the task. In these is by for the to specify a that is from the one that is by the et al. H. et and in the medial frontal PubMed Scopus Google Scholar] compared on a task with on the task. The in the pre-SMA for and the anterior cingulate cortex for the task. Nachev et al. P. et and in human medial frontal PubMed Scopus Google Scholar] required to make a to the or and by using a that a change in The was in the of the supplementary as by the for anterior to the that is for et for the supplementary fields in human with Cortex. 9: PubMed Scopus Google Scholar]. the task involved it from the task in that there were no visual et al. [14Nachev P. et al.Functional role of the supplementary and pre-supplementary motor areas.Nat. Rev. Neurosci. 2008; 9: 856-869Crossref PubMed Scopus (1255) Google Scholar] have suggested that the for action are more for self-initiated than for We that actions as by is the cues are not We also with the fact that there are for and that it is necessary to in We in the of the for self-generated action can be by the that has as in the task et of to act in to of activity The initiation of a voluntary Scopus Google Scholar], and there are cells in the pre-SMA and SMA that for et by in the and supplementary motor areas.Nat. Neurosci. PubMed Scopus Google action can as a for the action as in a of actions on the task [1Passingham R.E. Lau H.C. Free choice and the human brain.in: Pockett S. Does Consciousness Cause Behavior? MIT Press, 2006: 53-72Google Scholar], and there are cells in the pre-SMA and SMA that for K. Tanji J. Neuronal activity in the supplementary and motor areas for of movements.J. Neurophysiol. PubMed Scopus Google of the can to the action that is appropriate in that as in the arm task [5Thaler D. et al.The functions of the medial premotor cortex (SMA). I. Simple learned movements.Exp. Brain Res. 1995; 102: 445-460Crossref PubMed Scopus (145) Google change in the can to a change in the For example, one of food is by the animal on it to the animal to the action that is appropriate for the other food et J. Google Scholar]. In lesions of the medial frontal cortex this effect Lesions of medial cortex the but not the of Neurosci. PubMed Scopus Google Scholar]. In lesions in the anterior cingulate cortex the of the animals to the action that is appropriate for the of food that is presented et al.The anterior cingulate and of Neurophysiol. PubMed Scopus Google to the or can a change in the appropriate as in the motor task by et al. et al.The functions of the medial premotor cortex (SMA). The and of learned movements.Exp. Brain Res. 1995; 102: PubMed Scopus Google Scholar]. with lesions in the SMA and pre-SMA were to et al.The functions of the medial premotor cortex (SMA). The and of learned movements.Exp. Brain Res. 1995; 102: PubMed Scopus Google Scholar], as were monkeys with lesions in the anterior cingulate sulcus and cortex et and the anterior cingulate Neurosci. 9: PubMed Scopus Google Scholar], but not monkeys with pre-motor lesions [7Passingham R.E. The Frontal Lobes and Voluntary Action. Oxford University Press, 1993Google Scholar]. If there are no external cues to specify the appropriate the animal its on an of the It is a crucial that there are many cells in the anterior cingulate sulcus that the of in terms of and et in the frontal the of Neurosci. PubMed Scopus Google Scholar]. there are more such cells in that area than in the frontal cortex. the anterior cingulate cortex and frontal cortex are involved in with but both et cortex lesions in monkeys by both and Neurosci. PubMed Scopus Google Scholar, et cortex in between actions and Neurosci. 2008; PubMed Scopus Google Scholar] and imaging evidence J. et a role for cortex in Cortex. PubMed Scopus Google Scholar] the frontal cortex in learned between and and the anterior cingulate cortex in learned between actions and The anterior cingulate cortex is involved in between actions only there is no external to specify a Thus, there are cells in the anterior cingulate sulcus in monkeys K. Tanji J. for cingulate motor area cells in voluntary movement on PubMed Scopus Google Scholar] and et anterior cingulate and the of with motor Neurosci. PubMed Scopus Google Scholar] that when an action is not by the that an action be on the these cells not or less there is an external to specify the a in the anterior cingulate human were to a reduction in the these findings, it is not that there is activity in the anterior cingulate cortex when human not the as when they make et al.The basis of and the Rev. PubMed Scopus Google Scholar]. et al. et in and over the medial frontal PubMed Scopus Google Scholar] these in terms of the of whether or between and Neurosci. PubMed Scopus Google Scholar]. proposal is that the anterior cingulate cortex is involved in the of but et al. and on anterior cingulate Neurosci. PubMed Scopus Google Scholar] that this might be the more of the of et al. et and decisions in the medial frontal 8: PubMed Scopus Google Scholar] on to the functions of the anterior cingulate cortex and pre-SMA SMA by that the pre-SMA and SMA are involved in or self-generated and the anterior cingulate cortex in the of If it is a change in that as in the could to a change in The that to when human learn motor than in the cortex and the pre-SMA J. et a role for cortex in Cortex. PubMed Scopus Google Scholar]. In the human cingulate area and to the cortex area a ventral area can be in the monkey there not to be of the human dorsal cortex D. et of the human and medial PubMed Scopus Google Scholar]. is evidence that, in the human this area is also involved in the of et al. et to well to and Cortex. PubMed Scopus Google Scholar] the task, one that this was a of and the other that the task was as to the In the task were as to that was between the In both as there was in the cortex when they made However, in the that the task to be a to there was also in the dorsal cortex on The was as to reflection on own It was that the of the in the same area as the when these were required to own et to well to and Cortex. PubMed Scopus Google Scholar]. The the task It was that were of more for the the to be a of monkeys can it is that they can in this to a of no of monkeys to but can be It is that, when whether or not to the of in the dorsal cortex in a similar et al.The of and PubMed Scopus Google Scholar]. The is that in on human own For example, can whether by on the actions that have in the and on when to actions H.C. et the of Neurosci. PubMed Scopus Google Scholar] or H.C. et to PubMed Scopus Google Scholar], the are in the medial frontal cortex. a has been used to that one can the of from activity in the and medial cortex et in the human PubMed Scopus Google Scholar]. suggested by to action and of other PubMed Scopus Google Scholar], the the cortex is also when on the mental of might be that it is also involved in on own mental states. Two of the medial frontal cortex and Rev. Neurosci. PubMed Scopus Google Scholar, et al.Functional cortex a Neurosci. PubMed Scopus Google Scholar] suggest that there is a to anterior on the medial frontal with the more anterior cortex involved in as in on mental states. is a that, in an one the that other It is for this that are and et in the effects of task and motor on in the human anterior cingulate cortex: a review of 9: PubMed Scopus Google Scholar, of the anterior cingulate cortex the human frontal a Brain Res. PubMed Scopus Google Scholar] have of in the medial frontal and it is that regions the medial frontal cortex can be on a of are findings in that to be The first to from memory. are required to of actions as to there are in the as well as the anterior and cingulate cortex. The are in the medial frontal cortex when of action is with of but on the lateral when than actions D. et to the basis of Neurosci. PubMed Scopus Google Scholar, in PubMed Scopus Google Scholar]. In these might although in these experiments they were by a However, the had to for 20 which they had and the in PubMed Scopus Google Scholar]. In other they the We suggest that on the medial frontal of action one action as the for the The is that, when are as to a task, there is more activity in the medial frontal which of the of brain function: a of an PubMed Scopus Google Scholar, et cortex and mental to a of brain U. S. PubMed Scopus Google Scholar]. It has been suggested that activity in the medial frontal cortex et al.The and in U. S. PubMed Scopus Google Scholar] or et the and PubMed Scopus Google Scholar]. are with this the authors that the were in However, in a study by et al. et activity Google Scholar], the and the in the medial frontal cortex was the the number of such The is that, as et al. et on the and PubMed Scopus Google Scholar] could the or of the or other aspects of the However, the crucial issue for the is whether these cues uniquely the of the an but in there can be only one that is in there might be in such the to The role of the cortex: evidence from the effects of in a PubMed Scopus Google Scholar]. an external stimulus, such as a can a is a as to as the is It to be shown it is that self-generated and of own on on the medial frontal Passingham et al. R.E. et al.The anatomical basis of in the Rev. Neurosci. Google Scholar] have that the to the function of brain area in its that the in which the pattern of connections from that of other areas. the connections of the medial and lateral of the It is that there is only one area that a to the medial frontal and this is the cortex H. et are by connections with and by from areas in the 1994; Scopus Google Scholar]. It could be that this for the fact that there is a cingulate motor area in the monkey brain that is involved in U. Cerebral of in the Brain Res. PubMed Scopus Google Scholar, U. et in the monkey Brain Res. PubMed Scopus Google Scholar]. the medial frontal cortex no from other areas M. The of frontal in the 2008; PubMed Scopus Google Scholar]. These to the lateral and cortex. the cortex from three the lateral and frontal cortex from external R.E. et cortex and to in Action. Press, Google Scholar]. The dorsal pre-motor cortex from the cortex et and of cortical cells to three of the premotor PubMed Scopus Google Scholar], the ventral pre-motor cortex visual from the anterior cortex E. et connections of the anterior Cortex. 2008; PubMed Scopus Google Scholar]. the medial frontal cortex little the external it has a from area et of fields in the motor and cortex of monkeys.J. PubMed Scopus Google Scholar]. the same it is well the environment as the of Limbic connections of the and medial cortex in monkeys.J. 1995; PubMed Scopus Google Scholar] and, in it is with the in the monkey (Macaca PubMed Scopus Google Scholar, et of for on the between cortex and PubMed Scopus Google Scholar]. Finally, there are between the medial motor the cingulate areas and the cingulate areas that are involved in from H. et are by connections with and by from areas in the 1994; Scopus Google Scholar]. The connections suggest that the external is on the lateral they are also consistent with the proposal that the of the actions and of the animal on the medial frontal However, it is to be we are and are not we are not that the is only on the medial We for example, that the cortex also a from area and cortical connections in Cortex. PubMed Scopus Google Scholar], and the frontal cortex the Limbic connections of the and medial cortex in monkeys.J. 1995; PubMed Scopus Google Scholar]. only the of with actions and of actions with we are not that, one one operates in the brain and that the other We for example, that the SMA and the lateral pre-motor cortex are G. et connections of arm in the agranular frontal cortex.Exp. Brain Res. PubMed Scopus Google Scholar], and that cells in the areas [2Romo R. Schultz W. Neuronal activity preceding self-initiated or externally timed arm movements in area 6 of monkey cortex.Exp. Brain Res. 1987; 67: 656-662Crossref PubMed Scopus (133) Google Scholar]. This is be given that the between self-generated and externally actions is but one of for example, the task on which from Here, the as a but it not specify a the has a of to and the to the of that be this task there is both in the lateral cortex and the anterior cingulate cortex et in human brain Cortex. 1994; PubMed Scopus Google Scholar]. lesions can systems in the brain as a are only of they suggest further for The crucial are in for the to which external cues or not specify the action to the of in the medial frontal the of medial frontal cortex to the of sequences of action and, it whether is by external in the to the of of whether they are the external environment or the the to which external cues or not specify the action to the of in the medial frontal the of medial frontal cortex to the of sequences of action and, it whether is by external in the to the of of whether they are the external environment or the We and for on an early of this The was in to by and is for