Citation for published version (APA): Gazzola, V. (2007). Action in the brain: shared neural circuits for action observation and execution s.n.

Transcription

1 University of Groningen Action in the brain Gazzola, Valeria IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 2007 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Gazzola, V. (2007). Action in the brain: shared neural circuits for action observation and execution s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 Aplasics born without hands mirror the goal of hand actions with their feet 4 V. Gazzola, H. van der Worp, T. Mulder, B. Wicker, G. Rizzolatti, and C. Keysers Current Biology 17 (2007) The premotor and parietal mirror neuron system (MNS) is thought to contribute to the understanding of observed actions by mapping them onto corresponding motor programs of the observer [1 24], but how would the MNS respond to the observation of hand actions if the observer never had hands? Would it not show changes of blood-oxygen-level dependent (BOLD) signal, because the observer lacks motor programs that can resonate [12, 25, 26], or would it show significant changes because the observer has motor programs for the foot or mouth with corresponding goals [15, 17, 19, 27, 28]? We scanned two aplasic subjects, born without arms or hands, while they watched hand actions and compared their brain activity with that of 16 control subjects. All subjects additionally executed actions with different effectors (feet, mouth, and, for controls, hands). The BOLD signal of aplasic individuals within the putative MNS was augmented when they watched hand actions, demonstrating the brain s capacity to mirror actions that deviate from the embodiment of the observer by recruiting voxels involved in the execution of actions that achieve corresponding goals by different effectors. This sheds light on the functional organization of the MNS and predominance of goals in imitation. Results and Discussion A detailed description of the experiments can be found in the Supplemental Methods. Briefly, 2 aplasic and 16 typically developed (TD) individuals participated in two functional magnetic resonance imaging (fmri) experiments. In the first, they observed movies of hands manipulating various objects (HandAction, e.g. grasping a cocktail glass or scooping soup out of a bowl) as well as static images of the hands resting behind the same objects (HandStatic). In the second experiment, all participants were asked to manipulate an object with their lips (MouthExe), their toes (FootExe) or, for the typically developed individuals, their hands (HandExe). The execution experiment was always performed after the observation to avoid biasing visual activations towards motor areas. We initially contrasted the brain activity resulting from viewing the hand actions against the one resulting from viewing the static hand and object (HandAction-HandStatic; Fig.1). Both aplasic subjects and the typically developed (TD) controls activated, during action observation, a bilateral circuit composed mainly of frontal, parietal and temporal clusters corresponding to that reported in the literature [4,7,10-13,15,18,19,29] (Fig. 1B and Table S1; all results also survived FDR correction even at over the entire brain). No significant differences were found between the 2 aplasic subjects and the 16 TD participants (2-sample t-test, voxel-by-voxel, using a very lenient threshold of p<0.5 FDR corrected), suggesting that the visual activation in the aplasic subjects where in the range of normal variability. A graphical comparison of the activation patterns of the 16 TDs (Fig. S1) reveals substantial variation in the location and extent of visual activations. The brain activity of the two aplasic individuals fell within the range of this variability, with some TDs showing relatively less and others more activations. The classical definition of the MNS requires an overlap of brain activations related to the observation of an action and the execution of a similar action [1,11,17,18]. For the TD group we therefore inclusively masked the visual contrast HandAction-HandStatic with the brain activation elicited during the execution of hand actions (Fig. 2, lower right), finding a bilateral putative mirror circuit composed of frontal (BA6, SFG/ MFG, precg and left BA44), parietal (SI, SII, SPL and the Supramarginal Gyrus of the inferior parietal lobule) and temporal (MTG/ITG) cortices (see Tab. S3). Half of the TDs had voxels involved during observation and execution also in mesial BA6 (MNI: x=0, y=6, z=54) in locations considered to belong to the supplementary/pre-supplementary motor cortex and, supporting the idea that these regions may also be part of the human MNS[30]. In aplasic subjects masking with activation maps related to hand execution is obviously impossible. In their seminal work, Gallese and coworkers [1] described that the most frequent subtype of mirror neurons ( broadly congruent ) often responded during the execution of an action with a particular effector (e.g. grasping with the hand) and during the observation of a similar action performed by the same or a different effector (e.g. grasping with the hand or the mouth). Voxels that are active during the observation of manipulative hand actions and during the execution of manipulative foot or mouth actions should therefore also be considered to be part of the MNS. We therefore used brain activation during foot or mouth execution (Fig. 2, upper left) to examine if the vision of hand actions recruited the putative MNS in aplasic subjects. For comparison we also masked the data of the TD group with their brain activations during foot or mouth execution (Fig. 2, lower left). In all cases fronto-parieto-temporal areas were activated both during observation and execution, suggesting the existence of a MNS for observed hand actions also in aplasic subjects (Tab. S2,S3). In the same TD group, we also contrasted the vision of hand actions against the vision of meaningless hand movements not involving manipulations [28]. Results indicated that this fronto-parieto-temporal system responds significantly more to the observation of manipulation than other biological motion (hand movements not involving an object), supporting the interpretation of these areas as part of the MNS [28]. In addition, to examine if the vision of hand actions in aplasic individuals recruited regions that in TDs would respond during the execution of hand actions, we also masked the visual activations in aplasic individuals with the hand execution data of the TDs (Fig. 2, upper right). To directly compare the amplitude of mirror activations in aplas- 37

3 Chapter 4 A HandAction HandStatic 13.5 sec... ~ 3 sec ~ 3 sec... B TD 13.5 sec t-values 0 11 Figure 1. Design and Results of the Visual Experiment. (A) Experimental stimuli and design during action observation. Four video clips from the HandAction or the HandStatic category formed a block. (B) Activations during action observation. The upper four renders show the activations resulting from the contrast HandAction-HandStatic for the two aplasic subjects ( and ), and the lowest two show the activations resulting from the contrast HandAction-HandStatic for the typically developed individuals (TDs). All activations are rendered on the average anatomy of all 18 subjects (16 TDs + 2 APL, punc < and pfdr < 0.05). HandAction-HandStatic Masked FeetExe OR MouthExe A HandAction-HandStatic Masked TDHandExe B L R L R TD C D t-values 0 11 Figure 2. Putative Mirror System for Actions. The left two columns show putative hand MNS for aplasic subjects ( and ) and typically developed individuals (TD), defined by inclusively masking the visual contrast HandAction-HandStatic with their FeetExecution or MouthExecution. The right two columns show the same but defined by masking with TD s HandExecution. For -2, activations are rendered on the individual s own anatomy, and for TDs, activationsare rendered on the average anatomy of the 16TDs (punc < for the visual and motor contrast separately, and pfdr < 0.05 applied afterinclusively masking observation by execution). ics and TDs, we extracted the BOLD signal for the contrast HandAction-HandStatic in the putative mirror regions of the TDs (Fig. 3). In all regions (premotor, parietal and temporal) the contrast values of the two aplasics fell between the first and third quartile of the TDs, i.e. in these regions at least four of the TDs showed less and four more activations than the aplasic individuals. The non-parametric Mann-Whitney-U test, examining the rank order of the aplasic individual s contrasts within the distribution of all 18 subjects, identified no evidence for hypo-activation in aplasic individuals (all p>0.39). This indicates that, at least within the putative MNS, the lack of significant differences observed using the voxel-wise 2-sample t-test was not due to a lack of statistical power. By defining the mirror regions on the TDs, this test was systematically bi- 38

4 Aplasia and the mirror neuron system contrast values (HandAction-HandStatic) in arbitrary units p>0.67 TDs p>0.88 quartiles Mid-temporal Premotor Parietal p>0.39 Figure 3. Visual Activations in the Putative MNS. The render shows the location of the three regions of interest derived from Figure 2 (right bottom row, HandAction-HandStatic inclusively masked with HandExe for TDs). In each case, the right and left regions of interest were combined. For each region, the graph plots the value of the observation contrast (HandAction-HandStatic) for each TD subject as a cross in the left column and for the two aplasics as a circle and a square in the right column. The dashes in the middle column represent the first, second (median), and third quartile of the TDs. Two sided, nonparametric Mann-Whitney U test comparing the contrast values of the TDs and aplasics had probabilities of p > 0.67, p > 0.88, and p > 0.39 for the midtemporal, premotor, and parietal cluster, respectively, showing that there is no evidence for hypoactivation of the mirror system in aplasic individuals. ased in favor of the TDs, strengthening the significance of the absence of hypo-activation in the aplasics. To examine the nature of the motor programs activated by the sight of hand actions we differentiated, in both TDs and aplasics, putative mirror areas which during execution were selective for a particular effector from those that were not (see Supplemental Data Motor Decomposition and Characterization of Visual Activations and Figure S2). While viewing hand actions, both TDs and aplasics activated a combination of effector-unspecific areas and regions devoted to the effector the observer would use to perform the observed action: the hand for TDs and the foot or mouth for aplasics. Our main finding, that during the observation of hand actions both aplasic individuals robustly activated regions generally attributed to the MNS[1,4,6,8-11,15,16,18,20,21,23] and involved in the execution of foot or mouth actions, has important implications for our understanding of the MNS. As pointed out in the introduction, the MNS is generally assumed to associate observed actions with corresponding motor programs of the observer[1-24]. What though is exactly meant by corresponding? Two aspects of actions can be distinguished: their goals and their means. If I remove the cap of a fountain-pen with my mouth, my hands or my toes, the goal of the action (i.e. what is being immediately achieved) remains the same ( removing the cap ), while the means (i.e. effector and kinematics) used to achieve this goal differ. This pragmatic definition of goal does not necessarily refer to a further purpose (e.g. removing the cap to write a love letter) or sense of intentionality. Distinguishing goals and means raises the question of whether the MNS associates observed actions with (a) motor programs for corresponding goals, (b) for corresponding means, or (c) a combination of a and b. In the monkey, the MNS is composed of at least two types of mirror neurons. Strictly congruent mirror neurons (scmn), only responding to the sight of an action if it has the same goal and uses the same effector as the effective executed action and broadly congruent mirror neurons (bcmn) that also respond when the observed action involves a different effector - as long as it has the same goal. Such bcmns may respond during the execution of grasping with the hand, and during the observation of grasping with the hand (same goal, same effector) or the mouth (same goal, different effector) but not during the vision of placing with the hand (different goal, same effector [1,9]. The bcmns are approximately twice as frequent as the scmn in the monkey[1]. Jointly, in the monkey, bcmn and scmn thus associate observed actions with a combination of: actions with corresponding goals and means, and action with corresponding goals but dissimilar means. The human MNS literature, in contrast, seduced by the potential contribution of the MNS for imitation[4,7,11,22,31-39], has focused onto the capacity of the MNS to associate observed actions with motor programs corresponding in terms of both means and goals. Concepts such as direct matching [4,7] or motor resonance [22] reflect this focus, and experiments that show that the sight of actions performed with different effectors specifically recruit regions of the cortex that are involved in the execution of actions with the same effector [6,13,18,26] have fueled this focus to the point where goal matching had almost been forgotten. Our main finding that both aplasic individuals robustly activate regions involved in mouth and foot execution provides direct evidence for the potential of goal matching to recruit the putative MNS even in the absence of a matching effector. Broadly congruent MNs similar to those found in monkeys could provide the neural substrate for this goal matching. Evolutionary, what counts is achieving goals: if you are hungry, being capable of opening a nut matters, doing so using your teeth or a stone does not. Indeed, young children tend to imitate the goal not the way in which an action is performed, unless the instructions clearly ask for that[27,40-42]. In addition, while we observe other individuals, differences between the details of our bodies always introduce disparities between observed and executed actions. Overcoming such disparities, of which aplasia can be considered an extreme example, may be the evolutionary necessity that lead the MNS to also match goals of actions in an effector independent fashion. A number of recent experiments on the mirror [5,15,17,19,28] and motor systems[43] concord with this interpretation. How may the brain of aplasic individuals have developed the capacity to associate motor programs of the foot or mouth with the vision of hand actions? A speculative possibility is that aplasic individuals often interact with TDs and during joint actions the hand actions of the TDs would often occurre in synchrony with the foot and to a lesser extent the mouth actions of the aplasic individuals. This synchrony could have lead to the enhancement of Hebbian associations between the sight of hand actions and motor programs for corresponding mouth or foot actions[20,44]. A secondary finding of our study is that the amplitude of the putative MNS activations in the aplasic individuals during observation was within the range of normal variability in the TDs. If the activations of aplasic individuals within the putative MNS would have fallen within the lower quartile of the TDs activations, this lack of significant differences would have been simply attributable to a lack of statistical power. Instead, the activations of both aplasic individuals fell within the center two quartiles of the TDs range, suggesting that a lack of power was not the reason for our negative finding. By itself, this negative finding, as any negative finding, has to be interpreted with care: it does not proof that the average amplitude or spatial distribution of the visual activations within the aplasic individuals was equal to that of the TDs, but simply that the difference between these patterns was small compared to the variance between the subjects. In a separate experiment, the same TDs [28], were shown movies of (a) a human and (b) an industrial robot performing the same actions. 39

5 Chapter 4 During the vision of human actions, TDs have motor programs that match both in terms of means and goals, while during the vision of robotic actions, they only have motor programs with corresponding goals: robotic actions differ in terms of effector (robotic claw vs human hand) and kinematics from motor programs TDs would use to perform these actions. The contrast HumanAction-RobotActions should thus quantify the amount of effector specific motor programs that the putative MNS recruits during the vision of hand actions, and be conceptually similar to the contrast between aplasic individuals and TDs in the present experiment. In agreement with the aplasic data, we found robust activations to the sight of actions for which the observer had no directly matching effector (RobotAction) in regions involved in the execution of hand actions in all our 16 TDs and there was again no significant differences between the observation of actions with and without matching effector (Human - RoboticActions) within the putative MNS. Jointly, the lack of significant differences in both experiments indicates that when you observe actions of which you have achieved the goal yourself in the past, you will recruit your MNS to a similar degree whether your body includes the observed effector or not - where by similar, we mean that if differences exist, they are small compared to inter-individual differences within the population. The emphasis on goals in our study is in apparent contrast with the observation of somatotopy in the MNS[6,13,18]. Why, if goals are so important, do we activate more dorsal sectors of the premotor cortex while listening to hand actions and more ventral sectors while listening to mouth actions[18]? As shown in the monkey [1], the human MNS is likely to perform both goal and effector matching. In addition, goal matching does not preclude the recruitment of effector specific motor programs. As evidenced in the case of aplasia, effector specific motor programs (for the foot or mouth) are indeed recruited, even though the observer lacks the effector used by the observer. These recruited effector specific programs may reflect the response of bcmn that match the goal of observed actions onto whatever effector the observer would use to perform these actions. What is remarkable in aplasia, is that this most likely effector is not the one they most often see other people use to achieve the same goals. In this case, the activations therefore cannot be explained by a direct matching of observed effector onto corresponding effector (effectoreffector route), and the data suggests the presence of a route mediate by the existence of other actions that achieve the same goal (effector-goaleffector route). In all experiments on somatotopy so far, this mismatch was not present[6,13,18]: participants viewed or heard actions performed by the same effector that they would most likely use to perform these actions, and the proposed effector-goal-effector route would then activate the same effector specific motor programs as a direct effector-effector route thereby also contributing to the observed somatotopy. Experiments in which subjects observe actions performed by unusual effectors would help dissociate the contribution of the two pathways to somatotopy[18]. In the context of the role attributed to the MNS in imitation and learning[15], the idea that the observation of an action also recruits motor programs of actions with corresponding goals but differing means endows the observer with the flexibility of mapping the observed action onto the behavioral alternative that is most suited under his present circumstances. It resolves a long standing enigma: why do monkeys not imitate despite the presence of a mirror neuron system? If the MNS also matches goals on goals, one would expect individuals to learn to reach goals without necessarily imitating the way in which the goals are achieved. Recent primate studies demonstrate the fact that monkeys indeed learn to reach goals by observation[45]. True imitation then becomes a relatively rare, although sometimes important, phenomenon linked to the minority of strictly congruent neurons that closely match the details of observed and executed actions. In humans infants and possibly monkeys [46], such imitation is possible but may require appropriate instructions and training [27,40,47]. One could thus think of motor execution following learning by observation as a competition between the motor alternatives that the combination of scmn and bcmn activate during observation. Without specific instructions, this competition will lead to execution of the most economical alternative, which might often only have the goal in common with the observed action. If instructed to imitate the details of the actions, or if detailed imitation is the only way to reach the observed goal (e.g. producing intricate stone tools), this competition would be biased towards motor programs with matching means, possibly by enhancing the response of scmn. Finally these findings provide direct evidence for the long standing question of how we could comprehend actions that we never performed ourselves: contrary to what has been assumed[25], if the goal of the observed actions relate to goals that are part of our motor vocabulary we may comprehend them through the mirror of our own actions. Acknowledgments We are grateful to Dr. Tom Mulders of De Hoogstraat Rehabilitation Center in Utrecht, the Netherlands for helping us with finding and selecting the aplasia subjects. We thank Anita Kuipers for help with scanning, Michael Spezio and all members of the social brain lab for providing helpful comments. We thank Geoff Bird for suggesting the Hebbian explanation of our data as a reviewer and Karl Friston for statistical advice. The research was financed by a Marie Curie Excellence Grant and a N.W.O. Vidi Grant to CK. VG developed and conducted all experiments and CK helped in all steps. HvdW and ThM helped design and conduct the experiment on aplasic subjects. The paper was written by VG and CK with comments from the other authors. We are particularly grateful to all our subjects for their willingness to participate over a period of multiple days in our experiments. Reference List 1. Gallese V, Fadiga L, Fogassi L, Rizzolatti G (1996) Action recognition in the premotor cortex. Brain 119 (Pt 2): Fadiga L, Fogassi L, Pavesi G, Rizzolatti G (1995) Motor facilitation during action observation: a magnetic stimulation study. J Neurophysiol 73: Hari R, Forss N, Avikainen S, Kirveskari E, Salenius S, et al. (1998) Activation of human primary motor cortex during action observation: a neuromagnetic study. Proc Natl Acad Sci U S A 95: Iacoboni M, Woods RP, Brass M, Bekkering H, Mazziotta JC, et al. (1999) Cortical mechanisms of human imitation. Science 286: Umilta MA, Kohler E, Gallese V, Fogassi L, Fadiga L, et al. (2001) I know what you are doing. a neurophysiological study. Neuron 31: Buccino G, Binkofski F, Fink GR, Fadiga L, Fogassi L, et al. (2001) Action observation activates premotor and parietal areas in a somatotopic manner: an fmri study. Eur J Neurosci 13: Iacoboni M, Koski LM, Brass M, Bekkering H, Woods RP, et al. (2001) Reafferent copies of imitated actions in the right superior temporal cortex. Proc Natl Acad Sci U S A 98: Kohler E, Keysers C, Umilta MA, Fogassi L, Gallese V, et al. (2002) Hearing sounds, understanding actions: action representation in mirror neurons. Science 297: Keysers C, Kohler E, Umilta MA, Nanetti L, Fogassi L, et al. (2003) Audiovisual mirror neurons and action recognition. Exp Brain Res 153: Grezes J, Armony JL, Rowe J, Passingham RE (2003) Activations related to mirror and canonical neurones in the human brain: an fmri study. Neuroimage 18: Buccino G, Vogt S, Ritzl A, Fink GR, Zilles K, et al. (2004) Neural circuits underlying imitation learning of hand actions: an event-related fmri study. Neuron 42: Buccino G, Lui F, Canessa N, Patteri I, Lagravinese G, et al. (2004) Neural circuits involved in the recognition of actions performed by nonconspecifics: an FMRI study. J Cogn Neurosci 16: Wheaton KJ, Thompson JC, Syngeniotis A, Abbott DF, Puce A (2004) Viewing the motion of human body parts activates different regions of premotor, temporal, and parietal cortex. Neuroimage 22: Calvo-Merino B, Glaser DE, Grezes J, Passingham RE, Haggard P (2005) Action observation and acquired motor skills: an FMRI study with expert dancers. Cereb Cortex 15: Iacoboni M, Molnar-Szakacs I, Gallese V, Buccino G, Mazziotta JC, et al. (2005) Grasping the intentions of others with one s own mirror neuron system. PLoS Biol 3: e Nelissen K, Luppino G, Vanduffel W, Rizzolatti G, Orban GA (2005) Observing others: multiple action representation in the frontal lobe. Science 310: Fogassi L, Ferrari PF, Gesierich B, Rozzi S, Chersi F, et al. (2005) Parietal lobe: from action organization to intention understanding. Science 308: Gazzola V, Aziz-Zadeh L, Keysers C (2006) Empathy and the Somatotopic Auditory Mirror System in Human. Current Biology 16: Hamilton AF, Grafton ST (2006) Goal representation in human anterior intraparietal sulcus. J Neurosci 26: Keysers C, Perrett DI (2004) Demystifying social cognition: a Hebbian perspective. Trends Cogn Sci 8: Gallese V, Keysers C, Rizzolatti G (2004) A unifying view of the basis of social cognition. 40

6 Aplasia and the mirror neuron system Trends Cogn Sci 8: Rizzolatti G, Craighero L (2004) The mirror-neuron system. Annu Rev Neurosci 27: Keysers C, Gazzola V (2006) Towards a unifying neural theory of social cognition. Progress in Brain Research 156: Gallese V (2006) Intentional attunement: a neurophysiological perspective on social cognition and its disruption in autism. Brain Res 1079: Jacob P, Jeannerod M (2005) The motor theory of social cognition: a critique. Trends Cogn Sci 9: Tai YF, Scherfler C, Brooks DJ, Sawamoto N, Castiello U (2004) The human premotor cortex is mirror only for biological actions. Curr Biol 14: Bekkering H, Wohlschlager A, Gattis M (2000) Imitation of gestures in children is goaldirected. Q J Exp Psychol A 53: Gazzola V, Rizzolatti G, Wicker B, Keysers C (2007) The anthropomorphic brain: The mirror neuron system responds to human and robotic actions. Neuroimage. 29. Rizzolatti G, Fadiga L, Matelli M, Bettinardi V, Paulesu E, et al. (1996) Localization of grasp representations in humans by PET: 1. Observation versus execution. Exp Brain Res 111: Koski L, Iacoboni M, Dubeau MC, Woods RP, Mazziotta JC (2003) Modulation of cortical activity during different imitative behaviors. J Neurophysiol 89: Press C, Bird G, Flach R, Heyes C (2005) Robotic movement elicits automatic imitation. Brain Res Cogn Brain Res 25: Nishitani N, Hari R (2000) Temporal dynamics of cortical representation for action. 97: Aziz-Zadeh L, Koski L, Zaidel E, Mazziotta J, Iacoboni M (2006) Lateralization of the human mirror neuron system. J Neurosci 26: Heiser M, Iacoboni M, Maeda F, Marcus J, Mazziotta JC (2003) The essential role of Broca s area in imitation. Eur J Neurosci 17: Leslie KR, Johnson-Frey SH, Grafton ST (2004) Functional imaging of face and hand imitation: towards a motor theory of empathy. Neuroimage 21: Molnar-Szakacs I, Iacoboni M, Koski L, Mazziotta JC (2005) Functional segregation within pars opercularis of the inferior frontal gyrus: evidence from fmri studies of imitation and action observation. Cereb Cortex 15: Jackson PL, Meltzoff AN, Decety J (2006) Neural circuits involved in imitation and perspectivetaking. Neuroimage 31: Nishitani N, Hari R (2002) Viewing lip forms: cortical dynamics. Neuron 36: Oberman LM, Hubbard EM, McCleery JP, Altschuler EL, Ramachandran VS, et al. (2005) EEG evidence for mirror neuron dysfunction in autism spectrum disorders. Brain Res Cogn Brain Res 24: Gergely G, Bekkering H, Kiraly I (2002) Rational imitation in preverbal infants. Nature 415: Thompson DE, Russell J (2004) The ghost condition: imitation versus emulation in young children s observational learning. Dev Psychol 40: Nielsen M (2006) Copying actions and copying outcomes: social learning through the second year. Dev Psychol 42: Rijntjes M, Dettmers C, Buchel C, Kiebel S, Frackowiak RS, et al. (1999) A blueprint for movement: functional and anatomical representations in the human motor system. J Neurosci 19: Heyes C (2001) Causes and consequences of imitation. Trends Cogn Sci 5: Subiaul F, Cantlon JF, Holloway RL, Terrace HS (2004) Cognitive imitation in rhesus macaques. Science 305: Kumashiro M, Ishibashi H, Uchiyama Y, Itakura S, Murata A, et al. (2003) Natural imitation induced by joint attention in Japanese monkeys. Int J Psychophysiol 50: Wohlschlager A, Gattis M, Bekkering H (2003) Action generation and action perception in imitation: an instance of the ideomotor principle. Philos Trans R Soc Lond B Biol Sci 358: Supplemantal Materials and Methods Subjects and general procedures (a) Typically developed individuals (the same subjects used in Ref. 18 and 28): 16 healthy volunteers (14 right and 2 left handed; 9 female and 7 male; mean age 31yrs ranging 25-45yrs) with normal or corrected to normal vision were tested on two different days. On the first day of scanning, they viewed movies of robotic and human hand-object interactions contrasted against static controls (for space constraints, the data on the robot stimuli are presented in full details in a separate paper[28]). On the second day, they executed mouth, hand and feet actions in the scanner. (b) Aplasic Subjects. Aplasia is a defective development resulting in the absence of all or part of an organ or tissue. In our case, we searched for subjects born without hands and arms. We found only three subjects in the Netherlands showing this extremely rare form of limb aplasia but one had to be excluded from the experiment because of chronic back pain that would have made scanning too painful. The remaining two subjects were invited to participate in the study. At the time of the experiment, they were aged 32 and 33 years respectively, both held cognitively demanding jobs, were otherwise healthy and had normal or corrected to normal vision. They performed daily motor-tasks mainly by substituting their missing hands with their feet (with a predominance of the left foot) and their mouth. We tested them on two different sessions: observation of human hand-object interactions (the same movies showed during the first scanning day of the typically developed group, except for the robotic actions) and execution of mouth and feet actions in the scanner. All subjects were informed about the content of the study on a session by session basis, remaining naïve to the content of the tasks to follow. Subjects signed an informed consent. All experiments were approved by the Medical Ethical Commission (METc) of the University Medical Center Groningen (NL). Movies Movies were of 2 categories: (a) HandActions: a hand-object interaction on a table, with a hand entering to grasp and/or manipulate an objects; (b) HandStatic: a static control stimulus, in which the arm of the agent was placed behind the same objects used in (a). Each stimulus category came in 6 variants, corresponding to 6 complex objects and appropriate actions: moving a tea bag out of a cup, grasping a small coffee cup from the handle, grasping a cocktail glass by the stem, placing the lid on a sugar box, taking a spoon full of soup from a bowl, grasping and swirling a wine glass. Each example was presented in two variants: one with the right hand of the agent entering from the right, and one with the right hand entering from the left. TDs were additionally tested with the following stimuli: (c) Hand- Movement: the same hand used in HandAction entered the screen to rest on the table without engagng in any manipulation; (d) Robotic Movies: the human agent was replaced by an industrial robot for the categories HandAction, HandMovement and HandStatic[28]. Given the non-biological kinematics of the robotic agent, these stimuli enabled the quantification of the impact of differences in the way actions/movement are performed on the response of the MS. Movies were recorded using a digital video camera, elaborated using AdobePremiere ( presented using Presentation (www. neuro-bs.com), projected using an LCD projector on a semi-opaque screen placed at the head end of the bore and seen through a mirror placed on the head coil. The duration of the action movies ranged from 2.5s-4s and the static controls were presented for a matched duration. Visual experimental design All conditions were presented in a block design, with four exemplars of each condition picked out pseudo-randomly to form 13.5s blocks containing 4 different actions or 4 different static images, separated by 200ms intervals of blank screen (Fig. 1A). The order of blocks was pseudo-random and consecutive blocks were separated by a 10s pause of blank screen with a fixation cross. The experiment was split in 2 runs with a total of 12 repetitions per condition. Half of the blocks depicted only hands entering from the right of the screen, and half only hands entering from the left of the screen. Subjects were instructed to watch the movies carefully, paying particular attention to the relationship between the hands and the objects. These data were acquired prior to the action execution to avoid biasing the observation of the actions towards the motor system. Motor task Three different tasks (a-c) were performed with the typically developed individuals in three different sessions (previously described in ref. 18,28) while only 2 of them (b, c) were used with the aplasic subjects. Sessions were acquired in counterbalanced order. (a) HandExecution (typically developed group only): Subjects performed a single run of motor testing. Before scanning the subject was shown the T-shaped table that would be placed on his/her lap during scanning. The table contained 4 objects. The two lateral branches of the T contained a high-stemmed plastic glass. The intersection of the T contained a plastic bowl with a plastic spoon. The bottom of the T contained a plastic cup with a handle. Subjects were then trained on their task. The task sequence was as follows: at the commencement of each trial subjects 41

7 Chapter 4 viewed a diagram of the table on the screen, with a pink rectangle at the left or right to indicate what hand to use, and a red cross in one of the four object locations indicated which objects they had to act upon. When the red cross turned to green, subjects had to perform the action compatible with the object. For the glass, they had to reach for the glass, grasp it, bring it towards their mouth, but stop before reaching the mouth, and then replace it in its original location. For the cup of coffee, they had to do the same action, but grasping the cup by the handle. For the bowl, they had to perform the same action as above, but with the spoon, as if drinking soup with a spoon. Subject were then placed in the scanner, with their heads and lower arms firmly strapped onto the scanner bed to avoid that the actions would lead to significant head motion (in all subject within session head motion remained lower than 1mm of translation and 3deg of rotation). We ensured that subjects were unable to see their own actions and they were then trained to perform the actions blindly. The timing of the actions was rehearsed to last approximately 5s, but an experimenter within the scanner room documented the beginning and end of each action using a button box to determine the actual duration of the action, that was then used to define the design matrix for data analysis. Within a single scanning session of 500s, subjects performed eighteen ~5s actions with their right hand and 18 with their left. Their arms never crossed the table (i.e. right hand only grasped the right glass, and left hand only the left glass), and the 18 actions were composed of 6 actions involving each of the three objects. Conditions were fully randomized with 13±2s lapsing between the onset of two conditions. (b) MouthExecution (all subjects): subjects had to manipulate a small object hanging from a wooden rod by only moving their lips. The appearance of a central green cross indicated the beginning of the action whilst its disappearing, indicated the end. The experimenter lowered the rod based on acoustical instructions matched in time with the appearance of the green cross. Each single manipulation lasted for 4 sec and was repeated 16 times. (c) FeetExecution (all subjects): subjects had to manipulate a pencil or a spoon using their 1st and 2nd toe. Again the appearance of a green cross indicated the beginning of the action and its disappearing, the end. The position of the cross relative to the side of the screen (left or right) indicated the foot to be used. The experimenter received acoustical instructions with the side and the object to use. Each manipulation lasted 4 sec and was repeated 16 times for each foot (8 times the pencil, 8 the spoon for the right side and 8 times the stick, 8 the spoon for the left side). FMRI Scanning was performed using a Philips Intera 3T Quaser, a synergy SENSE head coil, 30 mt/m grandients and a standard single shot EPI with TE=30ms, TA=TR=2s, 39 axial slices of 3mm thickness, with no slice gap and a 3x3mm in plane resolution acquired to cover the entire brain and cerebellum. a Random Effect analysis. All conditions were modeled using box-car function convolved with the hemodynamic response function (hdr). Observation sessions were modeled using 2 predictors for the aplasic individuals (HandAction and HandStatic) and 4 for the TDs (HandAction, HandStatic, RobotAction, RobotStatic). Execution sessions were modeled using a single predictor for MouthExecution and 2 predictors (Left and Right) each for HandExecution and FeetExecution. Statistical Thresholding All results will be reported using a threshold of p<0.001 at the voxel level, corrected at p<0.05 for the entire brain using false discovery rate (FDR) and requiring that clusters be composed of at least 10 contiguous voxels. When masking an image with another or applying logical AND s between images, the voxel level threshold was applied separately to all images and the FDR correction and minimum voxel size requirement applied after the operation. Analyses of motor data Inspection of the realignment parameters revealed that head motion within the motor run was below one voxel and below 3deg of rotation in all subjects. Data were analyzed both with and without taking the 6 motion parameters (x,y,z translation; and rotations around the three axes) into account as covariates. There were only minimal differences between the two models and we only report the analyses with motion covariates here. (a) HandExecution. Data were analyzed using a general linear model with two experimental conditions (actions performed with the right and action performed with the left hand), modeled as box-car function convoluted with the hdr. For each subject, this resulted in separate parameter-estimates for actions of the right and left hand, that were then averaged (i.e. weight of 0.5 in the contrasts definition) together to provide a single estimate of the brain activity during hand actions and carried to the second level of analyses. (b) MouthExecution. Data were analyzed using a general linear model with a single experimental condition, modeled as box-car function convoluted with the hdr. (c) FeetExecution. As in a. Analysis of visual stimuli and mirror areas Data from each scanning session of the movies were analyzed using a general linear model with a separate predictor for the left and right versions of each stimulus category convoluted with the hemodynamic response function. Parameter estimates for the right and left versions of each category over all runs were then averaged to provide a single estimate of the brain activity for each stimulus category. Brain activity to the vision of the actions was contrasted against the static control stimulus. General data processing Data were preprocessed using SPM2 ( EPI images from all sessions were slice time corrected and aligned to the first volume of the first session of scanning. High quality T1 images were coregistered to the mean EPI image and segmented. The co-registered gray matter segment was normalized onto the gray matter template and the resulting normalization parameters applied to all EPI images. Smoothing using 6x6x6mm FWHH was applied to all normalized EPI images. Data were then analyzed by applying a general linear model separately for each individual, either voxel-by-voxel, using SPM, or to the mean signal of the voxels contained in a region of interest (ROI), using MarsBar ( The contrast values obtained for each subject were then analyzed at the second level using t-tests to implement Voxel-wise between group comparison: To compare aplasic and TDs, for each voxel of the normalized brain separately, the HandAction- HandStatic contrast values were entered in two sample (unmatched) t- test that compares the 16 TD s against the two aplasic individuals using SPM at the second level of analysis (random effect analysis). This analysis, as implemented in SPM, uses the general linear model instead of a classic 2-sample t-test. This approach calculates a t-statistics based on the comparison of the difference between the means of the two groups (i.e. HandAction-HandStatic APL vs HandAction-HandStatic TDs) with the residual error. This residual error reflects how much any subject deviates from the mean of his/her group and therefore pools the variance of the aplasic group and the TD group. This approach remains valid for unequal sample sizes in the two groups. Because it does not estimate the variance of the two groups separately, it remains valid even if one of the group 42

8 Aplasia and the mirror neuron system contains a single subject (see reference S4 and Karl Friston, SPM Mailing List, &O=D&F=&S=&X=740FB861DA7A6B8242&Y=c.keysers%40med.rug. nl&p=23636) Results of this test were first thresholded at p<0.005 uncorrected in each voxel and then corrected using false discovery rate (FDR) correction at pfdr<0.5 to correct for the increase in false positive when repeating this test for all the voxels in the brain. Given that no significant differences were observed, we repeated this analysis without any uncorrected threshold and only a very liberal threshold of p<0.5 FDR corrected, to examine for the presence of smaller differences between the groups, but still no significant differences were observed. Mirror Regions: To define putative mirror regions, the results of the contrast HandAction-HandStatic were inclusively masked with results from motor execution as detailed in the text. To be considered putatively mirror a voxel thus had to show a significant response during action observation and action execution. The visual contrast had to survive a threshold of p<0.001 uncorrected, and the motor contrast had to survive the same threshold. FDR correction at p<0.05 was then applied to the result of the inclusive masking. This analysis was performed at the level of the single subject using fixed effect analyses, and at the level of the group for the 16 TD s by entering their contrast values into a one-tailed, one-sample t-test to implement a random effect analysis. ROI analysis: to avoid the problem of multiple comparison and increase the sensitivity of the between group comparison, we defined three regions of interest in the group of typically developed individuals. These regions were obtained at the second level in the TDs using the contrast HandAction-HandStatic at p<0.001 inclusively masked with HandExecution at p<0.001, FDR corrected at p<0.05. The 4 clusters attributed to the premotor cortex (right ventral, right dorsal, left ventral and left dorsal premotor cortex) were combined into a single ROI. The two main clusters in the parietal cortex (left and right) into a single ROI, and the right and left mid-temporal clusters into the third ROI. For each of these ROIs, the BOLD signal of all the voxels within the ROI were averaged for each acquired volume, and analyzed using Marsbar with the same GLM used for the voxel-wise analysis. The contrast values were then compared using non-parametric measures (quartiles) and the Mann-Whitney U-test using the program Statistica (StatSoft, Tulsa, USA) based on the procedure of Dinneen and Blakesley [S3]. The Mann-Whitney U-test is a non-parametric test examining the rank order of the data and is a powerful test to examine data from small groups, based on the idea that if the TDs and Aplasic individuals do not differ from each other, there should be similar numbers of TDs with contrast values larger than those of the aplasics than TDs with values smaller than the Aplasic. Anatomical descriptions Anatomical description are based on the probabilistic cytoarchitectonic maps of the brain mapping group in Juelich, germany ( as implemented in the SPM anatomy toolbox ( This approach is based on probabilistic cytoarchitectonic maps of various brain areas. At present such maps do not exist for all brain areas, and outside of the identified areas, locations are described anatomically (e.g. precentral gyrus ). This means that a reference to precentral gyrus means that the activation was in a sector of the precentral gyrus that did not fall within any of the cytoarchitectonically identified areas (e.g. BA6). Supplemental Data Motor Decomposition and Characterization of Visual Activations In order to characterize the motor nature of the neural representations activated by the sight of actions, we decomposed the visually activated clusters in sub-clusters with particular motor properties. Decomposition of visual activations in aplasic individuals: We inclusively masked the HandAction-HandStatic contrast with the results of the individual s own motor (mouth/feet) execution. Figure S2A shows, on the first row, visually responsive voxels (HandAction-HandStatic, punc <0.001) that show effector preference for the feet in yellow (FeetExe-Rest, punc <0.001, AND FeetExe-MouthExe, punc<0.001) and for the mouth in red (MouthExe-Rest, punc <0.001, AND MouthExe-FeetExe, punc<0.001). On the second row, Figure S2A shows those that respond significantly and similarly during execution of actions with both effectors in green (FeetExe-Rest, punc <0.001, AND MouthExe-Rest, punc <0.001, AND FeetExe-MouthExe, punc>0.001, AND MouthExe-FeetExe, punc>0.001). The vast majority ( 80% and for 70%; Table S4) of the visual activations fell in regions common to both effectors (green area), suggesting the importance of effector indipendent motor programs in the processing of the sight of hand actions in these subjects. According to the idea of a genetical blueprint[s1, S2], one might expect the vision of hand actions to activate regions that in the group of typically developed individuals are selectively involved in HandExecution. To examine this possibility, we used the data from the TD s motor execution (HandExe-Rest, punc<0.001, AND HandExe-FeetExe, punc<0.001, AND HandExe-MouthExe, punc<0.001) to mask the visual contrast HandAction-HandStatic, with the visual activation determined at the single subject level as above, but motor execution properties defined at the group level, using the random effect analysis of the TDs. Figure S2C illustrates the fact that areas that are hand selective in TDs were indeed visually activated in the aplasic individuals. What motor function do these blue areas have in the aplasic subjects that never had an arm? To examine this question, we first compared the location of the blue areas with that of the yellow, red and green areas of the first two rows of the figure S2A. For both aplasic subjects a substantial part of the blue area fell within the green areas common to mouth and foot execution (50% for and 81% for ). For the remainder of the blue area fell mainly in the foot selective area (14%) with only a negligible part (5%) falling outside of the areas of Figure S2A. For the other 50% of the blue area fell outside of the regions of Figure S2A. Extraction of the mean signal in these remaining 50% revealed: significant responses during FeetExe (punc<0.002), marginally above our voxelwise threshold; no significant response during MouthExe (punc>0.6), and significantly larger activation during FeetExe compared to MouthExe (punc<0.05). In conclusion, the motorically hand selective mirror representation of TDs is replaced, in aplasic individuals, by a combination of effector independent and foot selective areas. The latter is in line with the fact that they tend to perform traditional hand actions mainly with their feet. This suggests some form of motor plasticity[s1,s2]: depending on the availability, or not, of a hand, these patches of cortex specialize in hand execution, or remain effectorindependent/feet-selective, respectively. Decomposition of visual activations in typically developed individuals: For comparison, the visually active voxels in the group of typically developed individuals was also split into voxels showing effector preference during execution and those responding similarly to all effectors (Fig. S2B). As can be seen from the Table S4, as for aplasics, areas common to all effectors play an important role in the mirror responses. The second most important contribution stems from areas dedicated to the efector (hand) that TDs would use to perform the observed actions. In an additional experiment [28], we compared visual activations in these same subjects while viewing robotic and human hand actions. Their was no difference (all punc>0.05) in the mean bold response in the motorically hand selective regions of Fig. S2B between the vision of robotic and human actions, suggesting that this area does not reflect a strictly congruent mapping of hand observation on hand execution, but rather 43

9 Chapter 4 the broadly congruent mapping of an observed action (be it robotic or human) on the effector most likely used by the observer to perform a similar action (hand). Conclusions: Together this data suggests that,for both aplasic and TD participants, there are two main contribution to the mirror responses to the observation of manipulative hand actions. The first contribution stems from regions common to all tested effectors. The second from region devoted to the effector the observer would use to perform the observed action: the hand for TDs and the foot/mouth for aplasics. Technical Notes: Using this approach, in the interest of simplicity, we do not visualize voxels that have intermediate motor responses, e.g. responding during execution of all effectors but with Hand>Mouth but Mouth=Feet. All thresholds were applied either at the single subject level ( or 2) or second level (TD s). The same threshold used in the main paper was used throughout (i.e uncorrected applied to all elements of a masking or logical operation, followed by FDR correction at 0.05 and an extent threshold of 10 voxels per cluster). Table S1. H andaction-handstatic (p unc <0.001, k=10) Hem. Anatomical Description k P(corr) T MNI (x,y,z) Aplasic Subject 1 R Fusiform/Cerebellum/ITG/MTG/IOG/MOG/ Hippocampus/Amygdala/Putamen/Temporal- Pole/Insula/SII/SI/M1/BA6/IPL/SPL/MCC/ PCC/SFG/MFG/Cuneus/Precuneus/SMG R MFG L IFG R MCC L MTS L ITG L Superior-Orbital-Gyrus R Mid-Orbital-Gyrus R/L Rectal-Gyrus R ACC L ITG L Hippocampus R TE L MTG Aplasic Subject 2 R MTG/MOG/STG/IPL(supramarginal) R SI/SPL/Precuneus/IPL L MTG/STG/iPL(supramarginal) L SI/SPL/IPL R MFG/SFG/Precentral/BA L MFG/SFG/Precentral/BA L MOG/BA L ITS/Fusiform R SOG/BA17-18/Cuneus R SOG/Cuneus R BA6/MCC R Fusiform Gyrus L Supramarginal Gyrus R Amygdala L IPL/SI L BA6 mesial L SPL/IPL R BA17/ L Amygdala R SFG R Precuneus L IFG L SI L Cerebellum R BA6/SMA R Hippocampus R Fusiform Gyrus L Fusiform Gyrus Typically Developed Group L SI/SPL/Precuneus/IPL/IPC/SII/STG/MTG/ ITG/Fusiform/Cerebellum R M1/SI/SPL/IPL L BA6/SFG/MFG/Precentral/BA R BA17/ R MTG/ITG R BA6/SFG/MFG/Precentral R Cerebellum R Cerebellum L SOG R/L Thalamus R SOG L MCC R BA R MCC L Cerebellum R BA17/ L MFG L Putamen R BA L Insula R Cerebellum L Cerebellum R IPL R Amygdala R IFG R Fusiform Gyrus L MFG R BA L IFG Visual Activity. Results of the contrast HandAction-HandStatic for aplasic subject 1 and 2 (single-subject analysis) and the typically developed group (random effect analysis of n = 16 subjects) thresholded at punc < and k = 10 are shown. For each cluster, from left to right, we describe the following: the hemisphere containing the cluster; the anatomical/cytoarchitectonic description of the cluster; the number (k) of mm voxels contained in the cluster; information regarding the peak voxel in the cluster, namely its fdr-corrected p and T values; and its coordinates in MNI space. The following abbreviations are used: MFG (midfrontal gyrus), IFG (inferior frontal gyrus), SFG (superior frontal gyrus), SMG (superior medial gyrus), MCC (mid-cingulate cortex), ACC (anterior cingulate cortex), PCC (posterior cingulate cortex), ITG (inferior temporal gyrus), MTG/ S (midtemporal gyrus/sulcus), TE (temporal area E), STG/S (superior temporal gyrus/sulcus), IOG (inferior occipital gyrus), MOG (midoccipital gyrus), SOG (superior occipital gyrus), SII (secondary somatosensory cortex), SI (primary somatosensory cortex), M1 (primary motor cortex), BA6 (Brodmann Area 6, premotor cortex), IPL (inferior parietal lobule), SPL (superior parietal lobule), BA17/18 (primary visual cortices), SMA (supplementary motor area), BA44 (Brodmann Area 44, premotor cortex), and BA45 (Brodmann Area 45). 44

10 Aplasia and the mirror neuron system Table S2. Results of the Contrasts HandAction-HandStatic Masked with FeetExe or MouthExe for Aplasic Subject 1 and 2 and the Typically Developed Group Hem. Anatomical Description K P(corr) T MNI (x,y,z) Aplasic Subject 1 L/R ITG, right MTG, STG, right Supramarginal, right SII, SI, BA6, M1, SPL, MFG, SFG, right BA44, right Putamen L/R BA17/ L SII,SI,BA6,BA R Fusiform/cerebellum L Putamen L Rolandic L Cerebellum R Precuneus L MTG L Cerebellum L Cerebellum,BA R MCC R Hippocampus R Insula R SOG R MTG R Thalamus Aplasic Subject 2 R STS/STG(V5/MT),Supramarginal,MTG/ MTS,ITG,MOG R SI,SPL L STS/STG(V5/MT),MTG/MTS,ITG L SI,SPL,IPL R Precentral,MFG,SFG,BA L Precentral,MFG,SFG,BA L BA17/18,MOG R SOG L Fusiform,ITG R BA L SII,Supramarginal R Fusiform R SOG L SI R Amygdala L BA R BA17/ R SFG R SI L SPL L Cerebellum L IFG R BA Typically Developed Group L SI,SPL,IPL,Postcentral,Supramarginal,SII,Pre cuneus R SI,SPL,STS,Postcentral,Supramarginal,SII,Pr ecuneus L/R BA17/ L SFG,Precentral,MFG,BA R MTG(V5/MT),ITG R SFG,MFG,BA L Cerebellum, Fusiform L Precentral, BA L/R Thalamus R Cerebellum L MTG(V5/MT),MOG L MCC R Cerebellum R BA R MCC R BA17/ L Insula L/R Pons L Cerebellum L Cerebellum HandAction-HandStatic Masked FeetExe or MouthExe (p < 0.001, k = 10). Conventions as in Table S1. Table S3. Results of the Contrasts HandAction-HandStatic Masked with HandExe of the Typically Developed Group Hem. Anatomical Description K P(corr) T MNI (x,y,z) Aplasic Subject 1 R/L ITG, right MTG, STG, right Supramarginal, right SII, SI, BA6, M1, SPL, MFG, SFG, right BA44/45, right Putamen, Precuneus, right Insula, right IFG/MFG, right Fusiform, Cerebellum, right Lingual, BA17/18, Thalamus L/R MCC,SMA/BA6,M L BA44,Rolandic,IFG,Insula R Putamen,Insula L MTG/MOG,ITG/IOG L SII,STG,Supramarginal R MTS L Cerebellum R MCC L ITG,Fusiform L Insula L BA L/R BA L MTS Aplasic Subject 2 R SI,SPL,Precuneus,IPL R MTG,ITG/S(V5/MT),STG,Supramarginal,SII L SI,Postcentral,IPL L MTG(V5MT) R MFG,Precentral,SFG,BA L MFG,Precentral,SFG R BA6/SMA L Supramarginal,SII R Cuneus,SOG L SI L Precentral R IFG/BA L BA L SI R Fusiform R Amygdala L Fusiform Typically Developed Group L SI,SPL,IPL,Postcentral,Supramarginal,SII,Pr ecuneus R SI,SPL,STS,Postcentral,Supramarginal,SII,P recuneus L SFG,Precentral,MFG,BA L/R BA17/ L MTG,MOG,ITG,V5/MT R MTG(V5/MT),ITG R SFG,MFG,BA L Cerebellum,Fusiform R Cerebellum L BA44,Precentral L Thalamus L MCC R Cerebellum R Thalamus R MCC L SOG R BA R Cerebellum L Putamen L Cerebellum L Cerebellum L Insula R STS R SOG R BA HandAction-HandStatic Masked TDHandExe (p < 0.001, k = 10). Results are shown separately for aplasic subject 1 and 2 and the typicallydeveloped group. Conventions are as shown in Table S1. TDHandExe refers to the hand action execution of the TD subjects. Table S4. Motor Decomposition of Mirror Responses in Aplasics and TDs Motor selectivity Definition (all at α unc = 0.001) Color in Fig. S2B Feet FeetExe-Rest AND FeetExe-MouthExe Yellow 6 Mouth MouthExe-Rest AND MouthExe-FeetExe Red 15 Common to Feet and Mouth MouthExe-Rest AND FeetExe-Rest AND NO significant difference between effectors Green 79 Feet FeetExe-Rest AND FeetExe-MouthExe Yellow 30 Mouth MouthExe-Rest AND MouthExe-FeetExe Red 1 Common to Feet and Mouth TD group Feet Mouth Hand Common to Feet, Mouth and Hand MouthExe-Rest AND FeetExe-Rest AND NO significant difference between effectors FeetExe-Rest AND FeetExe-MouthExe AND FeetExe-HandExe MouthExe-Rest AND MouthExe-FeetExe AND MouthExe-HandExe HandExe-Rest AND HandExe-FeetExe AND HandExe-MouthExe HandExe-Rest AND MouthExe-Rest AND FeetExe-Rest AND NO significant difference between effectors Green 70 Yellow 4 Red 4 Blue 16 Green 76 % of total volume 45

11 Chapter 4 Supplemental Figures HandAction-HandStatic S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 Figure S1. Subject-by-Subject Activations in the MNS. Activation during action observation defined subject by subject (HandAction-HandStatic; punc < 0.001, minimum of ten contiguous voxels). S1 S16 represent the TD individuals, and and represent the two aplasic subjects. All activations are rendered on the average T1 weighted anatomy of all 18 subjects to facilitate comparison. A. HandAction obs. incl. masked with Aplasics s execution L FeetExe-MouthExe B. R L MouthExe-FeetExe R MouthExe & FeetExe & FeetExe=MouthExe HandAction obs. incl. masked with TDs execution TDs HandExe-FeetExe & HandExe-MouthExe FeetExe-MouthExe & FeetExe-HandExe MouthExe-FeetExe & MouthExe-HandExe MouthExe & FeetExe & HandExe & MouthExe=FeetExe=HandExe C. APL s HandAction obs. incl. masked with TDs HandExecution HandExe-FeetExe & HandExe-MouthExe Figure S2. Motor Decomposition of Visual activations. Results of masking the visual contrasts HandAction-HandStatic inclusively with results of the motor execution of the aplasic individuals (A) or the TDs (B and C) as defined in the color legend below each panel. See Supplemental Experimental Procedures section Motor Decomposition and Characterization of Visual Activations for further details. Results are shown on the aplasic individuals normalized anatomy (A and C) and on the averaged T1 of only the 16 TDs (B) in order to provide a realistic illustration of the aplasic subjects brains. Equal signs stand for no significant difference at