Neurofunctional correlates of body-ownership and sense of agency: A meta-analytical account of self-consciousness

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

Available online at www.sciencedirect.com

ScienceDirect Journal homepage: www.elsevier.com/locate/cortex

Research Report

Neurofunctional correlates of body-ownership and sense of agency: A meta-analytical account of selfconsciousness Silvia Seghezzi a,b, Gianluigi Giannini a and Laura Zapparoli c,* a Department of Psychology and NeuroMi e Milan Centre for Neuroscience, University of Milano-Bicocca, Milano, Italy b PhD Program in Neuroscience, School of Medicine and Surgery, University of Milano-Bicocca, Milano, Italy c IRCCS Istituto Ortopedico Galeazzi, Milano, Italy

article info

abstract

Article history:

Self-consciousness consists of several dissociable experiences, including the sense of

Received 27 May 2019

ownership of one's body and the sense of agency over one's action consequences. The

Reviewed 29 July 2019

relationship between body-ownership and the sense of agency has been described by

Revised 31 July 2019

different neurocognitive models, each providing specific neurofunctional predictions. Ac-

Accepted 29 August 2019

cording to an “additive” model, the sense of agency entails body-ownership, while an

Action editor Stephen Jackson

alternative “independence” hypothesis suggests that they represent two qualitatively

Published online 19 September 2019

different processes, underpinned by distinct brain systems. We propose a third “interactive” model, arguing the interdependence between body-ownership and the sense of

Keywords:

agency: these constructs might represent different experiences with specific and exclusive

Body-ownership

brain correlates, but they also could partly overlap at the neurofunctional level. Here we

Sense of agency

test these three neurocognitive models by reviewing the available neurofunctional litera-

Motor control

ture of body-ownership and the sense of agency, with a quantitative meta-analytical

Self-awareness

approach that allowed us to compare their neural correlates statistically. We identified

Meta-analysis

(i) a body-ownership-specific network including the left inferior parietal lobule and the left extra-striate body area, (ii) a sense-of-agency-specific network including the left SMA, the left posterior insula, the right postcentral gyrus, and the right superior temporal lobe and (iii) a shared network in the left middle insula. These results provide support for the interactive neurocognitive model of bodyownership and the sense of agency. Body-ownership involves a sensory network in which multisensory inputs are integrated to be self-attributed. On the other hand, the sense of agency is specifically associated with premotor and sensory-motor areas, typically involved in generating motor predictions and in action monitoring. Finally, body-

List of abbreviations: AAL, Automatic Anatomical Labeling; ALE, Activation Likelihood Estimation; CL, Cluster; fMRI, Functional Magnetic Resonance Imaging; EBA, Extra-striate Body Area; FWE, Family-Wise Error; HC, Hierarchical Clustering; MNI, Montreal Neurological Institute; PET, Positron Emission Tomography; SMA, Supplementary motor area; TAL, Talairach. * Corresponding author. E-mail addresses: [email protected], [email protected] (L. Zapparoli). https://doi.org/10.1016/j.cortex.2019.08.018 0010-9452/© 2019 Elsevier Ltd. All rights reserved.

170

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

ownership and the sense of agency interact at the level of the left middle insula, a highlevel multisensory hub engaged in body and action awareness in general. © 2019 Elsevier Ltd. All rights reserved.

1.

Introduction

1.1. The minimal-self: body-ownership and the sense of agency Self-consciousness is a complex mental state involving several dissociable experiences. Gallagher (2000) suggested that the so-defined “minimal self”, i.e., the consciousness of oneself as an immediate subject of an experience, includes two different levels of awareness: the sense of ownership of one's body (i.e., body-ownership) and sense of agency over one's action consequence (Gallagher, 2000). Body-ownership concerns the feeling of the body as one's own (Tsakiris, Schu¨tz-Bosbach, & Gallagher, 2007), while the “sense of agency” refers to the feeling of controlling one's movements and, through them, the events in the outside world (Haggard, 2017). The sense of ownership is mainly generated by the integration of afferent signals from different sensory modalities: when visual, proprioceptive and kinesthetic information received from a limb all match, a feeling of ownership arises for this body-part (Kalckert & Ehrsson, 2012; Tsakiris, Schu¨tzBosbach, et al., 2007). Conversely, the sense of agency is generated by a complex matching process between the efferent motor commands and the afferent sensory feedbacks of the movement. It has been argued that the comparator monitoring process used for the optimal motor control may also be responsible for the sense of agency over the action (Blakemore, Wolpert, & Frith, 2002; Frith, 1987; Frith, Friston, Liddle, & Frackowiak, 1991; Haggard, 2017; Wolpert & Ghahramani, 2000). Accordingly, an internal “forward model” for action monitoring uses the so-called “efference copy” of the current motor program to make predictions of the sensory consequences of the ongoing movement. These predictions are then compared with the actual outcomes of the actions: if the predicted and estimated actual states are congruent, the sensory event is attributed to one's agency. A mismatch prevents the consequence from being assigned to one's action (Haggard, 2017).

1.2. The neurofunctional correlates of body-ownership and the sense of agency Several neuroimaging studies have investigated the neurofunctional correlates of body-ownership and the sense of agency. Body-ownership has been mainly studied through the rubber hand illusion paradigm (Botvinick & Cohen, 1998). Typically, synchronous (but not asynchronous) touches, delivered to both the participant's hidden hand and a rubber hand placed in a position congruent with the participant's body, induce a feeling of ownership over the artificial hand.

The strength of the illusion is typically investigated by means of explicit questions (e.g., questionnaire about the subjective experience of the illusion, Botvinick & Cohen, 1998) or by means indirect measures (e.g., the “proprioceptive drift”, the perceived mislocalization of the real hand's positions towards the artificial hand, Tsakiris & Haggard, 2005). Neuroimaging studies conducted by using the rubber hand illusion report specific activations at the level of the bilateral premotor cortex, the insular cortex and the parietal lobules, related with the strength of the illusion (Ehrsson, Holmes, & Passingham, 2005; Ehrsson, Spence, & Passingham, 2004; for a metaanalytical review see; Grivaz, Blanke, & Serino, 2017). A similar network, involving fronto-parietal and insular regions, resulted also consistently activated in neuroimaging studies investigating the sense of agency (for a metaanalytical review see Seghezzi, Zirone, Paulesu, & Zapparoli, 2019). The usual manipulation in these experiments consists of providing incongruent action feedbacks. Typically, participants perform hand movements and see video feedbacks showing their motor acts, which could be veridical or distorted at varying degrees. They are then requested to judge whether they are viewing their actions or not, basing their decisions either on spatial (Daprati, Wriessnegger, & Lacquaniti, 2007) or temporal (Farrer et al., 2008) features of the observed movements. Activity in the right posterior insula was correlated with the degree of match between the performed and viewed action and with the self-attribution judgment (Farrer et al., 2003). Alternatively, the intentional binding effect (Haggard, Clark, & Kalogeras, 2002), i.e., the subjective compression of the perceived time interval between a voluntary action and its effect, has been used as a putative marker of the experience of agency. One study taking advantage of this effect has highlighted the crucial role of the supplementary motor area (SMA) in the sense of agency experience (Ku¨hn, Brass, & Haggard, 2013). At the nominal level, the comparison between the neuroimaging findings concerning body-ownership and the sense of agency reveals a high degree of overlap. One possibility is that the experimental manipulations employed so far have not been able to separate the body-ownership from the sense of agency experience (Tsakiris, Longo, & Haggard, 2010). For example, neuroimaging studies addressing the sense of agency by manipulating the visual feedback to either match or mismatch the participant's manual action may confound the neurofunctional correlates of the sense of agency and bodyownership. In particular, the distortion of the visual feedback could result not only in a mismatch between the agent's action predictions and actual feedback of the movement but could also cause an inter-sensory conflict between the proprioceptive afferent signals and vision (Tsakiris et al., 2010). According to the comparator model for voluntary action

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

(Wolpert & Ghahramani, 2000), the first mismatch can be related to agency manipulation. However, the concurrent proprioceptive-visual mismatch has implications for the normal body-ownership experience. As a result, any difference between match and mismatch conditions of these experimental paradigms could reflect both agency and ownership manipulations. The remaining question on whether these two aspects of self-consciousness could also be dissociated on a finer anatomical grain represents the issue that we will address in the empirical part of this manuscript.

1.3. The relationship between body-ownership and the sense of agency Body-ownership and sense of agency are both essential for the complete feeling of self-awareness. However, it is still unclear how and to which extent they are related. In the everyday experience of voluntary action, the body-ownership experience and the sense of agency are intimately coupled and indistinguishable. Nevertheless, whereas we experience a sense of agency only for voluntary movements, we feel bodyownership as continued and omnipresent, regardless of whether the body is acting, either voluntary or involuntary. Consequently, it has been proposed that the sense of agency entails the sense of body-ownership, plus an additional experience of voluntary control (Haggard, 2017; Tsakiris, Schu¨tz-Bosbach, et al., 2007). Accordingly, the sense of agency consists of the sense of body-ownership (‘Is my body the one that is moving’) and, also, the experience of voluntary control over that body movement (‘I voluntarily made it move’). However, even if both body-ownership and agency experiences might depend on similar proprioceptive feedbacks, they are likely to be physiologically and psychologically very different. Indeed, voluntary action depends on a complex succession of preparatory cognitive-motor processes supported by the brain's frontal lobes (Haggard, 2008; Zapparoli et al., 2018). These preparatory processes contribute to the sense of agency over the action, although they are unnecessary for the body-ownership experience. As a consequence, body-ownership and sense of agency could be regarded as qualitatively different experiences, with dissociable underlying cognitive processes and neurofunctional correlates (Tsakiris et al., 2010). To account for the complex relation between bodyownership and sense of agency, Tsakiris et al. (2010)

171

proposed two different neurocognitive models, each providing specific neuroanatomical predictions. The first “additive” model posits that the sense of agency includes the sense of body-ownership. The sense of agency should always entail the sense of body-ownership, plus an additional specific component for the agency proper. Accordingly, there should be a brain network shared by both the sense of agency and the body-ownership and additional brain regions specifically related to the sense of agency (see Fig. 1A). Conversely, the second “independence” model argues that the sense of agency and body-ownership represent qualitatively different and independent experiences. Based on this model, body-ownership and the sense of agency should be associated with different and non-overlapping brain networks (see Fig. 1B). In our opinion, a third neurocognitive model might be proposed. We suggest that body-ownership and the sense of agency might be regarded as different experiences, but still closely interdependent and interacting at the neurofunctional level. This hypothesis is supported by the fact that, on the one hand, the sense of agency depends to some extent on the body-ownership experience (‘I'm the agent of this action since the body that produced it is my own’). On the other hand, however, the sense of agency represents a reliable cue to the sense of ownership itself (‘This should be my own body since I voluntarily made it move’) and influences it (Tsakiris, Prabhu, & Haggard, 2006). This “interactive” model suggests a specific neurofunctional scenario: first, there should be brain regions specifically associated with either the sense of bodyownership and the sense of agency; moreover, there should be a shared set of brain regions at the interface between the body-ownership and the agency-specific processes (see Fig. 1C).

1.4.

Aim of the study

We adopted a meta-analytical approach that allowed us to formally compare, with statistical measures, the neural features of body ownership and sense of agency. The assessment of the degree of overlap and dissociations of the neurofunctional correlates of the two constructs would be useful in disentangling their anatomical foundations and tell whether we should favor the additive model, the independence model, or the interactive model (see Fig. 1 for a graphical representation of the three models). First, we computed an activation likelihood estimation (ALE) conjunction analysis between body-ownership and the

Fig. 1 e Graphical description of the three models of interaction between body-ownership and sense of agency.

172

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

sense of agency. The absence of any overlap between the neural correlates of the two experiences would support the “independence” model, namely the idea that the feeling of body-ownership and the sense of agency are associated with different brain networks, without any common area activated by both the experiences. Conversely, the existence of a brain network shared by the two processes would encourage to further explore the hypotheses of the “additive” or the “interactive” models. The presence of additional brain regions specifically related to the sense of agency, and the absence of areas exclusively linked to body-ownership not included in the conjunction areas between them, would support the idea that the sense of agency entails the sense of body-ownership, as suggested by the “additive” model. On the other hand, the presence of brain networks specifically associated with either the feeling of body-ownership and the sense of agency, together with a shared set of brain regions at the interface between them, would favor the “interactive” model.

2.

Materials and methods

2.1.

Data collection and preparation

We interrogated the Pubmed database (www.pubmed.com) in March 2019, by using both general-domain (“fMRI” or “PET”) and specific-domain (“body-ownership” or “sense of agency”) keywords. We performed a detailed inspection of the resulting manuscripts, and we included in the current meta-analysis only PET or fMRI studies, conducted on healthy subjects, reporting results either in standardized Montreal Neurological Institute (MNI) or Talairach (TAL) coordinates. All TAL coordinates were then transformed into MNI through the Talairach to MNI (SPM) transformation implemented in the software CluB (Clustering the Brain, see below for further details) and all the subsequent analyses were performed in the MNI space. The final “body-ownership” dataset included 17 studies (Tables S1 and S2) investigating neural correlates related to self-attribution of a body part or the whole body, for a total of 205 peaks of activation. Most studies employed variations of the original rubber-hand illusion paradigm (Botvinick & Cohen, 1998), in which asynchronous stroking (Ehrsson et al., 2004), stroking in incongruent positions (Gentile, Guterstam, Brozzoli, & Ehrsson, 2013; Limanowski, Lutti, & Blankenburg, 2014), stroking a body-part in an incongruent posture (Tsakiris, Hesse, Boy, Haggard, & Fink, 2007) or detached from the body (Petkova et al., 2011) served as control conditions. We employed activation foci resulting from the simple comparison between the factor of interest and the control condition (e.g., synchronous > asynchronous stroking conditions), as well as from interactions [e.g., (synchronous congruent condition > asynchronous congruent condition) versus (synchronous incongruent condition > asynchronous incongruent condition)]. For a detailed description of the included contrasts, see Table S1. The final “sense of agency” dataset included 14 experiments (Tables S3 and S4) whose aims were to assess which

brain regions were specifically involved in the self-agency attribution, for a total of 106 peaks of activation. Most studies manipulated the visual feedback of performed movements (Daprati et al., 2007; Farrer et al., 2008) to either match or mismatch the participant's manual action and compared visuo-motor congruency conditions (self-agency condition) with visuo-motor discrepancy conditions (external-agency condition). One study took advantage of a parametric regression with the intentional binding effect by comparing active (agency condition) and passive (no-agency condition) movements conditions (Ku¨hn et al., 2013). We included activation foci resulting from the simple comparison between the factor of interest and the control condition (e.g., visuo-motor congruency > rest or visuo-motor congruency > visuo-motor discrepancy), as well as from parametric regressions (e.g., the parametric function of the BOLD response as a function of visuo-motor congruency degree). For a detailed description of the included contrasts, see Table S3. All the raw data necessary to replicate the results of this study are contained in the Supplementary Information. The final dataset (comprising both the body ownership and the sense of agency datasets) included a total of 562 participants (mean age ¼ 29.07 ± 3.79), 35 contrasts and 311 activation peaks.

2.2. Conjunction activation likelihood estimation analysis of body-ownership and sense of agency As mentioned in the introduction, we first explore the possibility to dissociate body-ownership and sense of agency by computing an activation likelihood estimation (ALE) conjunction analysis between body-ownership and sense of agency. The ALE conjunction analysis, provided by the Ginger-ALE software (Eickhoff et al., 2009; Turkeltaub et al., 2012), creates a conjunction image by using the voxel-wise minimum value of the input ALE images (see, for example, Murray, Schaer, &
, 2012). Individual meta-analyses were first conductDebbane ed for the body-ownership and the sense of agency dataset by using the Turkeltaub Non-Additive method (Turkeltaub et al., 2012). Clusters in each meta-analysis were thresholded at p < .05 family-wise error (FWE) corrected. Conjunction analysis was then carried out to determine the intersection between the meta-analyses on body-ownership and the sense of agency. We employed a cluster-size threshold of 300 mm3, such that only clusters of contiguous voxels exceeding a volume of 300 mm3 were considered (Eickhoff et al., 2009; Turkeltaub et al., 2012). The maps of the ALE values were overlaid on a ch2better.nii.gz template using MRIcron software (Rorden & Brett, 2000). The absence of any overlap between the neural correlates of the two experiences would support the idea at the basis of the “independence” model. Conversely, the existence of a brain network shared by the two constructs would suggest to further explore the hypotheses of the “additive” or the “interactive” models, by looking for specific systems associated with each of the two constructs with a hierarchical clustering approach.

173

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

2.3.

Cluster analysis and cluster composition analysis

To identify anatomically coherent regional effects specifically associated with the sense of agency and the body ownership experience, we first performed a hierarchical clustering analysis (HC) using the unique-solution clustering algorithm developed by Cattinelli et al. (Cattinelli, Valentini, Paulesu, & Borghese, 2013). This method is implemented in a suite of MATLAB (2014a MathWorks) called “CluB” (Berlingeri et al., 2019; the software can be found here: https://osf.io/4b2pc/; no customized analysis code, other than that available in the CluB toolbox, were used in the present study). It takes into account the squared Euclidian distance between each couple of foci included in the dataset; the clusters with minimal dissimilarity are then recursively merged by means of Ward's criterion (Ward, 1963), with the aim of minimizing the intracluster variability and maximizing the between-cluster variability. The spatial resolution of our analyses was set to 5 mm, corresponding to the maximum mean spatial variance within each cluster in the three directions, in line with recent metaanalytical studies performed with the same software (Devoto et al., 2018; Seghezzi et al., 2019; Zapparoli, Seghezzi, & Paulesu, 2017). The output of the HC analysis was then entered as an input for the subsequent cluster composition binomial analysis. This procedure allows a non-parametric post-hoc exploration of the composition of each cluster, providing a statistical account of the degree of association of each cluster with the levels of the factor of interest. In particular, after having extracted the proportion of peaks associated with the levels (“body-ownership” and “sense of agency”) of the factor of interest (“self-awareness”) within the whole dataset, the software computes the proportion of activation peaks belonging to either level within each cluster. Then, the proportions observed within each cluster are compared to the overall prior likelihood with a binomial test. A significant binomial test (p < .05) indicates that, for a given factor level, the proportion of activation peaks included in the specific cluster is higher than the ratio computed all over the brain. After the clustering procedure, the centroid coordinates of each resulting cluster were labeled according to the Automatic Anatomic Labeling (AAL) and then manually checked by visual inspection using the MRIcron (https://www.nitrc.org/projects/mricron, Rorden & Brett, 2000) software. The presence of additional brain regions specifically related to the sense of agency, accompanied by the absence of areas exclusively linked to the body-ownership, would

support the idea that the sense of agency includes the sense of body-ownership, as suggested by the “additive” model. On the other hand, the presence of brain networks specifically associated with both the feeling of body-ownership and the sense of agency, together with a shared set of brain regions at the interface between the two processes, would favor the “interactive” model. We employed the HC procedure in this phase of the study to overcome the disproportion of peaks associated with the two levels of the factor of interest within our source dataset (205 peaks of activation for the body ownership dataset and 106 peaks for the sense of agency dataset). The HC method has the advantage of allowing the statistical exploration of each resulting cluster (cluster composition analysis) by comparing the proportion of the foci belonging to either level of the factor of interest within each cluster with the overall distribution of foci in the whole dataset. By applying this prior likelihood estimate of the expected number of foci in any cluster under the null hypothesis, given the overall proportions of construct specific foci, we were able to efficiently handle the higher number of foci associated with the body-ownership level compared with the sense of self-agency one and still perform our inferences.

3.

Results

3.1. Conjunction activation likelihood estimation analysis of body-ownership and sense of agency Only one cluster located between the left middle insula and the left globus pallidum was activated consistently for both body-ownership and sense of agency. See Table 1 and Fig. 2.

3.2.

Cluster analysis and cluster composition analysis

The cluster analysis yielded 96 clusters (CL), containing a number of foci between 1 and 10 peaks. The mean standard deviation along the three axes was 4.25 mm (x-axis), 4.01 mm (y-axis) and 4.89 mm (z-axis). See Table S5. The cluster composition binomial analysis revealed two clusters significantly associated with body ownership and four clusters significantly related with the sense of agency. Body-ownership-specific clusters were located in the left inferior parietal lobule (CL96) and the left inferior occipital gyrus (CL39). Sense-of-agency-specific clusters were described in the left SMA (CL56), in the left posterior insula (CL76), in the

Table 1 e Results of the ALE conjunction analysis between body-ownership and the sense of agency. We report: the cluster ID, the ALE score and the centroid coordinates in the MNI stereotaxic space. Cluster ID

Brain regions (BA)

ALE score

MNI coordinates of the Weighted Center Left Hemisphere

1 1 1

Middle insula Middle insula Pallidum

.0023 .0022 .0022

Right Hemisphere

x

y

z

38 34 24

8 10 4

4 0 2

x

y

z

174

.05 6.5 3.3 9 59 14

3 .02 5.3 2.6 4.9 3 6 76

39

1

2.2 6.4 3.5 3 3 56

7

73

.02 .04 4.8 6.4 6.8 5.6 3.4 5.2 55 2 46 67

y

z x z y

40 46 10 8 96 39

a. Body-ownership Inferior parietal lobule (40) Inferior occipital gyrus (19) b. Sense of agency Supplementary motor area (6) Postcentral gyrus (3) Posterior insula Superior temporal pole (38)

x

.04

35

5

x

38

y

26

z

52

4

.6

x

y

2.1

z

5.4

5.3

.04

p value Standard deviation

Right hemisphere

MNI coordinates k Cluster ID p value Standard deviation

To deepen the relationship between body-ownership and the sense of agency, we reviewed three alternative neurocognitive models, each one providing specific neurofunctional predictions. The first “independent” model (Tsakiris et al., 2010) argues that body-ownership and the sense of agency represent two qualitatively different processes. Accordingly, this model predicts the existence of specific brain activity patterns associated with either body-ownership or the sense of agency, without any region shared by the two experiences. On the other hand, the “additive” model (Tsakiris et al., 2010) hypothesizes that the sense of agency for voluntary movements always includes the sense of body-ownership, plus an additional specific component for the sense of agency proper. Accordingly, this model predicts the existence of some shared activations between body-ownership and sense of agency and additional brain regions specifically related to the sense of agency. In our opinion, a third neurocognitive model might be hypothesized, namely an “interactive” model, suggesting the interdependence between body-ownership and the sense of agency: they would be partly different experiences, each associated with specific brain correlates, but they also would interact somehow at the neurofunctional level. Accordingly,

Left hemisphere

4.1. Neurocognitive models underpinning bodyownership and the sense of agency

MNI coordinates

In this paper, we investigated the neurofunctional correlates of body-ownership and sense of agency, by reviewing the existent neuroimaging literature with a quantitative approach. We conducted a meta-analysis based on PET/fMRI studies in order (i) to assess which brain regions consistently process body-ownership, (ii) which areas are associated with the sense of agency and (iii) whether and to which extent body-ownership and the sense of agency share common brain regions.

k

Discussion

Cluster ID

4.

MNI coordinates

right postcentral gyrus (CL35) and the right superior temporal lobe (CL14). See Table 2 and Fig. 3.

Brain regions (BA)

Fig. 2 e Results of the ALE conjunction analysis showing common consistently activated regions in bodyownership and sense of agency experiments (areas depicted in red).

Table 2 e Results of the Cluster Composition Binomial Analysis on body-ownership and the sense of agency. For each cluster, we report: the cluster ID; the number of foci falling within the cluster (k); the centroid coordinates in the MNI stereotaxic space and the standard deviation of the Euclidean distance from the centroid along the three axes; the p values associated with the binomial test.

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

175

Fig. 3 e Results of the cluster composition binomial analysis on body-ownership (yellow clusters) and the sense of agency (blue clusters) and of the ALE conjunction analysis (red clusters) between body-ownership and the sense of agency.

this model predicts the existence of specific networks for body-ownership and the sense of agency and a set of brain areas commonly activated by both the two processes. To formally test these alternative models, we performed two different analyses, each one adapted to the available data and the tested hypotheses. We first used an ALE conjunction meta-analysis, to explore the brain areas commonly associated with body-ownership and the sense of agency. This analysis revealed the existence of one cluster located between the left middle insula and the left globus pallidum significantly activated for both the two processes. On the one hand, this first meta-analytical exploration is sufficient to rule out the independent model: indeed, according to this model, no conjunction effect should have been described. On the other hand, this result is consistent with both the additive and the interactive models, as it indicates that there is at least one brain region commonly activated by body-ownership and the sense of agency. To investigate which one of the two remaining models was more likely, we conducted a cluster composition binomial analysis. This analysis allowed us to investigate whether body-ownership and sense of agency are related to specific sets of brain regions outside the conjunction areas, by functionally characterizing each cluster of spatial convergence as more strongly associated with one process than the other one. The results demonstrate that both body-ownership and the sense of agency had specific patterns of activations. Crucially, contrary to the predictions of the additive framework, the brain regions related to body-ownership were not limited to those emerging from the conjunction with the sense of agency. Taken together, our results provide support for the interactive model, since they show the existence of (i) a network located in the left middle insula shared between bodyownership and the sense of agency, and (ii) specific and exclusive sets of activations for both each of the two processes (see Fig. 3).

4.2. A brain network shared by body-ownership and the sense of agency The role of the insular cortex in high-level sensory-motor integration is widely acknowledged, as testified by functional connectivity (Cauda et al., 2011), brain stimulation (Showers & Lauer, 1961) and imaging (Kurth, Zilles, Fox, Laird, & Eickhoff, 2010) studies. In particular, the middle insula is functionally

connected with premotor, sensorimotor, supplementary motor and middle-posterior cingulate cortices, indicating its role in sensorimotor integration (Cauda et al., 2011). This region might, therefore, represent an integration hub, where different functional systems interact. On the one hand, a match between afferent inputs may contribute to the feeling of body-ownership. On the other hand, a match between the afferent feedbacks of the movement and the efferent motor commands and predictions may provide a baseline from which the sense of agency for the performed act can be generated. In line with this hypothesis, the left middle insula has also been identified as an anatomical conjunction point between the sense of agency and motor intentions, in our recent meta-analytical study (Seghezzi et al., 2019). Therefore, the middle insula could be part of a high-level sensorimotor hub underpinning the several subcomponents of action and body awareness: motor intentionality, sense of agency and body-ownership. These results appear in contrast with the only neuroimaging study manipulating the sense of agency and bodyownership in the same experimental paradigm (Tsakiris et al., 2010); indeed, the authors did not find any shared activation between body ownership and the sense of agency. However, in this study, participants were not asked to explicitly rate the body-ownership sensation and the sense of agency for their produced actions during the fMRI scans. Conversely, explicit judgments were reported only following the fMRI experiment. Moreover, no implicit measure of bodyownership and sense of agency were collected during the fMRI scans. As a consequence, no measures related to bodyownership and sense of agency can be directly related to the neuroimaging results. This methodological choice could partly account for this unexpected result.

4.3. Body-ownership and sense of agency brain-specific networks As mentioned, our results also show the existence of specific neurofunctional underpinnings for body-ownership and the sense of agency. Body-ownership-specific clusters were located in the left inferior parietal lobule and the left inferior occipital gyrus. This last cluster was located in a region anatomically compatible with the left extra-striate body area (EBA, Astafiev, Stanley, Shulman, & Corbetta, 2004; Downing, Jiang, Shuman, & Kanwisher, 2001). The importance of vision for body-

176

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

ownership is widely acknowledged, and it has been supported by the observation that viewing a rubber hand being stroked synchronously to one's hidden hand modulates the feeling of ownership of the rubber hand (Botvinick & Cohen, 1998). The role of high-level visual cortices in the body-ownership experience is not limited to the representation of the limb's position in space; instead, it is also involved in the integration of visual representations of the body with somatosensory information about body parts (Limanowski et al., 2014). This latter function seems to be supported by the EBA, a brain region that selectively responds to vision of bodies and body parts (Downing et al., 2001), and changes in limb position (Astafiev et al., 2004). Also, the left inferior parietal lobule has proved to be involved in mapping the position and orientation of limbs in space in strength interconnection with the vision of the body. In particular, an fMRI study demonstrated that the tactile stimulation of the right hand activated the right inferior parietal cortex when the hand was placed across the midline; activations shifted to a left parieto-frontal network when the eyes were open (Lloyd, Shore, Spence, & Calvert, 2003). The left inferior parietal lobule and the left EBA may, therefore, be part of a network in which multisensory inputs are integrated to maintain an up-to-date representation of the position of the body parts and to attribute the body to the self. On the other hand, sense-of-agency-specific clusters were described in the left SMA, in the left posterior insula, in the right postcentral gyrus and the right superior temporal lobe. These results partly overlap with those emerging from our previous meta-analysis, which investigated the neurofunctional correlates of motor intention and the sense of agency (Seghezzi et al., 2019). The left SMA and the left posterior insula turned out to be specific for the sense of agency in the comparison between the sense of agency and motor intention. These areas could be considered as crucial nodes of a sensorymotor comparator network, which predicts the sensory consequences of a given motor program and compares these predictions with the real outcome of the movement to attribute agency (Haggard, 2017).

5.

Conclusion

Our findings show that, even if body-ownership and the sense of agency are partly grounded on the same basic multisensory integration processes, they can be regarded as different experiences underpinned by partially distinct brain networks. The integration of sensory-related signals would be a process shared between body-ownership and sense of agency, whereas their specific experience would rely on additional processes exclusive for each phenomenon (Pyasik, Burin, & Pia, 2018). In line with the idea of “dissociable” processes, it has been suggested that body-ownership and sense of agency differ from each other at the level of their subjective experience, investigated by explicit measure. Indeed, whereas the proprioceptive drift and the sensory attenuation phenomenon, i.e., the implicit measure for body-ownership and sense of agency respectively, are meant to be positively related, no correlations were described between the explicit measure of the two senses (Pyasik et al., 2018). Consequently, Pyasik and

collaborators hypothesize that body-ownership and sense of agency share the low-level processes of sensory and multisensory integration, whereas they differ by their high-level subjective processes (Pyasik et al., 2018). The hypothesized interactive relationship between agency and body-ownership is supported by neuropsychological syndromes, affecting different aspects of self-consciousness. For example, patients with anarchic hand syndrome explicitly report a preserved sense of ownership for body-parts whose voluntary movements and their subsequent consequences are denied. In other words, these patients report a retained sense of body-ownership over the anarchic hand, while they do not experience an overt sense of agency (Berti, Bottini, & Paulesu, 2019; Della Sala, Marchetti, & Spinnler, 1991). Conversely, cases of patients with somatoparaphrenia, the denial of ownership for a paralyzed limb, who do not show anosognosia for hemiplegia and display a veridical sense of agency for the affected limb (Invernizzi et al., 2013), may represent the reverse dissociation. Bodyownership and sense of agency might also be both delusively attributed to an “alien” limb, in an indissociable manner. For example, hemiplegic patients can manifest a pathological embodiment of other people's body parts and erroneously attribute to them both body-ownership and sense of agency (Garbarini et al., 2013, 2014). It is not clear whether the illusory body-ownership induces a non-veridical sense of agency or vice-versa, indicating the close relationship between body and motor awareness, and suggesting not only their independence but also their interactive nature (Berti et al., 2019). Further evidence on the proposed interactive model comes from behavioral findings in healthy subjects. Kalckert and Ehrsson (2012), by taking advantage of a modified version of the rubber hand illusion paradigm in which participants could also control the movements of the model hand, provide behavioral evidence for a double dissociation of bodyownership and the sense of agency, suggesting that they represent distinct cognitive processes. Interestingly, a specific interaction between the two processes was also observed: the sense of agency was stronger when the hand was perceived to be a part of the body. This interactive scenario could be advantageous from an evolutionary perspective since single damage could be not able to disrupt all the crucial abilities that allow us to trace back to us the causal origin of actions and to sense the boundary between one's body and the external world.

Conflict of interest statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding This paper was supported by a grant funded by the Italian Ministry of Health to LZ (Ricerca Corrente; Project L3025).

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

Pre-registration statement No part of the study procedures was pre-registered prior to the research being conducted.

Open practices The study in this article earned an Open Data badge for transparent practices.

CRediT authorship contribution statement Silvia Seghezzi: Conceptualization, Data curation, Investigation, Methodology, Writing - original draft, Writing - review & editing. Gianluigi Giannini: Data curation. Laura Zapparoli: Supervision, Methodology, Writing - original draft, Writing review & editing.

Acknowledgments We are grateful to Prof. Manuela Berlingeri and Prof. Eraldo Paulesu for giving us the opportunity to use the software CluB. We are also grateful to the authors of the original empirical works that were submitted to meta-analysis here. Without their works, we would have not been able to produce this manuscript.

Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.cortex.2019.08.018.

references

Astafiev, S. V., Stanley, C. M., Shulman, G. L., & Corbetta, M. (2004). Extrastriate body area in human occipital cortex responds to the performance of motor actions. Nature Neuroscience, 7(5), 542e548. https://doi.org/10.1038/nn1241. Berlingeri, M., Devoto, F., Gasparini, F., Saibene, A., Corchs, S. E., Clemente, L., et al. (2019). Clustering the brain with “CluB”: A new toolbox for quantitative meta-analysis of neuroimaging data. Frontiers in Neuroscience, 13, 1037. https://doi.org/10.3389/ fnins.2019.01037. Berti, A. M., Bottini, G., & Paulesu, E. (2019). Consapevolezza  e patologia. In Zanichelli (Ed.), Manuale di corporea: normalita  e patologia dei processi cognitivi (3 ed., neuropsicologia. Normalita pp. 677e698). Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (2002). Abnormalities in the awareness of action. Trends in Cognitive Sciences, 6(6), 237e242. Botvinick, M., & Cohen, J. (1998). Rubber hands ‘feel’ touch that eyes see. Nature, 391(6669), 756. https://doi.org/10.1038/35784. Cattinelli, I., Valentini, G., Paulesu, E., & Borghese, N. A. (2013). A novel approach to the problem of non-uniqueness of the

177

solution in hierarchical clustering. IEEE Transactions on Neural Networks and Learning Systems, 24(7), 1166e1173. https:// doi.org/10.1109/TNNLS.2013.2247058. Cauda, F., D'Agata, F., Sacco, K., Duca, S., Geminiani, G., & Vercelli, A. (2011). Functional connectivity of the insula in the resting brain. Neuroimage, 55(1), 8e23. https://doi.org/10.1016/ j.neuroimage.2010.11.049. Daprati, E., Wriessnegger, S., & Lacquaniti, F. (2007). Kinematic cues and recognition of self-generated actions. Experimental Brain Research, 177(1), 31e44. https://doi.org/10.1007/s00221006-0646-9. Della Sala, S., Marchetti, C., & Spinnler, H. (1991). Right-sided anarchic (alien) hand: A longitudinal study. Neuropsychologia, 29(11), 1113e1127. Devoto, F., Zapparoli, L., Bonandrini, R., Berlingeri, M., Ferrulli, A., Luzi, L., et al. (2018). Hungry brains: A meta-analytical review of brain activation imaging studies on food perception and appetite in obese individuals. Neuroscience and Biobehavioral Reviews, 94, 271e285. https://doi.org/10.1016/ j.neubiorev.2018.07.017. Downing, P. E., Jiang, Y., Shuman, M., & Kanwisher, N. (2001). A cortical area selective for visual processing of the human body. Science, 293(5539), 2470e2473. https://doi.org/10.1126/ science.1063414. Ehrsson, H. H., Holmes, N. P., & Passingham, R. E. (2005). Touching a rubber hand: Feeling of body ownership is associated with activity in multisensory brain areas. The Journal of Neuroscience, 25(45), 10564e10573. https://doi.org/10.1523/JNEUROSCI.080005.2005. Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). That's my hand! Activity in premotor cortex reflects feeling of ownership of a limb. Science, 305(5685), 875e877. https://doi.org/10.1126/ science.1097011. Eickhoff, S. B., Laird, A. R., Grefkes, C., Wang, L. E., Zilles, K., & Fox, P. T. (2009). Coordinate-based activation likelihood estimation meta-analysis of neuroimaging data: A randomeffects approach based on empirical estimates of spatial uncertainty. Human Brain Mapping, 30(9), 2907e2926. https:// doi.org/10.1002/hbm.20718. Farrer, C., Franck, N., Georgieff, N., Frith, C. D., Decety, J., & Jeannerod, M. (2003). Modulating the experience of agency: A positron emission tomography study. Neuroimage, 18(2), 324e333. Farrer, C., Frey, S. H., Van Horn, J. D., Tunik, E., Turk, D., Inati, S., et al. (2008). The angular gyrus computes action awareness representations. Cerebral Cortex, 18(2), 254e261. https:// doi.org/10.1093/cercor/bhm050. Frith, C. D. (1987). The positive and negative symptoms of schizophrenia reflect impairments in the perception and initiation of action. Psychological Medicine, 17(3), 631e648. Frith, C. D., Friston, K., Liddle, P. F., & Frackowiak, R. S. (1991). Willed action and the prefrontal cortex in man: A study with PET. Proceedings Biological Sciences, 244(1311), 241e246. https:// doi.org/10.1098/rspb.1991.0077. Gallagher, S. (2000). Philosophical conceptions of the self: Implications for cognitive science. Trends in Cognitive Sciences, 4(1). Garbarini, F., Fornia, L., Fossataro, C., Pia, L., Gindri, P., & Berti, A. (2014). Embodiment of others' hands elicits arousal responses similar to one's own hands. Current Biology, 24(16), R738eR739. https://doi.org/10.1016/j.cub.2014.07.023. Garbarini, F., Pia, L., Piedimonte, A., Rabuffetti, M., Gindri, P., & Berti, A. (2013). Embodiment of an alien hand interferes with intact-hand movements. Current Biology, 23(2), R57eR58. https://doi.org/10.1016/j.cub.2012.12.003. Gentile, G., Guterstam, A., Brozzoli, C., & Ehrsson, H. H. (2013). Disintegration of multisensory signals from the real hand

178

c o r t e x 1 2 1 ( 2 0 1 9 ) 1 6 9 e1 7 8

reduces default limb self-attribution: An fMRI study. The Journal of Neuroscience, 33(33), 13350e13366. https://doi.org/ 10.1523/JNEUROSCI.1363-13.2013. Grivaz, P., Blanke, O., & Serino, A. (2017). Common and distinct brain regions processing multisensory bodily signals for peripersonal space and body ownership. Neuroimage, 147, 602e618. https://doi.org/10.1016/j.neuroimage.2016.12.052. Haggard, P. (2008). Human volition: Towards a neuroscience of will. Nature Reviews Neuroscience, 9(12), 934e946. https:// doi.org/10.1038/nrn2497. Haggard, P. (2017). Sense of agency in the human brain. Nature Reviews Neuroscience, 18(4), 196e207. https://doi.org/10.1038/ nrn.2017.14. Haggard, P., Clark, S., & Kalogeras, J. (2002). Voluntary action and conscious awareness. Nature Neuroscience, 5(4), 382e385. https://doi.org/10.1038/nn827. Invernizzi, P., Gandola, M., Romano, D., Zapparoli, L., Bottini, G., & Paulesu, E. (2013). What is mine? Behavioral and anatomical dissociations between somatoparaphrenia and anosognosia for hemiplegia. Behavioural Neurology, 26(1e2), 139e150. https://doi.org/10.3233/BEN-2012-110226. Kalckert, A., & Ehrsson, H. H. (2012). Moving a rubber hand that feels like your own: A dissociation of ownership and agency. Frontiers in Human Neuroscience, 6, 40. https://doi.org/10.3389/ fnhum.2012.00040. Ku¨hn, S., Brass, M., & Haggard, P. (2013). Feeling in control: Neural correlates of experience of agency. Cortex, 49(7), 1935e1942. https://doi.org/10.1016/j.cortex.2012.09.002. Kurth, F., Zilles, K., Fox, P. T., Laird, A. R., & Eickhoff, S. B. (2010). A link between the systems: Functional differentiation and integration within the human insula revealed by metaanalysis. Brain Structure & Function, 214(5e6), 519e534. https:// doi.org/10.1007/s00429-010-0255-z. Limanowski, J., Lutti, A., & Blankenburg, F. (2014). The extrastriate body area is involved in illusory limb ownership. Neuroimage, 86, 514e524. https://doi.org/10.1016/j.neuroimage.2013.10.035. Lloyd, D. M., Shore, D. I., Spence, C., & Calvert, G. A. (2003). Multisensory representation of limb position in human premotor cortex. Nature Neuroscience, 6(1), 17e18. https:// doi.org/10.1038/nn991.
, M. (2012). Degrees of Murray, R. J., Schaer, M., & Debbane separation: A quantitative neuroimaging meta-analysis investigating self-specificity and shared neural activation between self- and other-reflection. Neuroscience and Biobehavioral Reviews, 36(3), 1043e1059. https://doi.org/10.1016/ j.neubiorev.2011.12.013. € rnsdotter, M., Gentile, G., Jonsson, T., Li, T. Q., & Petkova, V. I., Bjo Ehrsson, H. H. (2011). From part- to whole-body ownership in the multisensory brain. Current Biology, 21(13), 1118e1122. https://doi.org/10.1016/j.cub.2011.05.022. Pyasik, M., Burin, D., & Pia, L. (2018). On the relation between body ownership and sense of agency: A link at the level of sensory-

related signals. Acta Psychologica (Amst), 185, 219e228. https:// doi.org/10.1016/j.actpsy.2018.03.001. Rorden, C., & Brett, M. (2000). Stereotaxic display of brain lesions. Behavioural Neurology, 12(4), 191e200. Seghezzi, S., Zirone, E., Paulesu, E., & Zapparoli, L. (2019). The brain in (Willed) action: A meta-analytical comparison of imaging studies on motor intentionality and sense of agency. Frontiers in Psychology, 10, 804. https://doi.org/10.3389/ fpsyg.2019.00804. Showers, M. J., & Lauer, E. W. (1961). Somatovisceral motor patterns in the insula. Journal of Comparative Neurology, 117, 107e115. Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: Visuotactile integration and self-attribution. Journal of Experimental Psychology Human Perception and Performance, 31(1), 80e91. https://doi.org/10.1037/0096-1523.31.1.80. Tsakiris, M., Hesse, M. D., Boy, C., Haggard, P., & Fink, G. R. (2007). Neural signatures of body ownership: A sensory network for bodily self-consciousness. Cerebral Cortex, 17(10), 2235e2244. https://doi.org/10.1093/cercor/bhl131. Tsakiris, M., Longo, M. R., & Haggard, P. (2010). Having a body versus moving your body: Neural signatures of agency and body-ownership. Neuropsychologia, 48(9), 2740e2749. https:// doi.org/10.1016/j.neuropsychologia.2010.05.021. Tsakiris, M., Prabhu, G., & Haggard, P. (2006). Having a body versus moving your body: How agency structures bodyownership. Consciousness and Cognition, 15(2), 423e432. https:// doi.org/10.1016/j.concog.2005.09.004. Tsakiris, M., Schu¨tz-Bosbach, S., & Gallagher, S. (2007). On agency and body-ownership: Phenomenological and neurocognitive reflections. Consciousness and Cognition, 16(3), 645e660. https:// doi.org/10.1016/j.concog.2007.05.012. Turkeltaub, P. E., Eickhoff, S. B., Laird, A. R., Fox, M., Wiener, M., & Fox, P. (2012). Minimizing within-experiment and withingroup effects in activation likelihood estimation metaanalyses. Human Brain Mapping, 33(1), 1e13. https://doi.org/ 10.1002/hbm.21186. Ward, J. H. (1963). Hierarchical Grouping to optimize an objective function. Journal of the American Statistical Association, 58(301), 236e244. Wolpert, D. M., & Ghahramani, Z. (2000). Computational principles of movement neuroscience. Nature Neuroscience, 3(Suppl), 1212e1217. https://doi.org/10.1038/81497. Zapparoli, L., Seghezzi, S., & Paulesu, E. (2017). The what, the when, and the whether of intentional action in the brain: A meta-analytical review. Frontiers in Human Neuroscience, 11, 238. https://doi.org/10.3389/fnhum.2017.00238. Zapparoli, L., Seghezzi, S., Scifo, P., Zerbi, A., Banfi, G., Tettamanti, M., et al. (2018). Dissecting the neurofunctional bases of intentional action. Proceedings of the National Academy of Sciences of the United States of America. https://doi.org/ 10.1073/pnas.171889111.