Running head: NEURAL CORRELATES OF METACOGNITION OF AGENCY 1 Dissociating Neural Correlates of Action Monitoring and Metacognition of Agency

نویسندگان

  • David B. Miele
  • Tor D. Wager
  • Jason P. Mitchell
  • Janet Metcalfe
چکیده

Judgments of agency refer to people’s self-reflective assessments concerning their own control: their assessments of the extent to which they themselves are responsible for an action. These self-reflective metacognitive judgments can be distinguished from action monitoring, which involves the detection of the divergence (or lack of divergence) between observed states and expected states. Presumably, people form judgments of agency by metacognitively reflecting on the output of their action monitoring and then consciously inferring the extent to which they caused the action in question. While a number of previous imaging studies have been directed at action monitoring, none have assessed judgments of agency as a potentially separate process. The present fMRI study used an agency paradigm that not only allowed us to examine the brain activity associated with action monitoring, but that also enabled us to investigate those regions associated with metacognition of agency. Regarding action monitoring: we found that being ‘out of control’ during the task (i.e., detection of a discrepancy between observed and expected states) was associated with increased brain activity in the right temporoparietal junction (rTPJ), while being 'in control' was associated with increased activity in the presupplementary motor areas (pre-SMA), rostral cingulate zone (RCZ), and dorsal striatum (regions linked to self-initiated action). In contrast, when participants made self-reflective metacognitive judgments about the extent of their own control (i.e., judgments of agency) compared to when they made judgments that were not about control (i.e., judgments of performance), increased activity was observed in the anterior prefrontal cortex (aPFC), a region associated with self-reflective processing. These results indicate that action monitoring is dissociable from people’s conscious self-attributions of control. NEURAL CORRELATES OF METACOGNITION OF AGENCY 3 Dissociating Neural Correlates of Action Monitoring and Metacognition of Agency The present study examines the neurocognitive basis of action monitoring and metacognitive judgments of agency. Action monitoring involves the ability to detect whether action outcomes are in line with the expected consequences of our intended actions (i.e., with our action plans). It plays an important role in the way we coordinate our movements with objects and events in the environment, including other agents. Although action monitoring is a critical component of voluntary action, empirical evidence suggests that it may (at least part of the time) operate at a spontaneous or automatic level of processing. For example, a number of studies have demonstrated that we can respond to minor violations of our action plans (such as unpredicted jumps in target position) at both a neural and behavioral level without awareness of having done so (e.g., Pisella et al., 2000). Thus, it appears that we can (and sometimes do) use the output of our monitoring to successfully coordinate our actions without first consciously interpreting whether this output indicates the presence or absence of control. In contrast, in order to make an explicit judgment concerning our own agency, we do need to consciously assess whether or not we were in control. For instance, to determine our responsibility for a particular action, we must be able to metacognitively reflect on the output of our action monitoring, as well as other relevant cues, and then consciously infer the extent to which we were or were not the cause of the action in question (Haggard & Tsakiris, 2009; Synofzik, Vosgerau, & Newen, 2008; Wegner, 2002, 2003; Wegner, Sparrow, & Winerman, 2004). Therefore, when it comes to understanding human agency, the question is not simply how we monitor disturbances in the execution of our actions, but also how we use the output of this monitoring to make metacognitive judgments about whether or not we were in control. Although we as well as others (Georgieff & Jeannerod, 1998; Haggard & Tsakiris, 2009; Metcalfe & Greene, 2007; Pacherie, in press; Synofzik et al., 2008) have proposed that there may be NEURAL CORRELATES OF METACOGNITION OF AGENCY 4 a distinction between action monitoring and metacognition of agency, brain imaging studies have focused exclusively on identifying patterns of neural activity that are associated with action monitoring (see David, Newen, & Vogeley, 2008, for a review). Typically, the participants in these studies have attempted to perform a specific action or behavior and then received sensory feedback concerning the spatial trajectory or temporal sequence of their movement that either coincides or conflicts with their action plan. When the feedback was discrepant with expectations, the imaging results have consistently shown increased activation in areas surrounding the right temporoparietal junction (rTPJ, e.g., Farrer et al., 2003; Leube et al., 2003; Nahab et al., in press; Spengler, von Cramon, & Brass, 2009). The TPJ is a region with boundaries that cannot be structurally defined and is often described as encompassing the supramarginal gyrus, the inferior parietal lobe/angular gyrus, caudal parts of the superior temporal gyrus, dorsal-rostral parts of the occipital gyri, and posterior parts of the superior temporal sulcus (Decety & Lamm, 2007). We later report the results of a meta-analysis of recent agency studies that we conducted in order to define this broad region of interest (ROI) for the present study. The fact that increases in TPJ activation are reliably associated with decreases in control has led many researchers to speculate that the TPJ receives input from a comparator mechanism that is responsible for determining whether the predicted consequences of an intended action (based on the output of an internal “forward model”) match sensory feedback about the actual trajectory of the action (Blakemore et al., 2002; Hohwy & Frith, 2004; Wolpert & Ghahramani, 2000; cf. Synofzik et al., 2008). When the predicted and actual trajectory do not match, the TPJ is thought to receive a signal indicating that a mismatch exists and that corrective action may be required. On a somewhat less consistent basis, the results of previous action monitoring studies have sometimes shown increased activation in the cerebellum (Blakemore, Frith, & Wolpert, 2001), precuneus (David et al., 2007), and extrastriate body area (EBA; David et al., 2007) in response to a mismatch, and in NEURAL CORRELATES OF METACOGNITION OF AGENCY 5 the insula (Farrer et al., 2003; Tsakiris, Longo, & Haggard, 2010) and pre-SMA (Tsakiris et al. 2010) when actions proceed as planned. Though participants in some of the studies on action monitoring provided judgments of agency after they completed the focal task (e.g., Nahab et al., in press; Spengler, von Cramon, & Brass, 2009; Tricomi, Delgado, & Fiez, 2004), these judgments were used only to identify neural activity during the task that was related to feeling in or out of control. Imaging data corresponding to the judgments themselves, though, were either not collected or not analyzed. Thus, as of yet, no studies have been published that examine the neural activity associated with metacognition of agency itself—that is, with reflecting on the output of one’s action monitoring after an action has been performed. In the present experiment, each trial consisted of both a game phase, in which participants’ task was to move the cursor (via the trackball) along a horizontal track to touch each of the downward scrolling X’s, and a judgment phase, in which they judged either how much “in control” they felt during the game or how well they had performed (see Figure 1). In a manipulation somewhat similar to those used in previous action-monitoring studies (Blakemore, Frith, & Wolpert, 1999; Farrer et al., 2003; Franck et al., 2001) in which the trajectory of the person's apparent movement was altered, we disrupted participants’ objective control by adding random noise, or ‘turbulence’, to the position of the cursor. Alternatively, we sometimes distorted control by having the X's disappear 'by magic' (i.e., without the cursor having actually touched the X). Our behavioral question was whether participants’ judgments of agency would be sensitive to these disruptions of control. Our imaging question was whether similar patterns of brain activation would be observed during the game and judgment phases of the task or whether action monitoring and judgments of agency involve different sets of neural processes. NEURAL CORRELATES OF METACOGNITION OF AGENCY 6 Because our previous research using a similar paradigm has shown that the judgments of agency made by college students, older adults, and even (to some extent) children, tend to accurately reflect these manipulated disruptions of control (Kirkpatrick, Metcalfe & Greene, 2008; Metcalfe & Greene, 2007; Metcalfe, Eich, & Castel, 2010), we hypothesized that when participants experienced such disruptions during the game phase of the present experiment, they would exhibit increased neural activity in regions that are associated with action monitoring, particularly the rTPJ. In contrast, we hypothesized that when they reflected back on this disruption (i.e., when they made their agency judgments) compared to when they reflected back on performance, there would be increased activation in areas that subserve key processes involved in metacognition of agency, such as self-reflection and self-attribution. Perhaps the most extensive line of research on self-reflection has focused on identifying brain regions that underlie the evaluation of one’s own traits, abilities, and mental states (for reviews, see Northoff et al., 2006; Schmitz & Johnson, 2007; van der Meer et al., 2010). In these studies, participants are typically presented with trait or mental state adjectives and are asked to indicate whether the adjectives apply to them (self-attribution condition) or whether they apply to another individual, such as a friend, relative, or celebrity (other-attribution condition). Results have consistently shown stronger activation of cortical midline structures and their surrounding regions, including the middle and medial anterior prefrontal cortex (aPFC; i.e., BA 10), during selfattribution trials compared to other-attribution trials (e.g., Jenkins & Mitchell, 2010; see also Vinogradov et al., 2006). A closely related line of research has shown increased activity in the aPFC when people reflect on their own feelings or intentions while forming judgments or making decisions (for reviews, see Amodio & Frith, 2006; Christoff and Gabrieli, 2000; Mitchell, 2009; Ramnani & Owen, 2004). Given the convergence of these findings, a number of investigators have theorized that the processing of self-relevant information (i.e., information about one’s attributes, NEURAL CORRELATES OF METACOGNITION OF AGENCY 7 experiences, and mental states) emerges from cortical midline structures, including regions of the aPFC, that are tightly interconnected and functionally distinct from structures that underlie domaingeneral cognitive processing (Northoff et al., 2006; Schmitz & Johnson, 2007; van der Meer et al., 2010). In the present experiment, then, it was our hypothesis that these structures, rather than the structures associated with action monitoring, would be particularly active when people made metacognitive judgments of agency compared to judgments of performance. Method Participants The participants were 11 members of the Columbia community (7 female; 1 left-handed; mean age, 27; range, 18-36) who provided consent in a manner approved by the Columbia University Institutional Review Board and were treated in accordance with APA ethical guidelines. An outlier analysis determined that the left-handed participant’s data did not differ significantly from the rest of the sample and so they were included in all analyses. Behavioral Procedure The primary task, which was framed as a computer game, involved using a trackball to move a cursor left and right along a horizontal track at the bottom of the screen as approximately 20-30 stimuli (Xs and Os) that were randomly distributed from left to right scrolled down from the top of the screen to the bottom (see Metcalfe et al., 2010; Metcalfe & Greene, 2007). During each trial of the task, the Xs or Os disappeared as soon as they were “touched” by the cursor, but continued to scroll past the horizontal track if they were not touched. In addition, a ping sound occurred each time an X was hit and a pong sound occurred each time an O was hit (see Figure 1). Before entering the scanner, participants were instructed to use the track ball in order to touch as many Xs with the cursor as possible and to avoiding touching any of the Os. They were also told that, after completing each trial, they would be asked to make either a judgment of agency NEURAL CORRELATES OF METACOGNITION OF AGENCY 8 or a judgment of performance. For judgments of agency, participants were presented with a visual analogue scale and instructed to pull the slider to the left if they felt they were “not in control” or to the right if they felt they were “in control.” For judgments of performance, participants were presented with the same scale, but were instructed to pull the slider to the left if they felt that their performance was “very low” or to the right if they felt their performance was “very high.” For both types of judgment, the left end of the scale was coded 0 and the right end was coded 100, with values in between being assigned a value based on the proportion of the distance between the two ends that was spanned by the slider. Once in the scanner, each experimental session consisted of 4-6 runs, with each run consisting of a single set of 24 trials, for a total of 96-144 trials per participant. All trials consisted of two phases, a game phase that lasted 15 s and a judgment phase that lasted from the end of the game phase (at which point the visual analogue scale immediately appeared) until the participant clicked an “OK” button after making an agency or performance rating (see Figure 1). The intertrial interval (during which a fixation cross appeared on the screen) was set to 2 s. Within each run, we manipulated two factors (termed “turbulence” and “magic”), such that the game phase of each trial belonged to 1 of 4 conditions: a Control condition in which the participants experienced a direct temporal and spatial correspondence between their movement of the mouse and the movement of the cursor on the screen, a Turbulence condition in which pseudo-random noise was added to the spatial correspondence between the mouse and the cursor, a Magic condition in which Xs disappeared and the corresponding ping noise sounded whenever the cursor was moved within 10 pixels of the outer edge of a target (such that participants often received credit for Xs without actually touching them), and a combined TurbMagic condition in which both manipulations occurred together. The 24 trials in each run were grouped into 6 blocks, such that each block contained trials from all 4 conditions. Within each block, the condition order was pseudorandom. NEURAL CORRELATES OF METACOGNITION OF AGENCY 9 Across runs, we manipulated a single factor (judgment type), such that participants made only judgments of agency for all of the trials in half of the runs and made judgments of performance for all of the trials in the other half of the runs. Judgment type alternated from one run to the next, with the judgment type of the first run counterbalanced across participants. The purpose of manipulating judgment type across runs, as opposed to having participants make both types of judgment for each trial or on consecutive trials, was to avoid carryover effects. In addition to recording participants’ judgments, the program recorded the number of Xs and Os participants touched during the game phase of each trial, the number of Xs and Os that appeared on the screen but were not touched, and the length of the judgment phase. Behavioral Data Analysis Performance. Trial by trial hit rates (i.e., percentage of Xs touched), false alarm rates (i.e., percentage of Os touched), and d’ values were computed for each participant. Because analyses of these performance variables yielded similar patterns of results, only the analysis of hit rate (which was the primary measure of performance in previous studies; e.g., Metcalfe et al., 2010; Metcalfe & Greene, 2007) is reported. To determine the effectiveness of the turbulence and magic manipulations, we submitted hit rate to a 2 (turbulence: turbulence vs. no turbulence) x 2 (magic: magic vs. no magic) repeated measures ANOVA. Judgments. Because there tends to be a strong correlation between participants’ judgments of agency and their hit rate (Metcalfe et al., 2010; Metcalfe & Greene, 2007), the primary question of interest was whether or not participants correctly inferred their lack of agency in conditions under which they were not fully in control (i.e., the Turbulence, Magic, and TurbMagic conditions), regardless of how well they thought they had performed in these conditions. That is, when controlling for differences in perceived performance across trials, do we still find effects of the Turbulence and Magic manipulations on judgments of agency? Affirmative answers would suggest NEURAL CORRELATES OF METACOGNITION OF AGENCY 10 that any brain activity resulting from these manipulations was related to the feeling of agency itself. To address the question, we conducted several different analyses. First, in keeping with our past experiments using this paradigm, we computed summary scores for each participant based on the following formula, (JOPC-JOAC)-(JOPE-JOAE), where JOP refers to the mean judgment of performance, JOA refers to the mean judgment of agency, the subscript C refers to the Control condition, and the subscript E refers to the particular experimental condition of interest (i.e., Turbulence, Magic, or TurbMagic). These summary scores, which compress the interactions between judgments of agency and judgments of performance into a single measure (see Metcalfe, Eich, & Castel, 2010), should have been negative, so long as participants realized that their performance in the experimental conditions (where agency was objectively decreased) was not entirely due to their own actions. Second, to take advantage of the factorial design used in the present study (and to be consistent with the contrast analyses of the imaging data reported below), we also computed another set of summary scores based on the main effects of the two manipulations. The formula used for these scores was (JOPA-JOAA)-(JOPP-JOAP), where the subscript A refers to the absence of the particular experimental manipulation of interest (e.g., the average of the Control and Magic conditions, when looking at Turbulence) and the subscript P refers to the presence of the manipulation (e.g., the average of the Turbulence and TurbMagic conditions, when looking at Turbulence). Finally, to analyze the behavioral results in a manner that is more consistent with our parametric analysis of the imaging data reported below, we submitted each participant’s judgments of agency to a separate regression test in which hit rate (which served as a proxy for perceived performance), turbulence (dummy-coded: 0 = no turbulence, 1 = turbulence), and magic (dummycoded: 0 = no magic, 1 = magic) served as simultaneous predictors. We then performed a series of one-sample t-tests to determine whether the mean beta coefficients for the group differed NEURAL CORRELATES OF METACOGNITION OF AGENCY 11 significantly from zero (Lorch & Myers, 1990). Next, to determine whether the manipulations uniquely targeted participants’ metacognition of agency, or whether they had similar effects on participants’ perceived performance, we repeated the previously described regression analyses and t-tests with judgments of performance as the dependent variable. We completed the analysis by directly comparing the beta coefficients associated with the two types of judgments in a series of paired-samples t-tests. fMRI Data Acquisition Images were acquired with a GE twin-speed 1.5-T scanner. Whole brain functional data were acquired in 27 contiguous axial slices (4.0 mm thick, 3 × 3 mm in-plane resolution) parallel to the anterior-posterior commissure (AC–PC) line with a T2*-weighted echo-planar imaging (EPI) sequence (TR = 2000, TE = 38, flip angle = 90°, FoV = 192, array size = 64 × 64). Each run acquired 275 whole brain volumes, the first 5 of which were discarded in order to allow the scanner to stabilize. Structural data were acquired with a high-resolution T1-weighted spoiled gradient echo (SPGR) scan which recorded 182 slices at a thickness of 1 mm and resolution of 1 × 1 mm. fMRI Data Analysis Preprocessing. Spatial preprocessing and statistical analyses were performed with SPM5 (Wellcome Trust Centre for Neuroimaging, University College London, UK). First, each functional brain volume was corrected for differences in acquisition time between slices and realigned to correct for head movement. The data were then normalized to a standard anatomical space using the Montreal Neurological Institute (MNI) EPI template, re-sampled into 2 mm isotropic voxels, and spatially smoothed with an 8 mm full width at half maximum [FWHM] Gaussian Kernel. Primary statistical analyses. Statistical analyses were performed to test two separate models using the general linear model framework implemented in SPM5 (Friston et al., 1995). In both cases, a high-pass filter (128 s) was applied to the data in order to remove low-frequency drifts NEURAL CORRELATES OF METACOGNITION OF AGENCY 12 in signal changes. The first model (i.e., the simple contrast model) was designed to identify brain regions that responded to disturbances (i.e., Turbulence and Magic) in agency during the game and judgment phases, as well as regions that showed increased activation when making judgments of agency compared to judgments of performance. Thus, eight within-run conditions (i.e., the four game conditions crossed with the two task phases) were modeled as boxcar functions (such that each phase of each trial was treated as a separate epoch) and convolved with the canonical hemodynamic response function (HRF) to create regressors of interest. For each participant, voxelwise statistical parametric maps were created for the game and judgment phase contrasts of turbulence vs. no turbulence (Turbulence + TurbMagic vs. Control + Magic), magic vs. no magic (Magic + TurbMagic vs. Control + Turbulence), and turbulence × magic (i.e., Turbulence + Magic vs. TurbMagic + Control). In addition, a map was created for the judgment phase contrast of judgment of agency vs. judgment of performance. Whereas the first four contrasts compared different regressors collapsing across all runs, the last contrast compared the same judgment phase regressor between two sets of runs. The second model (i.e., the parametric model) was setup to identify brain regions that varied in activation during the game phase as a function of perceived agency. Because we were primarily interested in participants’ sense of agency independent of their perceived performance on the game (i.e., their sense of being more or less in control than was indicated by the number of Xs they managed to touch), and because we did not collect judgments of performance on the same trials as judgments of agency (as that would have precluded the temporal separation necessary for computing the judgment of agency vs. judgment of performance contrast described above), we operationalized this construct as the difference between their judgment of agency and hit rate on each trial of the judgment of agency runs (i.e., judgment of agency – [Xhits / Xtotal * 100]). As a NEURAL CORRELATES OF METACOGNITION OF AGENCY 13 control, we also computed the difference between participants’ judgment of performance and hit rate on each trial of the judgment of performance runs. For each set of runs (i.e., the judgment of agency or judgment of performance runs), the corresponding difference score for each trial (i.e., judgment of agency – [Xhits / Xtotal * 100] or judgment of performance – [Xhits / Xtotal * 100]) was entered into the model as a parametric modulator of both task phases, each of which was modeled as a boxcar function and convolved with the canonical HRF. In addition, because we wanted to control for potential confounds associated with performance, such as motor activity and affect (e.g., participants may have moved the mouse more or been more pleased with their performance on the difficult turbulence trials than on other trials), each difference score was entered into the model after we had included hit rate as a secondary modulator of the two task phases. That is, because SPM5’s default option is the serial orthogonalization of parametric modulators, entering hit rate before the difference score ensured that none of the variance shared by the two regressors would contaminate activation that is uniquely associated with the difference score. Based on this model, subject-specific statistical parametric maps were created for the game phase regressors corresponding to each difference score. These maps represented the extent to which changes in the BOLD signal during the game phase were correlated with participants’ subsequent sense of agency or performance at each voxel in the brain. To explore each model, single-subject contrast images from the first-level were entered into a second-level random effects analysis, which involved computing one-sample t-tests across the contrast images of all subjects in order to create a series of group level statistical parametric maps (Mumford & Nichols, 2009). For whole-brain analysis, these maps were thresholded first at the voxel-level (p < .005, uncorrected) and then at the cluster level (p < .05, family-wise error [FWE] corrected) to protect against false-positive activations (Friston, Worsley, Frackowiak, Mazziotta, & Evans, 1993). All significant activations were overlaid on sections of the MNI canonical brain, with NEURAL CORRELATES OF METACOGNITION OF AGENCY 14 peak voxels reported in MNI coordinate space. For the anatomical labeling of these activations, we used the automated anatomical labeling (AAL) atlas (Tzourio-Mazoyer et al., 2002). Region of interest analyses. Brain regions that were identified based on theory and literature review were analyzed using small volume corrections (SVCs) as implemented in SPM5 (Worsley et al., 1996). Specifically, when testing for increased activation in regions of the anterior prefrontal cortex (aPFC) during the judgment phase of the agency task, we restricted the search volume to four 10 mm spheres that were centered at coordinates reported in a recent meta-analysis of self-reflection studies (van der Meer et al., 2010). The meta-analysis explored two sets of contrasts found in fMRI and PET studies of self-reflection: self-reflection vs. baseline contrasts and self-reflection vs. other-reflection contrasts. We based the ROI on the analysis of the latter set because, as the authors suggest, it identified regions “unique to self-reflective processing” as opposed to regions pertaining to “reflective processing on a broader scale.” The analysis of the 17 studies that included self-reflection vs. other-reflection contrasts yielded a large cluster with peaks in the pregenual part of the ACC (pACC; BA 24, 32; 2, 42, 20) and the bilateral aPFC (0, 50, -2; -2, 54, 8;-18, 50, 16)—the coordinates of these peaks were used to create the four 10 mm spheres. When testing for increased activity in the bilateral TPJ during the game phase of the task, we restricted the search volume to a 6745-voxel region derived from a meta-analysis of recent agency studies. The meta-analysis, which we conducted ourselves, was based on 21 TPJ coordinates from 10 agency studies originally included in a meta-analysis by Decety and Lamm (2007), as well as 25 additional coordinates reported in 7 recent agency studies (see Table 1). Analysis of the 46 activation peaks was conducted in MNI space using Multi-level Kernel Density Analysis (MKDA; Wager, Lindquist, Nichols, Kober, & Van Snellenberg, 2009). As shown in Figure 3, the resultant activation map, which was thresholded first at the voxel-level (p < .05) and then at the cluster-level (p < .05), yielded large clusters in the left (-50, -50, 34; KE = 1738) and right TPJ (54, -50, 32; KE = NEURAL CORRELATES OF METACOGNITION OF AGENCY 15 5007) that included portions of the angular gyus, supramarginal gyrus, inferior parietal lobule, superior temporal gyrus, and middle temporal gyrus. MKDA was used because it designates contrast maps (and not peaks) as the unit of analysis. This means that contrast maps are essentially treated as random effects, which ensures that no one contrast map can contribute disproportionately to the overall results, even if it contains numerous peaks in the same area. Unless otherwise indicated, all maps analyzed using SVCs were thresholded at a voxel-level of p < .005 (uncorrected) and a cluster level of p < .05 (FWE-corrected). Results Behavioral Results Performance. The results of the hit rate analysis revealed a significant main effect of turbulence (F(1, 10) = 57.35, p < .001, MSe = 707.42, ηp = .85), such that participants were less accurate when turbulence was present (M = 64.40, SE = 1.66) than when it was absent (M = 72.42, SE = 1.61), as well a significant main effect of magic (F(1, 10) = 1,146.72, p < .001, MSe = 12,087.35, ηp = .99), such that participants were more accurate when magic was present (M = 84.99, SE = 1.76) than when it was absent (M = 51.84, SE = 1.47). These main effects were qualified by a significant interaction (F(1, 10) = 148.80, p < .001, MSe = 543.84, ηp = .94). Bonferroni tests revealed that, although turbulence lowered the mean hit rate when magic was absent (t(10) = 11.42, p < .001), it had no effect on hit rate when magic was present (t(10) = .91, p = .38). It should also be noted that, across all conditions, there were strong within-participants correlations between hit rate and judgments of performance (rmean = .71), an important finding given that we used hit rate as a proxy for judgments of performance in some of the imaging analyses. Judgments. As shown in the left panel of Figure 2, the results of the first analysis replicated the findings from our previous agency studies. More specifically, the summary agency scores for the Turbulence ( = -27.21, t(10) = 4.68, p < .001), Magic ( = -13.35, t(10) = 3.79, p = .004), and NEURAL CORRELATES OF METACOGNITION OF AGENCY 16 TurbMagic conditions ( = -37.45, t(10) = 5.78, p < .001) were significantly less than zero, indicating that in all three instances people correctly recognized that they were not completely in control. Follow-up tests revealed that, although this effect was stronger in the Turbulence condition than in the Magic condition (t(10) = 2.81, p = .02), it was strongest when both manipulations were present in the same trial (i.e., in the TurbMagic condition; ts > 4.12, ps < .003). The results of the second analysis showed virtually the same pattern, as depicted in the right panel of Figure 2. The summary agency scores for the main effects of the Turbulence ( = -25.65, t(10) = 4.74, p < .001) and Magic manipulations ( = -11.80, t(10) = 6.05, p < .001) were both significantly less than zero, and this effect was again stronger for turbulence than for magic (t(10) = 2.81, p = .02). For the final analysis, within-subject regression tests revealed significant effects of hit rate (βmean = .34, t(10) = 9.03, p < .001) and turbulence (βmean = -.66, t(10) = -12.34, p < .001) on participants’ judgments of agency, but no effect of magic (βmean = -.05, t(10) = -.83, p = .43). Thus, consistent with the results of first two analyses, decreases in control due to turbulence (though not magic in this case) led to decreases in participants’ judgments of agency, even when controlling for task performance. Additional regression analyses revealed significant effects of hit rate (βmean = .64, t(10) = 10.69, p < .001) and turbulence (βmean = -.17, t(10) = -4.11, p = .002), but no effect of magic (βmean = .03, t(10) = .43, p = .78) on participants’ judgments of performance. However, a comparison of beta coefficients between the two types of judgments showed that turbulence had a significantly stronger negative effect on judgments of agency compared to judgments of performance (t(10) = -9.28, p < .001), whereas hit rate had a significantly weaker positive effect (t(10) = -8.40, p < .001). Once again, these results indicate that the main outcome of the turbulence manipulation was a reduction in participants’ sense of control. fMRI Results Simple contrast model. NEURAL CORRELATES OF METACOGNITION OF AGENCY 17 Turbulence contrasts. To determine whether the turbulence manipulation increased activity in the brain region that is most directly associated with detecting disruptions of control, we examined neural responses in the TPJ ROI during the game phase of the task for turbulence trials (Turbulence + TurbMagic) compared to no turbulence trials (Control + Magic). As shown in Figure 3, the results of the analysis yielded a large cluster of activity in the right TPJ (56, -34, 26; Z = 4.90; KE = 618; p < .05, cluster-corrected), as well as a smaller cluster in the right TPJ (48, -62, 4; Z = 3.96; KE = 72; p < .05, FDR-corrected) and two smaller clusters in the left TPJ (-44, -64, 12; Z= 3.37; KE = 65; p < .05, FDR-corrected; -50, -40, 24; Z= 3.48; KE = 21; p < .05, FDR-corrected) that did not meet the primary threshold. To explore the pattern of this activity across all four game conditions, we used MarsBar v0.42 software (http://marsbar.sourceforge.net) to extract percent signal change for each participant from the portion of the large right-sided cluster that overlapped with the TPJ ROI. As depicted in the bottom right inset of Figure 3, there was significantly more activity in the Turbulence condition than in the Magic or Control conditions (ts > 4.46, ps < .002); furthermore, these effects were not qualified by a significant Turbulence x Magic Interaction (F(1, 10) = .11, p = .75, MSe < .001, ηp = .01). Thus, it appears that (independent of any effects of Magic) our turbulence manipulation had a similar effect on participants’ neural activity as the manipulations of perceived control used in previous studies of agency. The corresponding whole-brain analysis revealed that much of the activity within the TPJ ROI belonged to a significant cluster of right-sided activation that included regions of the supramarginal gyrus, posterior superior temporal gyrus/sulcus (STG/STS), and the rolandic operculum (see Table 2). The analysis also revealed a similar, but smaller cluster of activation on the left side, as well as clusters in the bilateral posterior middle temporal gyrus, the left postcentral gyrus, and the right precentral gyrus. The reverse contrast (i.e., no turbulence > turbulence trials), did not reveal any suprathreshold brain regions during the game phase. NEURAL CORRELATES OF METACOGNITION OF AGENCY 18 During the judgment phase, the turbulence > no turbulence contrast did not reveal any suprathreshold brain regions. However, the reverse contrast revealed a large cluster of right-sided activity that extended across regions of the lingual gyrus, calcarine sulcus, and cuneus, as well as a smaller cluster that included regions of the right frontal operculum and precentral gyrus. Magic contrasts. To determine whether the magic manipulation also increased activity in the TPJ, we examined neural responses in the TPJ ROI during the game phase of the task for magic trials (Magic + TurbMagic) compared to no magic trials (Control + Turbulence). The results revealed several small clusters of activation in the left and right TPJ; however, these clusters did not survive correction. The corresponding whole-brain analysis showed increased activation in the bilateral putamen/piriform cortex and the right precuneus. The reverse contrast (i.e., performance during the no magic > magic trials), did not reveal any suprathreshold brain regions. For the judgments phase of the task, the magic > no magic contrast revealed a significant cluster of activation in the right lateral prefrontal cortex. The reverse contrast did not reveal any suprathreshold brain regions. The weak effects of Magic, relative to Turbulence, may correspond to the relatively weak behavioral effects that we observed for this variable (see above). Turbulence × Magic contrasts. To determine whether there were any regions of the TPJ that responded more strongly to the combined effects of Turbulence and Magic than to either manipulation alone, we examined neural responses in the TPJ ROI during the game phase of the task for the TurbMagic and Control trials compared to the Turbulence and Magic trials. The results of this analysis and of the corresponding whole brain analysis did not reveal any significant clusters. The reverse contrast (which would indicate regions that exhibited interference between the manipulations) revealed several minute clusters of activation in the right TPJ; however, these clusters did not survive correction. In addition, the corresponding whole-brain analysis did not NEURAL CORRELATES OF METACOGNITION OF AGENCY 19 reveal any suprathreshold brain regions. For the judgments phase of the task, neither of the two interaction contrasts revealed significant clusters of activation. Judgment of agency versus judgment of performance contrasts. To determine whether there was increased activity in brain regions associated with self-reflection when participants made judgments of agency compared to when they made judgments of performance, we examined neural responses in the aPFC ROI during the judgment phase of the task for judgment of agency trials compared to judgment of performance trials. The results of the analysis revealed a significant area of activation (-20, 50, 20; Z = 3.09; KE = 100; p < .05, cluster-corrected) that was part of a larger cluster (KE = 261) in the left aPFC (see Figure 4). In addition, a cluster was found in the right orbitofrontal cortex (OFC; i.e., BA 11) slightly outside of the aPFC ROI (18, 46, -12; KE = 97; p < . 0001, uncorrected). These activations suggest that judgments of agency may involve a greater degree of self-referential thought than typical metacognitive judgments. The whole brain-analyses for the contrasts of judgment of agency > judgment of performance and judgment of performance > judgment of agency did not reveal any suprathreshold brain regions. Parametric model. Regions during the game phase that were negatively correlated with sense of agency. The analysis examining the particular brain regions whose activity during the game phase was negatively correlated with participants’ reported sense of agency (i.e., judgments of agency minus hit rate) while controlling for actual performance did not yield any suprathreshold clusters of activation. However, small clusters of activity were found in the right (62, -36, 18; KE = 26; p < . 005, uncorrected) and left sides (-50, -40, 24; KE = 12; p < .005, uncorrected) of the TPJ ROI. In addition, a conjunction between the uncorrected map from the whole-brain analysis (p < .005) and the results of the turbulence > no turbulence contrast revealed that these clusters were part of common activity in the right (64, -36, 16; KE = 52; p < .005, uncorrected) and left TPJ (-46, -38, 18; NEURAL CORRELATES OF METACOGNITION OF AGENCY 20 KE = 85; p < .005, uncorrected). Thus, in accordance with previous findings, it appears that activity in the TPJ is associated with disruptions of control. As a control, a similar analysis of the regions that were negatively correlated with the difference between participants’ judgments of performance and their hit rate yielded a significant cluster in the medial precuneus/posterior cingulate cortex. And although a small cluster of activity in the TPJ ROI was also found (62, -20, 30, KE = 41; p < .005, uncorrected), a conjunction with the turbulence > no turbulence contrast showed that this cluster was not part of common activity in the TPJ (the conjunction did, however, show a small cluster of common activity in a neighboring region; 64, -36, 12; KE = 21; p < .005, uncorrected). Regions during the game phase that were positively correlated with sense of agency. The analysis examining the particular brain regions whose activity during the game phase was positively correlated with participants’ reported sense of agency while controlling for actual performance yielded significant clusters of activation in the right putamen/thalamus, bilateral RCZ/pre-SMA, right pre/postcentral gyrus, and right calcarine/precuneus (see Figure 5 and Table 3). A similar analysis of the regions that were positively correlated with the difference between judgments of performance and hit rate (which again served as a control) did not yield any suprathreshold clusters. Discussion In the present study, we dissociated the neural correlates of two distinct components of people’s sense of agency: the action monitoring that people engage in as they perform an action, and the metacognitive judgments of agency that they make when retrospectively assessing the extent of their control. These results are consistent with a model of agency (Georgieff & Jeannerod, 1998; Haggard & Tsakiris, 2009; Metcalfe et al., 2010; Metcalfe & Green, 2007; Synofzik et al., 2008; Wegner, 2002, 2003; Wegner, Sparrow, & Winerman, 2004) that says that people form their metacognitive judgments of agency by reflecting on the output of their action monitoring, as well as NEURAL CORRELATES OF METACOGNITION OF AGENCY 21 other relevant cues, and then consciously inferring the extent to which they caused the action outcomes in question. Previous studies, like the present study, used judgments of agency to identify neural activity that is associated with action monitoring; however, unlike the present study, they did not attempted to determine whether there are specific patterns of neural activity associated with the judgments of agency themselves. The present study addressed this gap in the literature by using an agency paradigm that not only allowed us to examine the neural activity associated with the presence or absence of disruptions of control during the performance of a motor task, but that also enabled us to determine which brain regions were active after the motor task had been completed and participants reflected back on how much control they had had. Disruptions of Control The results of both the simple contrast and parametric analyses showed increased activity in bilateral regions of the TPJ when the consequences of participants’ actions were not in line with what they intended or expected them to be (for example, when turbulence caused the cursor to move to the left after a participant moved the mouse to the right). An ROI analysis showed that these regions overlapped with brain areas identified in a meta-analysis of similar contrasts from previous agency studies, as is shown in Figure 3. One explanation for the role of the TPJ in agency processing comes from the comparator model of action monitoring, which proposes a mechanism by which the predicted consequences of an intended action are compared to sensory feedback about the actual consequences of the action (Blakemore et al., 2002; Hohwy & Frith, 2004; Wolpert & Ghahramani, 2000; cf. Synofzik et al., 2008). When the predicted and actual consequences do not match, the comparator sends a signal to the control mechanism responsible for executing motor corrections indicating that there has been a disruption of control. Because activity in the TPJ and its surrounding areas tends to be associated with disruptions of control, it is possible that this region is responsible for either monitoring the NEURAL CORRELATES OF METACOGNITION OF AGENCY 22 comparator signal or for executing corresponding motor corrections. Research on motor planning in monkeys (Mulliken, Musallam, & Andersen, 2008) indicates that the posterior parietal cortex (PPC; which is adjacent to the TPJ) uses multisensory input (including visual, auditory, and vestibular information) and efference copies of motor commands to predict the outcomes of self-initiated actions (e.g., the movement of a limb or the expected position of a cursor). Furthermore, TMS and lesion studies in humans (e.g., MacDonald & Paus, 2003) demonstrate that the PPC plays a causal role in detecting unexpected outcomes of self-generated movements. These and other findings provide converging evidence that increased TPJ activity may be associated with detecting mismatches between forward estimates and sensory feedback. However, the possibility that TPJ activity is also associated with executing motor corrections cannot be entirely ruled out. Although a number of previous agency studies showed that TPJ activation was present during monitoring when control was held constant (e.g., when participants merely observed disruptions of control and were not given the opportunity to perform corrective actions; e.g., Spengler, von Cramon, & Brass, 2009), none of these studies attempted to determine whether this TPJ activity was absent during control when monitoring was held constant. Furthermore, because the game phase of each trial in the present experiment involved a continuous stream of disruptions and corrective movements, our study is also unable to distinguish between these two interpretations of the observed TPJ activity. Being In Control Although previous agency studies (e.g., Farrer et al., 2003; Tsakiris, Longo, & Haggard, 2010) have occasionally identified patterns of brain activity that are associated with increases in control, these patterns have not been consistent across studies—suggesting, perhaps, that feeling in control is really just the absence of feeling out of control (Hohwy, 2007). However, the lack of a consistent neural signature for 'being in control' might have been due to the fact that prior studies did not control for variation in task performance (which tends to be strongly correlated with NEURAL CORRELATES OF METACOGNITION OF AGENCY 23 perceived agency). The results of the parametric analysis we conducted, which did control for task performance, yielded significant clusters of activation in the right putamen/thalamus, bilateral RCZ/pre-SMA, right pre/postcentral gyrus, and right calcarine/precuneus (see David et al., 2007, for a similar pattern of results). It is of interest, and perhaps even intuitive, that these are the same regions that have been indentified in studies of smoothly-executed, self-initiated actions (Boecker, Jankowski, Ditter, & Scheef, 2008; Cunnington, Windischberger, Deecke, & Moser, 2002; Soon, Brass, Heinze, & Haynes, 2008; Walton, Devlin, & Rushworth, 2004; and see Nachev, Kennard, & Hussain, 2008, for a review). Metacognitive Judgments of Agency: Self-Reflective Processing The primary new finding of the present study is that the left aPFC and right OFC increased in activation when participants made judgments of agency compared to when they made judgments of performance. Prior research on self-attribution and self-reflection indicates that these areas are part of cortical midline structures that subserve self-reflective processing. According to recent metaanalyses and theoretical reviews of published fMRI and PET studies on self-reflection, these structures can be segregated into several functionally distinct regions (Northoff et al., 2006; Schmitz & Johnson, 2007; van der Meer et al., 2010). The most ventral region, which includes portions of the aPFC, OFC, and the pACC, receives projections from a number of sensory and emotional processing centers such as the limbic system and the striatum (Ongur & Price, 2000) and, thus, is thought to be involved in the appraisal of sensory and affective stimuli in terms of their selfrelevance (Ochsner et al., 2004; Zysset et al., 2002; see Amodio and Frith, 2006, for a review). The more dorsal midline structures, including the dorsomedial prefrontal cortex (dMPFC), are tightly connected to areas of the lateral PFC that are involved in high-level reasoning and decision-making and, therefore, are believed to underlie the conscious formation of inferences and judgments based on information that may have been tagged as self-relevant (or not) by the aPFC/OFC. Finally, the NEURAL CORRELATES OF METACOGNITION OF AGENCY 24 posterior midline structures are considered part of a hippocampal network that is implicated in the encoding and retrieval of autobiographical memories and, thus, are thought to be responsible for putting self-relevant information within a temporal context that is informed by prior experience. The location of the brain activity observed for judgments of agency compared to judgments of performance suggests that judgments of agency may involve more self-related sensory-affective processing compared to other types of metacognitive judgments. This is consistent with a number of theories of agency (e.g., Haggard & Tsakiris, 2009; Metcalfe & Green, 2007; Synofzik et al., 2008). An alternative explanation for the present results could be that activity in the aPFC/OFC reflects a more general process of evaluating information about individuals, regardless of whether or not the information pertains to the self (see Amodio & Frith, 2006). Although a number of studies have shown that inferences about other people’s mental states and personality traits are typically associated with activity in the dMPFC, some of these studies have shown that such nonself-related inferences occasionally activate the aPFC/OFC as well (see Mitchell, Macrae, & Banaji, 2006, for a review). It is, however, important to note that in many of the studies that reported aPFC/OFC activation, participants were asked to mentalize about friends and similar others as opposed to strangers. A recent study (Mitchell et al., 2006) that specifically examined this distinction suggests that mentalizing about close or similar others is informed by knowledge about one’s own thoughts and feelings, which implies that the medial aPFC/OFC is indeed dedicated to processing self-relevant information. Additional support for this conclusion comes from a recent meta-analysis of studies in which judgments about the self were contrasted with judgments about others (van der Meer et al., 2010) which showed that activity in the aPFC/OFC is more strongly associated with judgments about the self. Finally, a study by Powell, Macrae, Cloutier, Metcalfe, and Mitchell (2009) seems, at first glance, to be inconsistent with the present results. Participants in this study completed a selfNEURAL CORRELATES OF METACOGNITION OF AGENCY 25 reflection task, in which they judged their own or another person’s personality traits, as well as an agency task in which they freely chose words to study or watched passively as words were selected for them by the computer. Although the results showed increased activation in the aPFC when participants judged their own traits as opposed to another’s, no ventral activity was found when they actively chose words to study (compared to when the computer chose for them). Instead, freely choosing words was associated with increased activity in the intraparietal sulcus (IPS). However, because choosing is quite different from retrospectively assessing how much control one had (i.e., it is an act and not a reflection), the fact that the choice task was not associated with aPFC/OFC activity may actually be consistent with the present findings. In fact, considering that the IPS overlaps with the TPJ, it appears that making voluntary choices may have more to do with action and action monitoring than with metacognition of agency. Conclusion Although a number of studies have attempted to identify the brain regions that underlie the online monitoring of action, ours is the first to dissociate such low-level action monitoring from metacognitive assessments of agency. While the results presented here were consistent with previous results on action monitoring, it is not action monitoring, but rather metacognition of agency, that is implicated in the attributions that underlie our feelings of responsibility for our own actions. The neural underpinnings of these metacognitions of agency have not been previously demonstrated. We found—by contrasting the neural activity associated with judging agency to the activity associated with a different type of metacognitive judgment (i.e., judging performance)— that there is a specific area (an area distinct from the regions involved in action monitoring, and related instead to self-referential processing) that is central to people's conscious, retrospective judgments of their own agency. NEURAL CORRELATES OF METACOGNITION OF AGENCY26 ReferencesAmodio, D. M., & Frith, C. D. (2006). Meeting of minds: The medial frontal cortex and socialcognition. Nature Reviews Neuroscience, 7, 268-277.Balslev, D., Nielsen, F. A., Lund, T. E., Law, I., & Paulson, O. B. (2006). Similar brain networksfor detecting visuo-motor and visuo-proprioceptive synchrony. Neuroimage, 31, 308-312. Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (1999). Spatio-temporal prediction modulates theperception of self-produced stimuli. Journal of Cognitive Neuroscience, 11, 551-559.Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (2001). The cerebellum is involved in predictingthe sensory consequences of action. Neuroreport, 12,1879-1884.Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (2002). Abnormalities in the awareness of action.Trends in Cognitive Science, 6, 237-242.Boecker, H., Jankowski, J., Ditter, P., & Scheef, L. (2008). A role of the basal ganglia and midbrainnuclei for initiation of motor sequences. Neuroimage, 39, 1356-1369.Chaminade, T., & Decety, J. (2002). Leader or follower? Involvement of the inferior parietal lobulein agency. Neuroreport, 13, 1975-1978.Christoff, K., & Gabrielli, J. D. E. (2000). The frontopolar cortex and human cognition: Evidencefor a rostrocaudal hierarchical organization within the human prefrontal cortex.Psychobiology, 28, 168-186.Cunnington, R., Windischberger, C., Deecke, L., & Moser, E. (2002). The preparation andexecution of self-initiated and externally-triggered movement: A study of event-relatedfMRI. Neuroimage, 15, 373-385.David, N., Bewernick, B. H., Cohen, M. X., Newen, A., Lux, S., Fink, G. R., et al (2006). Neuralrepresentations of self versus other: Visual-spatial perspective taking and agency in a virtualball-tossing game. Journal of Cognitive Neuroscience, 18, 898–910. NEURAL CORRELATES OF METACOGNITION OF AGENCY27 David, N., Cohen, M. X., Newen, A., Bewernick, B. H., Shah, N. J., Fink, G. R., & Vogeley, K.(2007). The extrastriate cortex distinguishes between the consequences of one's own andothers' behavior. Neuroimage, 36, 1004-1014.David, N., Newen, A., & Vogeley, K. (2008). The "sense of agency" and its underlying cognitiveand neural mechanisms. Consciousness and Cognition, 17, 523-534. Decety, J., Chaminade, T., Grezes, J., & Meltzoff, A. N. (2002). A PET exploration of the neuralmechanisms involved in reciprocal imitation. Neuroimage, 15, 265-272.Decety, J., & Lamm, C. (2007). The role of the right temporoparietal junction in social interaction:how low-level computational processes contribute to meta-cognition. Neuroscientist, 13,580-593.Farrer, C., Franck, N., Frith, C. D., Decety, J., Georgieff, N., d'Amato, T., et al. (2004). Neuralcorrelates of action attribution in schizophrenia. Psychiatry Research, 131, 31-44.Farrer, C., Franck, N., Georgieff, N., Frith, C. D., Decety, J., & Jeannerod, M. (2003). Modulatingthe experience of agency: A positron emission tomography study. Neuroimage, 18, 324-333.Farrer, C., Frey, S. H., Van Horn, J. D., Tunik, E., Turk, D., Inati, S., et al. (2008). The angulargyrus computes action awareness representations. Cereb Cortex, 18, 254-261.Farrer, C., & Frith, C. D. (2002). Experiencing oneself vs another person as being the cause of anaction: The neural correlates of the experience of agency. Neuroimage, 15, 596-603.Fink, G. R., Marshall, J. C., Halligan, P. W., Frith, C. D., Driver, J., Frackowiak, R. S., et al. (1999).The neural consequences of conflict between intention and the senses. Brain, 122, 497-512. Franck, N., Farrer, C., Georgieff, N., Marie-Cardine, M., Dalery, J., d’Amato, T., & Jeannerod, M. (2001). Defective recognition of one’s own actions in patients with schizophrenia. American Journal of Psychiatry, 158, 454-459. NEURAL CORRELATES OF METACOGNITION OF AGENCY28 Friston, K., Worsley, K., Frackowiak, R., Mazziotta, J., & Evans, A. (1993). Assessing thesignificance of focal activations using their spatial extent. Human Brain Mapping, 1, 210220.Friston, K. J., Holmes, A. P., Poline, J. B., Grasby, P. J., Williams, S. C., Frackowiak, R. S., et al.(1995). Analysis of fMRI time-series revisited. Neuroimage, 2, 45-53. Georgieff, N., & Jeannerod, M. (1998). Beyond consciousness of external reality. A “Who” systemfor consciousness of action and self-consciousness. Consciousness and Cognition, 7, 465–477.Haggard, P., & Tsakiris, M. (2009). The experience of agency. Current Directions in PsychologicalScience, 18, 242-246.Hohwy, J. (2007). The sense of self in the phenomenology of agency and perception. Psyche, 13(1).Retrieved from http://www.theassc.org/files/assc/2668.pdfHohwy, J., & Frith, C. (2004). Can neuroscience explain consciousness? Journal of ConsciousnessStudies, 11, 180-198.Jenkins, A. C., & Mitchell, J. P. (2010). Medial prefrontal cortex subserves diverse forms of self-reflection. Social Neuroscience.Kable, J. W., & Chatterjee, A. (2006). Specificity of action representations in the lateraloccipitotemporal cortex. Journal of Cognitive Neuroscience, 18, 1498-1517.Kirkpatrick, M., Metcalfe, J., Greene, M., & Hart, C. (2008). Effects of intranasalmethamphetamine on metacognition of agency. Psychopharmacology, 197, 137–144.Leube, D. T., Knoblich, G., Erb, M., Grodd, W., Bartels, M., & Kircher, T. T. (2003). The neuralcorrelates of perceiving one's own movements. Neuroimage, 20, 2084-2090.Lorch, R., & Myers, J. (1990). Regression analyses of repeated measures data in cognitive research.Journal of Experimental Psychology: Learning, Memory, and Cognition, 16, 149-157. NEURAL CORRELATES OF METACOGNITION OF AGENCY29 MacDonald, P. A., & Paus, T. (2003). The role of parietal cortex in awareness of self-generatedmovements: A transcranial magnetic stimulation study. Cerebral Cortex, 13, 962-967. Metcalfe, J., Eich, T. S., & Castel, A. D. (2010). Metacognition of agency across the lifespan.Cognition, 116, 267-282.Metcalfe, J., & Greene, M. J. (2007). Metacognition of agency. Journal of ExperimentalPsychology: General, 136, 184-199.Mitchell, J. P. (2009). Social psychology as a natural kind. Trends in Cognitive Science, 13, 246-251.Mitchell, J. P., Macrae, C. N., & Banaji, M. R. (2006). Dissociable medial prefrontal contributionsto judgments of similar and dissimilar others. Neuron, 50, 655-663.Mulliken, G. H., Musallam, S., & Andersen, R. A. (2008). Forward estimation of movement state inposterior parietal cortex. Proceedings of the National Academy of Sciences, 105, 8170-8177.Mumford, J. A., & Nichols, T. (2009). Simple group fMRI modeling and inference. Neuroimage,47, 1469-1475.Nachev, P., Kennard, C., & Husain, M. (2008). Functional role of the supplementary and pre-supplementary motor areas. Nature Reviews Neuroscience, 9, 856-869.Nahab, F. B., Kundu, P., Gallea, C., Kakareka, J., Pursley, R., Pohida, T., et al. (in press). Theneural processes underlying self-agency. Cerebral Cortex.Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., & Panksepp, J. (2006).Self-referential processing in our brain--a meta-analysis of imaging studies on the self.Neuroimage, 31, 440-457.Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S., Gabrieli, J. D., et al. (2004).For better or for worse: Neural systems supporting the cognitive downand up-regulation ofnegative emotion. Neuroimage, 23, 483-499. NEURAL CORRELATES OF METACOGNITION OF AGENCY30 Ongur, D., & Price, J. L. (2000). The organization of networks within the orbital and medialprefrontal cortex of rats, monkeys and humans. Cerebral Cortex, 10, 206-219. Pacherie, E. (in press). Sense of agency: Many facets, multiple sources. In H. S. Terrace & J.Metcalfe (Eds.), Agency and Joint Attention. Oxford: Oxford University Press.Pisella, L., Grea, H., Tilikete, C., Vighetto, A., Desmurget, M., Rode, G., et al. (2000). An 'automatic pilot' for the hand in human posterior parietal cortex: toward reinterpreting opticataxia. Nature Neuroscience, 3, 729-736.Powell, L. J., Macrae, C. N., Cloutier, J., Metcalfe, J., & Mitchell, J. P. (2009). Dissociable neuralsubstrates for agentic versus conceptual representations of self. Journal of CognitiveNeuroscience, 22, 2186-2197.Ramnani, N., & Miall, R. C. (2004). A system in the human brain for predicting the actions ofothers. Nature Neuroscience, 7, 85-90.Ramnani, N., & Owen, A. M. (2004). Anterior prefrontal cortex: Insights into function fromanatomy and neuroimaging. Nature Reviews Neuroscience, 5, 184-194.Ruby, P., & Decety, J. (2001). Effect of subjective perspective taking during simulation of action: APET investigation of agency. Nature Neuroscience, 4, 546-550.Saxe, R., Xiao, D. K., Kovacs, G., Perrett, D. I., & Kanwisher, N. (2004). A region of rightposterior superior temporal sulcus responds to observed intentional actions.Neuropsychologia, 42, 1435-1446.Schmitz, T. W., & Johnson, S. C. (2007). Relevance to self: A brief review and framework of neuralsystems underlying appraisal. Neuroscience & Biobehavioral Reviews, 31, 585-596.Schnell, K., Heekeren, K., Schnitker, R., Daumann, J., Weber, J., Hesselmann, V., et al. (2007). AnfMRI approach to particularize the frontoparietal network for visuomotor action monitoring: NEURAL CORRELATES OF METACOGNITION OF AGENCY31 Detection of incongruence between test subjects' actions and resulting perceptions.Neuroimage, 34, 332-341. Soon, C. S., Brass, M., Heinze, H. J., & Haynes, J. D. (2008). Unconscious determinants of freedecisions in the human brain. Nature Neuroscience, 11, 543-545.Spence, S. A., Brooks, D. J., Hirsch, S. R., Liddle, P. F., Meehan, J., & Grasby, P. M. (1997). A PET study of voluntary movement in schizophrenic patients experiencing passivityphenomena (delusions of alien control). Brain, 120, 1997-2011.Spengler, S., von Cramon, D. Y., & Brass, M. (2009). Was it me or was it you? How the sense ofagency originates from ideomotor learning revealed by fMRI. Neuroimage, 46, 290-298.Synofzik, M., Vosgerau, G., & Newen, A. (2008). Beyond the comparator model: A multifactorialtwo-step account of agency. Conscious and Cognition, 17, 219-239.Tricomi, E. M., Delgado, M. R., & Fiez, J. A. (2004). Modulation of caudate activity by actioncontingency. Neuron, 41, 281-292.Tsakiris, M., Longo, M. R., & Haggard, P. (2010). Having a body versus moving your body: Neuralsignatures of agency and body-ownership. Neuropsychologia, 48, 2740-2749.Tzourio-Mazoyer, N., Landeau, B., Papathanassiou, D., Crivello, F., Etard, O., Delcroix, N., et al.(2002). Automated anatomical labeling of activations in SPM using a macroscopicanatomical parcellation of the MNI MRI single-subject brain. Neuroimage, 15, 273-289.van der Meer, L., Costafreda, S., Aleman, A., & David, A. (2010). Self-reflection and the brain: Atheoretical review and meta-analysis of neuroimaging studies with implications forschizophrenia. Neuroscience & Biobehavioral Reviews, 34, 935-946.Vinogradov, S., Luks, T. L., Simpson, G. V., Schulman, B. J., Glenn, S., & Wong, A. E. (2006).Brain activation patterns during memory of cognitive agency. Neuroimage, 31, 896-905. NEURAL CORRELATES OF METACOGNITION OF AGENCY32 Wager, T. D., Lindquist, M. A., Nichols, T. E., Kober, H., & Van Snellenberg, J. X. (2009).Evaluating the consistency and specificity of neuroimaging data using meta-analysis.Neuroimage, 45, S210-221.Walton, M., Devlin, J., & Rushworth, M. (2004). Interactions between decision making andperformance monitoring within prefrontal cortex. Nature Neuroscience, 7, 1259-1265. Wegner, D. M. (2002). The illusion of conscious will. Cambridge, MA: MIT Press.Wegner, D. M. (2003). The mind’s best trick: How we experience conscious will. Trends inCognitive Sciences, 7, 65–69.Wegner, D. M., Sparrow, B., & Winerman, L. (2004). Vicarious agency: Experiencing control overthe movements of others. Journal of Personality and Social Psychology, 86, 838–848.Williams, J. H., Waiter, G. D., Gilchrist, A., Perrett, D. I., Murray, A. D., & Whiten, A. (2006).Neural mechanisms of imitation and 'mirror neuron' functioning in autistic spectrumdisorder. Neuropsychologia, 44, 610-621.Wolpert, D. M., & Ghahramani, Z. (2000). Computational principles of movement neuroscience.Nature Neuroscience, 3(Suppl.), 1212–1217.Worsley, K. J., Marrett, S., Neelin, P., Vandal, A. C., Friston, K. J., & Evans, A. C. (1996). Aunified statistical approach for determining significant signals in images of cerebralactivation. Human Brain Mapping, 4, 58-73.Yomogida, Y., Sugiura, M., Sassa, Y., Wakusawa, K., Sekiguchi, A., Fukushima, A., et al. (2010).The neural basis of agency: An fMRI study. Neuroimage, 50, 198-207.Zysset, S., Huber, O., Ferstl, E., & von Cramon, D. Y. (2002). The anterior frontomedian cortex andevaluative judgment: An fMRI study. Neuroimage, 15, 983-991. NEURAL CORRELATES OF METACOGNITION OF AGENCY33

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Dissociating Neural Correlates of Action Monitoring and Metacognition of Agency

Judgments of agency refer to people's self-reflective assessments concerning their own control: their assessments of the extent to which they themselves are responsible for an action. These self-reflective metacognitive judgments can be distinguished from action monitoring, which involves the detection of the divergence (or lack of divergence) between observed states and expected states. Presum...

متن کامل

Neural Correlates of Craving in Methamphetamine Abuse

Introduction: Methamphetamine is a powerful psychostimulant that causes significant neurological&nbsp;impairments with long-lasting effects and has provoked serious international concerns about public&nbsp;health. Denial of drug abuse and drug craving are two important factors that make the diagnosis&nbsp;and treatment extremely challenging. Here, we present a novel and rapid noninvasive method...

متن کامل

Neural Correlates of Boredom in Music Perception

Introduction: Music can elicit powerful emotional responses, the neural correlates of which have not been properly understood. An important aspect about the quality of any musical piece is its ability to elicit a sense of excitement in the listeners. In this study, we investigated the neural correlates of boredom evoked by music in human subjects. Methods: We used EEG recording in nine sub...

متن کامل

Metacognitive Facilitation of Spontaneous 3 Thought Processes : When Metacognition 4 Helps the Wandering Mind Find Its Way

8 undirected, spontaneous thought processes that often occur without our volition 9 and sometimes despite our intentions. Metacognition, by contrast, involves the 10 conscious, often intentional monitoring and evaluation of our own mental pro11 cesses and behaviors. The neural correlates of MW and metacognition may also 12 appear strictly distinct at first, considering the almost exclusive focu...

متن کامل

Metacognition in Early Phase Psychosis: Toward Understanding Neural Substrates

Individuals in the early phases of psychotic illness have disturbed metacognitive capacity, which has been linked to a number of poor outcomes. Little is known, however, about the neural systems associated with metacognition in this population. The purpose of this study was to elucidate the neuroanatomical correlates of metacognition. We anticipated that higher levels of metacognition may be de...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2011