Brain structures activated by overt and covert emotional visual stimuli

Brain structures activated by overt and covert emotional visual stimuli

Brain Research Bulletin 79 (2009) 258–264 Contents lists available at ScienceDirect Brain Research Bulletin journal homepage: www.elsevier.com/locat...

674KB Sizes 0 Downloads 138 Views

Brain Research Bulletin 79 (2009) 258–264

Contents lists available at ScienceDirect

Brain Research Bulletin journal homepage: www.elsevier.com/locate/brainresbull

Research report

Brain structures activated by overt and covert emotional visual stimuli Elisabetta Sabatini a,b,∗ , Stefania Della Penna c , Raffaella Franciotti c , Antonio Ferretti c , Pierluigi Zoccolotti d,e , Paolo M. Rossini f,b,g , Gian Luca Romani c , Guido Gainotti a a

Servizio di Neuropsicologia, Università Cattolica “Sacro Cuore” di Roma, Policlinico A. Gemelli, Largo A. Gemelli 8, 00168 Rome, Italy AfaR Dipartimento di Neuroscienze, Ospedale Fatebenefratelli, Isola Tiberina 39, 00186 Rome, Italy c Dipartimento di Scienze Cliniche e Bioimmagini, Università “G. D’Annunzio” and Istituto di Tecnologie Biomediche Avanzate, Fondazione dell’Università “G. D’Annunzio” di Chieti, Via dei Vestini 33, 66100 Chieti, Italy d IRCCS Fondazione Santa Lucia, Via Ardeatina 306, 00179 Rome, Italy e Dipartimento di Psicologia, Università degli Studi “La Sapienza”, Via dei Marsi 78, 00100 Rome, Italy f IRCCS Centro San Giovanni di Dio, Fatebenefratelli, Via Pilastroni 4, 25125 Brescia, Italy g Neurologia Clinica, Università “Campus Biomedico”, Via Emilio Longoni 83, 00155 Rome, Italy b

a r t i c l e

i n f o

Article history: Received 27 May 2008 Received in revised form 9 March 2009 Accepted 9 March 2009 Available online 19 March 2009 Keywords: Anterior insula Amygdala-related cortical network fMRI Unconscious perception Emotional faces Pain processing

a b s t r a c t Research data suggest that the amygdala and some related brain structures modulate the processing of emotional visual stimuli even when they are not consciously perceived. In this study, we examined neural responses to investigate whether and how other brain areas anatomically connected to the amygdala might become activated during both overt and covert presentation of conditioned emotional visual stimuli. In the covert presentation, a conditioned angry face was shown for 15 ms followed by a neutral masking face (CSmask). In the overt condition, an angry face associated with a painful stimulus (CS+), a happy (H) and a neutral face (N) were presented for 75 ms. Based on results of functional magnetic resonance imaging (fMRI) in 10 healthy volunteers, we show evidence that a network of brain structures anatomically connected to the amygdala (including the anterior insula, the fusiform gyrus and the superior temporal sulcus) are involved in the subliminal processing of visual emotional stimuli. Of particular interest was the dissociation between the anterior and posterior insula: the anterior insula responded to both overt and covert presentation of the conditioned stimulus, whereas the posterior insula responded only to the overt presentation of the face associated with a painful electrical stimulation. This response pattern suggests that the anterior insula, the fusiform gyrus and the temporal sulcus cooperate with the amygdala in the unconscious processing of pain-conditioned stimuli. © 2009 Elsevier Inc. All rights reserved.

1. Introduction In neuroscience research, there is increasing interest in elucidating the neural substrates of conscious and unconscious aspects of emotional learning [22,51,52]. In a classical aversive conditioning paradigm, activity in the amygdala might contribute to an emotional response that has not been consciously perceived [43,44]. Other studies confirmed that masked presentations [60] of emotional facial expressions modulate amygdala activity without explicit knowledge [35,75,57]. It is not known, however, whether

∗ Corresponding author. Tel.: +39 06 35501945; fax: +39 06 35501909. E-mail addresses: [email protected], [email protected] (E. Sabatini), [email protected] (S.D. Penna), [email protected] (R. Franciotti), [email protected] (A. Ferretti), [email protected] (P. Zoccolotti), [email protected] (P.M. Rossini), [email protected] (G.L. Romani), [email protected] (G. Gainotti). 0361-9230/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.brainresbull.2009.03.001

other brain structures closely connected with the amygdala participate in this unconscious processing of emotional stimuli. There are anatomical and functional reasons why the insula (in particular, its anterior part), the fusiform gyrus (FG), the superior temporal sulcus (STS) and, perhaps, the prefrontal cortex (PFC) and the cingulate gyrus (CG) are good candidates for this partnership. The insula, which is located in the paralimbic cortex in both monkeys and humans [42], is strongly interconnected with limbic structures and plays a key role in emotional functions involved in reactions to both internal and external stimuli [47]. There is also direct evidence for involvement of the insula in human pain processing [1,50,73]; indeed, some authors have shown that the insula is involved in the conscious perception of cognitive and emotional pain [53,7]. Furthermore, Simmons et al. [67,68] observed high insula activation in anxiety-prone subjects during anticipation of aversive visual stimuli, Büchel et al. [11] showed that the anterior insula is implicated in delayed fear conditioning, and Morris et al. [43] observed enhanced activity in the anterior insula during the processing of fearful faces. Finally, the existence

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

of an interaction between the amygdala and the insular cortex was shown in a recent study in which the amygdala and the insula showed parallel activation during an emotional face assessment task and parallel deactivation after lorazepam administration [48]. The fusiform gyrus (FG), which is part of the ventral stream of visual processing, includes the fusiform face area (FFA), which is specialized for facial recognition [29,21,33,34]. Studies that investigated the response of the temporal cortex to facial expressions found that the FG responds more to emotional than to neutral faces [10,23,74,76]. Dolan et al. [23] suggested that this effect might be modulated by connections from the amygdala to the FG. Studies in humans and in nonhuman primates [45,16,5] indicate that the superior temporal sulcus (STS) is involved in processing complex visual stimuli (including facial movements) and in the multimodal integration between visual and auditory information. Its activity is particularly increased during auditory-visual speech perception [17,32,61]; however, some results suggest that the STS is also activated by visual emotional stimuli [24,30]. From a neuroanatomical point of view, Stefanacci and Amaral [69] showed that rostral regions of the STS have multiple connections with the amygdala, and Augustine [2] stressed the connections between the STS and the insula. From a functional point of view, Sugase et al. [71] demonstrated that neurons in its upper and lower banks respond to pictures of different facial expressions; furthermore, neuroimaging studies showed that the STS is activated during perception of facial and vocal emotional expressions [36,37,54,57,58]. Finally, in both animals and humans, the amygdala has significant connections with the orbital and medial aspects of the prefrontal cortex (PFC) and the cingulate gyrus (CG). The PFC associates information from external and internal sensory modalities (including information about visceral activities) and can thus integrate a variety of inputs pertinent to moment-tomoment experience. Functional neuroimaging experiments have shown that the prefrontal cortex plays a role in the perception of facial expression [9,53,43], that it is preferentially involved in emotions reflecting complex states of mind, rather than basic emotions [66], and that it could be involved in the formation of emotional memories [65]. The CG is implicated in cognitive, emotional, motor and autonomic activities and seems necessary in adapting the autonomic state of arousal to concurrent cognitive and physical demands. It is strongly connected to the amygdala and is involved in integrating emotional and attentional processing [59,79]. Several functional neuroimaging experiments have shown that it is implicated in conditioned fear responses [12] and is activated in response to emotional faces [43,38,35]. The aim of the present study was to investigate activation of the amygdala and of other brain structures involved in processing facial emotional expressions during conscious and unconscious presentation of aversive visual stimuli. In addition to the above-mentioned structures, we also analyzed responses in the second somatosensory area (SII), which is located in the parietal opercular cortex, because painful electrical stimulation was always associated with one of the facial emotional expressions (namely, the aversively conditioned one). SII is critically involved in the processing of painful and non-painful somatosensory inputs [49,13,41]. It is well known that responses to emotionally laden stimuli show a marked pattern of habituation, known as repetition suppression, in the amygdala and related structures (e.g., [31,63,78]). Therefore, care was taken to examine activation to repeated stimuli as a function of the time period of the experimental session. Finally, several studies have reported significant, although inconsistent, hemispheric asymmetries in the amygdala and related structures [4,18,64]. Therefore, the possible role of hemispheric lateralization in modulating emotional responding during conscious and unconscious presentation of aversive visual stimuli was considered.

259

Overall, in this experiment, we used visual stimulations consisting of facial expressions in an aversive conditioning paradigm to clarify the following: (1) whether visual stimuli that are not consciously perceived can be processed at the level of the amygdala and of related brain structures, including the fusiform gyrus (FG), the superior temporal sulcus (SPS), the prefrontal cortex (PFC), the cingulate gyrus (CG) and the anterior insula; and (2) whether a different pattern of activation is observed in SII and in the posterior insula, which should be involved in the perceptual components of pain. We also examined (3) the influence of repetition suppression on the activation observed in these brain structures; finally, bilateral recordings allowed establishing (4) whether there are hemispheric asymmetries in brain activation and in which structures and conditions they are observed. 2. Materials and methods 2.1. Subjects Ten right-handed healthy volunteers, seven females and three males, took part in the study. All subjects (mean age 23.6 years) gave their written informed consent to participate in the study. All procedures were approved by the local ethics committee and were performed in accordance with the ethical standards established in the 1964 Declaration of Helsinki. 2.2. Stimuli Four different target faces (one angry, one happy and two neutral) were selected from a standard set of pictures of facial affect [25] and served as the experimental stimuli. Stimuli presentation was controlled by a MATLAB® code running on a PC placed in the scanner console room. Visual stimuli were projected onto a translucent screen placed at the back of the scanner bore using an LCD projector. A mirror fixed to the head coil inside the magnet allowed subjects to view the translucent screen. 2.3. Task 2.3.1. Conditioning phase The conditioning phase before scanning consisted of a sequence of 80 greyscale images of four faces: one angry (A), one happy (H) and two neutral (N). Images of a single face were presented on a computer screen for 75 ms at 12 s interstimulus intervals. Each of the four faces was shown twenty times in pseudorandom order. The angry expression (conditioned stimulus, CS+) was always associated with a painful electrical stimulation (10–15 mA intensity, 500 ␮s duration) delivered to the wrist of the left hand. None of the other faces was ever paired with the aversive stimulus. Throughout the experiment, the volunteers were instructed to press the mouse button on the right with the middle finger of the right hand if the angry face appeared and the one on the left with the index finger of the right hand for the other faces. No time pressure was given in response; however, a time limit of 10 s was used to accept the response as valid. The task proved very simple and errors were virtually absent. 2.3.2. Scanning phase After a 15 min interval, the fMRI was performed. During scanning, 96 stimuli were shown at 12 s interstimulus intervals. There were four different conditions: CS+, H and N were presented for 75 ms, and a fourth one, consisting of the same angry face, was presented for 15 ms, masked by a neutral face for 75 ms (CSmask; Fig. 1).

Fig. 1. Visual stimulus consisting of an angry face (CSmask) masked by a neutral face. After 12 s of gray screen fixation, the CSmask was presented for 15 ms followed by the neutral face (mask) for 75 ms and by another gray screen for 12 s.

260

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

The subjects’ perception of the angry face was explicit when it was associated with painful stimulation (CS+) and implicit when this conditioned stimulus was masked by a neutral face (CSmask). Following presentation of the single stimuli, a grey screen with a white fixation point was used as the rest condition. As in the conditioning phase, the volunteers pressed the right button if the angry face appeared and the left one for the other faces. 2.4. Imaging Neural activity while viewing emotional stimuli was measured by event-related fMRI. For each subject, Blood Oxygen Level Dependent (BOLD) contrast functional images were acquired using a Siemens Magneton Vision scanner at 1.5 T by means of T2* -weighted echo planar imaging free induction decay sequences (TE 60 ms, matrix size 64 × 64, FOV 256 mm, in-plane voxel size 4 mm × 4 mm, flip angle 90◦ , slice thickness 6 mm and no gap). A total of 870 functional volumes, consisting of 16 transaxial slices, were acquired with TR 2091 ms. A high resolution structural volume was acquired at the end of the session via a 3D MPRAGE sequence: sagittal, matrix 256 × 256, FOV 256 mm, slice thickness 1 mm, no gap, in-plane voxel size 1 mm × 1 mm, flip angle 12◦ , TR = 9.7 ms, TE = 4 ms. FMRI data were analysed using Brain Voyager QX software (Brain Innovation, The Netherlands). Pre-processing of functional data included slice scan time correction, motion correction, and removal of linear and nonlinear trends from voxel time series. Pre-processed functional volumes of each participant were coregistered with the corresponding structural dataset. As the 2D functional and 3D structural measurements were acquired in the same session, the coregistration transformation was determined using the Siemens slice position parameters of the functional images and the position parameters of the structural volume. Structural and functional volumes were transformed into the Talairach space using a piecewise affine and continuous transformation. Functional volumes were resampled at a voxel size of 3 mm × 3 mm × 3 mm. In the fMRI image analysis, the 96 stimuli were divided into 3 subsequent time periods (iterations) to account for possible repetition suppression effects. 2.5. Statistics FMRI analyses were performed by means of Brain Voyager QX software for individual subjects using the General Linear Model (GLM) [28] with correction for temporal autocorrelation [15,77] and considering twelve predictors of interest (4 faces × 3 iterations). To account for haemodynamic delay, the boxcar waveforms representing the rest and stimuli conditions were convolved with an empirically founded haemodynamic response [8]. Statistical activation maps were computed considering the contrasts of the experimental conditions vs. rest. The maps were thresholded at p < 0.0004 at the voxel level, and a minimum cluster size of at least four voxels was required. These thresholds and an estimate of the spatial correlation of voxels were used as input in a Monte Carlo simulation [27,19] in order to assess the overall significance level (the probability of false detection of a cluster in the entire functional volume). In this way, we obtained p < 0.05 as the significance level corrected for multiple comparisons. Thresholded maps were superimposed on the respective structural scans to localize significantly activated areas. ROIs were determined for each subject by considering the mask obtained from voxels activated in any contrast condition vs. rest. The maximum value of BOLD signal intensity variation with respect to rest in each ROI for each condition was calculated. To determine possible differences in activation for CS+ and CSmask for hemispheric predominance and for neural adaptation and extinction effects [12,70,39] we analysed the BOLD% signal change by means of a 3-way (condition, iteration and hemisphere) ANOVA for repeated measures.

3. Results 3.1. Behavioural results During the conditioning phase subjects’ responses were generally extremely accurate; during the scanning phase the masked angry face followed by the neutral one was associated with the right button press for ca. 1.8% of trials. Furthermore, at the end of the experiment, during subsequent debriefing, no subject verbally reported recognising the angry faces before the neutral masking face. 3.2. fMRI results All subjects showed a similar pattern of activation during perception of overt and covert stimuli. Bilateral responses deep within the lateral fissure, corresponding to the insula, were detected. Specifically, we found two different clusters of activation in the insular cortex, one located in the anterior part (orbital surface) and

Table 1 Talairach coordinates of ROIs from the group analysis with a significant contrast between the differential responses to the angry face associated with a painful stimulus (CS+) and the angry face masked to the neutral one (CSmask). STS = superior temporal sulcus; FG = fusiform gyrus; CG = Cingulate cortex; SII = second somatosensory area. Hemisphere and area

Left anterior insula Right anterior insula Left posterior insula Right posterior insula Left prefrontal cortex Right prefrontal cortex Left amygdala Right amygdala Left STS Right STS Left FG Right FG Left CG Right CG Left SII Right SII *

Talairach coordinates x

y

z

−32 35 −38 39 −38 39 −19 14 −38 43 −30 31 −6 7 −46 48

13 16 −1 −2 8 4 −1 −4 −62 −58 −49 −54 −5 1 −23 −27

13 11 12 14 31 32 −10 −13 −4 4 −22 −19 57 45 17 20

Volume (mm3 )

t test* p < 0.01

263 834 636 259 462 512 210 192 521 447 844 678 501 702 776 996

17.51 19.84 13.94 13.18 8.78 17.46 4.16 2.83 7.96 9.35 16.28 15.09 15.63 16.27 14.41 14.13

The t test is reported for the maximum in each cluster.

the other in the posterior part. Also, in addition to the insula, the fusiform gyrus (FG), the superior temporal sulcus (STS), the second somatosensory cortex (SII), the prefrontal cortex (PFC), the cingulate gyrus (CG) and the amygdala were considered regions of interest in our study for reasons detailed in the introductory part of the article. The Talairach coordinates [72] of the cluster centroids as well as the voxel volumes and t test values are shown in Table 1. The coordinates were obtained from the group map. 3.3. ROI analysis results 3.3.1. Amygdala In the amygdala the level of activation was significantly different among conditions (F(3, 27) = 3.24; p < 0.05): activation for CS+ was higher than for H (p < 0.01, Duncan test) and N (p < 0.05) stimuli (which did not differ from each other); activation for CSmask stimuli was intermediate and did not significantly differ from all other conditions (see Fig. 2a). The main effect of iteration was significant (F(2, 18) = 11.80; p < 0.01): activation during the first part of the experimental session was significantly higher than during the second (p < 0.01) that, in turn, was higher than the third (p < 0.05; see Fig. 2b). No other main effect or interaction was significant. 3.3.2. Insula In the anterior insula the level of activation was significantly different among conditions (F(3, 27) = 7.12; p < 0.01). Specifically, activation for CS+ and CSmask did not differ, but both were significantly higher than H and N stimuli (at least p < 0.01; see Fig. 2c). A trend for the iteration main effect was present (F(2, 18) = 3.14; p = 0.06; Fig. 2d): activation decreased significantly from the first to the third part of the experimental session (p < 0.05) No other main effect or interaction was present. In the posterior insula, the different conditions elicited significantly different responses (F(3, 27) = 6.21; p < 0.01; see Fig. 2e). Activation during CS+ was significantly higher than H (p < 0.001) and N (p < 0.001) stimuli; but, unlike the anterior insula, it was also significantly higher than the CSmask stimuli (p < 0.01). The level of activation was higher in the left than in the right hemisphere (F(1, 9) = 7.54; p < 0.05). There was no influence of iteration and no significant interaction.

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

261

Fig. 2. Analysis of the mean cortical activation resulting in different ROIs from the contrast between angry face associated with shock (CS+), angry face masked with neutral (CSmask), happy (H) and neutral (N) face. The maps were thresholded at p < 0.0004 and plots illustrate critical main effects (condition and iteration) and interactions in the various areas of interest. Plot values indicate the means ± S.E. across 10 subjects of the relative Bold signal intensity (% change) of the activations in the amygdala [AMY] for the condition (a) and iteration (b) effects; in the anterior insula [AI] for the condition (c) and iteration (d) effects; in the posterior insula [PI] for the condition effect (e); in the anterior and posterior insula as a function of condition (f); and in the secondary somatosensory cortex [SII] for the condition effect (g). The asterisks mark significant differences between means (* p < 0.05, ** p < 0.01, *** p < 0.001).

To make a direct comparison of the level of activation in the anterior versus the posterior insula during the critical conditions, an ANOVA was carried out on the activation values, with area (anterior, posterior) and condition (CS+ and CSmask) as factors (see Fig. 2f). The level of activation was significantly higher in the anterior than in the posterior insula (F(1, 9) = 8.31; p < 0.01). The interaction between area and condition was significant (F(1, 9) = 15.32; p < 0.01), indicating that the level of activation was greater in the anterior than in the posterior insula for CSmask (p < 0.01) but not for CS+ stimuli (Fig. 2f). 3.3.3. Secondary somatosensory cortex In the SII the level of activation was significantly different among conditions (F(3,27) = 3.75; p < 0.01). Similarly to the posterior insula, the activation for CS+ was higher than all other conditions (for H, p < 0.01; N, p < 0.01 and CSmask, p < 0.01) that did not differ from each other (see Fig. 2g). No other main effect or interaction was significant. 3.3.4. Fusiform gyrus In FG the level of activation was significantly different among conditions (F(3, 27) = 4.90; p < 0.01) and iterations (F(2, 18) = 8.97; p < 0.01). Activation for both CS+ and CSmask was higher than for H (at least p < 0.05) and N (at least p < 0.05) stimuli, whereas activation for CS+ and CSmask did not differ (see Fig. 3a). Furthermore, the activation detected during the first part of the session was significantly higher than that detected during the second (p < 0.01) and third (p < 0.001) parts (Fig. 3b). No other main effect or interaction was significant.

3.3.5. Superior temporal sulcus In STS the level of activation was significantly different among iterations (F(2, 18) = 18.76; p < 0.001). The activation detected during the first part of the session was significantly higher than the second and third (both ps < 0.001) which did not differ from each other (see Fig. 3c). The interaction between conditions and hemispheres was significant (F(3, 27) = 3.71; p < 0.05): the left hemisphere activation followed a pattern similar to the insula and fusiform gyrus, with activity for the CS+ (p < 0.01) and CSmask (p < 0.01) stimuli higher than for the H and N stimuli, while the right hemisphere did not modulate across conditions (see Fig. 3d). No other main effect or interaction was significant. 3.3.6. Prefrontal cortex In the prefrontal cortex (BA44) the level of activation was significantly different among iterations (F(2,18) = 3.71; p < 0.05): activation for the first part of the session was higher than for the third (p < 0.05); the second did not differ from the other two (Fig. 3e). No other main effect or interaction was significant. 3.3.7. Cingulate gyrus No main effect or interaction was significant for CG (the condition means are presented in Fig. 3g). 4. Discussion The main aim of our research was to clarify whether visual stimuli that are not consciously perceived can be processed at the level of the amygdala and other brain structures, such as the (anterior) insula, the FG, the STS, the PFC and the CG, that are involved in pro-

262

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

Fig. 3. Analysis of the mean cortical activation resulting in different ROIs from the contrast between angry face associated with shock (CS+), angry face masked with neutral (CSmask), happy (H) and neutral (N) face. The maps were thresholded at p < 0.0004 and plots illustrate critical main effects (condition and iteration) and interactions in the various areas of interest. Plot values indicate the means ± S.E. across 10 subjects of the relative Bold signal intensity (% change) of the activations in the Fusiform Gyrus [FG] for the condition (a) and iteration (b) effects; in the Superior Temporal Sulcus [STS] for the iteration effect (c) and the condition by hemisphere interaction (d); in the Prefrontal Cortex [PC] for the iteration effect (e) and Cingulate Cortex (CC) for the condition effect (f). The asterisks mark significant differences between means (* p < 0.05, ** p < 0.01, *** p < 0.001).

cessing facial emotional expressions. We also aimed at investigating whether a different pattern of activation could be observed in brain structures, such as SII and the (posterior) insula, that are mainly involved in processing the perceptual aspects of a painful stimulation. Our expectations, based on previous results reported in the literature [46,47,75,56] were confirmed only in part, because only the amygdala tended to be activated more by the conditioned angry face than by the neutral and happy faces. This unexpected result was probably due to the fast habituation process to conditioned stimuli typical of this structure [10,26,64], which was also confirmed in the present study. Furthermore, in fMRI studies the signal-to-noise ratio diminishes substantially at the level of the amygdala owing to the proximity of air-filled sinuses [14]. Both these factors may have hindered the possibility to detect selective responses to subliminal facial expressions in the amygdala. Nevertheless, other brain structures involved in the processing of visual emotional stimuli and anatomically connected to the amygdala (such as the insula, the fusiform gyrus, and the superior temporal sulcus) were activated more by the CS+ and CSmask than by neutral and happy faces. Of particular interest was the dissociation observed within the insula pointing at two different clusters of activation, one located in the anterior part (orbital surface) and the other one in the posterior part. The anterior insula activation was the same during overt viewing of the CS+ and covert presentation of the CSmask, whereas activation of the posterior insula depended only on the presence of the CS+ stimulus associated with a painful electrical stimulation. Conversely, during the other explicit targets (happy and neutral faces), activations in anterior and posterior insula were greatly reduced.

These results are consistent with previous reports on insula responses following a classical aversive conditioning paradigm [53] and with recent reports, based on transcranial magnetic stimulation, somato-sensory evoked potentials and functional neuroimaging studies, that the anterior and posterior parts of the insula are involved in neural circuits subserving different aspects of pain. Certainly, the posterior parts of the insula contribute to processing the sensory dimensions of pain [3,13,6], whereas the anterior parts contribute to the mental representation of pain in Self and others [20,21,53,55,62,40,6,18]. Our results provide further evidence that the anterior insula can be activated by stimuli that are not consciously perceived, as in the case of the amygdala [22] and of other brain structures (such as the fusiform gyrus and the superior temporal sulcus) involved in the processing of visual emotional stimuli. Furthermore, they confirm that the posterior insula processes the sensory dimensions of pain. Results obtained in the anterior insula, the fusiform gyrus, and (limited to left hemisphere activation) the superior temporal sulcus suggest that all these structures may be involved in subliminal processing of emotional stimuli, as previously found for the amygdala [46,47]. This interpretation is supported by findings related to the repetition suppression effect. In agreement with previous fMRI research [31], in this study activation decreased over time not only at the level of the amygdala, but also of the fusiform gyrus, the superior temporal sulcus, the prefrontal cortex and, to a lesser extent, the anterior insula. On the contrary, no decrement of activation over time was observed in the cingulate cortex, SII or posterior insula. These findings suggest the existence of a network (which includes the amygdala) involved in subliminal processing of emotional stim-

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

uli that shows a common pattern of activation and habituation. It is also possible that the activation induced by subliminal emotional stimuli (which, based on the literature, seems to spread from the amygdala to the anterior insula, the fusiform gyrus, and the superior temporal sulcus) may serve to focus attention on noxious physical stimuli in various sensory modalities (visual afferences in the case of the fusiform gyrus, auditory inputs in the case of the superior temporal sulcus and body-related information in the case of the insula). Notably, in this context the only brain structures anatomically related to the amygdala that were not influenced by the subliminal processing of the aversive affective stimulus were the prefrontal cortex and the cingulate gyrus, where no treatment of modality-specific information should occur. Only limited evidence of hemispheric asymmetries was detected. In the posterior insula, greater activation was present on the left side. In the superior temporal sulcus only the left side modulated with conditions. No asymmetries were detected in the amygdala, the FG, the PFC and the CG. Studies investigating hemispheric lateralization in the amygdala and related structures found inconsistent results with effects varying on the basis of various parameters including gender and stimulus valence of the stimuli and temporal dynamics. The pattern of hemispheric lateralization may prove clearer when examined across several studies by means of the meta-analysis approach [18,64] than in single investigations, as in our study. In conclusion, we found evidence that a network of brain structures anatomically connected to the amygdala (including the anterior insula, the fusiform gyrus and the superior temporal sulcus) are involved in the subliminal processing of visual emotional stimuli. Obviously, the fact that in our study the amygdala only tended to be activated more by the conditioned angry face than by neutral and happy faces is not very consistent with the hypothesis of an amygdala-centred network involved in subliminal processing of emotional stimuli. However, the cornerstone of this hypothesis, namely, the assumption that in the amygdala subliminal visual stimuli can elicit an emotional response without being consciously perceived, has been confirmed by many previous investigations [46,47,76,57]. This fact gives greater plausibility to our hypothesis and suggests that it deserves validation in further well-controlled investigations. Acknowledgements This study was supported by grants of the “Istituto Giuseppe Toniolo di Studi Superiori” - Università Cattolica del Sacro Cuore – Milan and AfaR Dipartimento di Neuroscienze, Ospedale Fatebenefratelli, Isola Tiberina – Rome, Italy. We thank Dr. M. Martelli for implementing the stimuli, Dr. D. Arienzo and Dr. M. Caulo for providing invaluable technical assistance, and Dr. M. Brunetti and Dr. L. Marzetti for their support. The author would like to thank especially Prof. Maurizio Romano and Dr. Gianlorenzo Romano for their friendly advice. References [1] A.V. Apkarian, M.C. Bushnell, R.D. Treede, J.K. Zubieta, Human brain mechanisms of pain perception and regulation in health and disease, Eur. J. Pain 9 (2005) 463–484. [2] J.R. Augustine, Circuitry and functional aspects of the insular lobe in primates including humans, Brain Res. Brain Res. Rev. 22 (3) (1996) 229–244. [3] A. Avenanti, D. Bueti, G. Galati, S.M. Aglioti, Transcranial magnetic stimulation highlights the sensorimotor side of empathy for pain, Nat. Neurosci. 8 (7) (2005) 955–960. [4] D. Baas, A. Aleman, R.S. Kahn, Lateralization of amygdala activation: a systematic review of functional neuroimaging studies, Brain Res. Brain Res. Rev. 45 (2) (2004) 96–103. [5] M.S. Beauchamp, K.E. Lee, B.D. Argall, A. Martin, Integration of auditory and visual information about objects in superior temporal sulcus, Neuron 41 (5) (2004) 809–823.

263

[6] F. Benuzzi, F. Lui, D. Duzzi, P.F. Nichelli, C.A. Porro, Does it look painful or disgusting? Ask your parietal and cingulate cortex, J. Neurosci. 28 (4) (2008) 923–931. [7] M. Botvinick, A.P. Jha, L.M. Bylsma, S.A. Fabian, P.E. Solomon, K.M. Prkachin, Viewing facial expressions of pain engages cortical areas involved in the direct experience of pain, NeuroImage 25 (2005) 312–319. [8] G.M. Boynton, S.A. Egel, J.H. Glover, D.J. Heeger, Linear system analysis of functional magnetic imaging in human V1, J. Neurosci. 16 (1996) 4207–4221. [9] J.C. Britton, S.F. Taylor, K.D. Sudheimer, I. Liberzon, Facial expressions and complex IAPS pictures: Common and differential networks, NeuroImage 31 (2) (2006) 906–919. [10] H.C. Breiter, N.L. Etcoff, P.J. Whalen, W.A. Kennedy, S.L. Rauch, R.L. Buckner, M.M. Strauss, S.E. Hyman, B.R. Rosen, Response and habituation of the human amygdala during visual processing of facial expression, Neuron 17 (5) (1996) 875–887. [11] C. Büchel, R.J. Dolan, J.L. Armony, K.J. Friston, Amygdala-hippocampal involvment in human aversive trace conditioning revealed through eventrelated functional magnetic resonance imaging, J. Neurosci. 19 (24) (1999) 10869–10876. [12] C. Büchel, J. Morris, R.J. Dolan, K.J. Friston, Brain systems mediating aversive conditioning: an event-related fMRI Study, Neuron 20 (1998) 947–957. [13] I. Bufalari, T. Aprile, A. Avenanti, F. Di Russo, S.M. Aglioti, Empathy for pain and touch in the human somatosensory cortex, Cereb Cortex 17 (11) (2007) 2553–2561. [14] E. Bullmore, M. Brammer, S.C. Williams, S. Rabe-Hesketh, N. Janot, A. David, J. Mellers, T. Howard, P. Sham, Statistical methods of estimation and inference for functional MR image analysis, Mang. Reson. Med. 35 (1996) 261–277. [15] H. Burton, M. Fabri, K. Alloway, Cortical areas within the lateral sulcus connected to cutaneous representations in areas 3b and 1: a revised interpretation of the second somatosensory area in macaque monkeys, J. Comp. Neurol. 355 (4) (1995) 539–562. [16] K.O. Bushara, T. Hanakawa, I. Immisch, K. Toma, K. Kansaku, M. Hallett, Neural correlates of cross-modal binding, Nat. Neurosci. 6 (2) (2003) 190–195. [17] G.A. Calvert, R. Campbell, Reading speech from still and moving faces: the neural substrates of visible speech, J. Cogn. Neurosci. 15 (1) (2003) 57–70. [18] S.G. Costafreda, M.J. Brammer, A.S. David, C.H. Fu, Predictors of amygdala activation during the processing of emotional stimuli: a meta-analysis of 385 PET and fMRI studies, Brain Res. Rev. 58 (1) (2008) 57–70. [19] R.W. Cox, AFNI: software for analysis and visualization of functional magnetic resonance neuroimages, Comput. Biomed. Res. 29 (1996) 162–173. [20] J. Decety, K.J. Michalska, Y. Akitsuki, Who caused the pain? An fMRI investigation of empathy and intentionality in children, Neuropsychologia 46 (11) (2008) 2607–2614. [21] S.W. Derbyshire, Exploring the pain “neuromatrix”, Curr. Rev. Pain 4 (6) (2000) 467–477. [22] R.J. Dolan, Emotion cognition and behaviour, Science 298 (2002) 1191–1194. [23] R.J. Dolan, J.S. Morris, B. de Gelder, Crossmodal binding of fear in voice and face, Proc. Natl. Acad. Sci. U.S.A. 98 (17) (2001) 10006–10010. [24] A.D. Engell, J.V. Haxby, Facial expression and gaze-direction in human superior temporal sulcus, Neuropsychologia 45 (14) (2007) 3234–3241. [25] P. Ekman, W.Y. Friesen, Pictures of Facial Affect, Consulting Psychologist Press, Palo Alto, 1976. [26] H. Fischer, C.I. Wright, P.J. Whalen, S.C. McInerney, L.M. Shin, S.L. Rauch, Brain habituation during repeated exposure to fearful and neutral faces: a functional MRI study, Brain Res. Bull. 59 (2003) 387–392. [27] S.D. Forman, J.D. Cohen, M. Fitzgerald, W.F. Eddy, M.A. Mintun, D.C. Noll, Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold, Mang. Reson. Med. 33 (1995) 636–647. [28] K.J. Friston, A.P. Holmes, K.G. Worsley, J.B. Poline, C.D. Frith, R.S. Frackowiak, Statistical parametric maps in functional imaging: a general linear approach, Hum. Brain Mapp. 2 (1995) 189–210. [29] C.G. Gross, J. Sergent, Face recognition, Curr. Opin. Neurobiol. 2 (2) (1992) 156–161. [30] J.V. Haxby, E.A. Hoffman, M.I. Gobbini, The distributed human neural system for face perception, Trends. Cogn. Sci. 4 (6) (2000) 223–233. [31] A. Ishai, L. Pessoa, P.C. Bikle, L.G. Ungerleider, Repetition suppression of faces is modulated by emotion, Proc. Natl. Acad. Sci. U.S.A. 101 (June (26)) (2004) 9827–9832. [32] J.A. Jones, D.E. Callan, Brain activity during audiovisual speech perception: an fMRI study of the McGurk effect, Neuroreport 14 (8) (2003) 1129–1133. [33] N. Kanwisher, J. McDermott, M.M. Chun, The fusiform face area: a module in human extrastriate cortex specialized for face perception, J. Neurosci. 17 (11) (1997) 4302–4311. [34] N. Kanwisher, G. Yovel, The fusiform face area: a cortical region specialized for the perception of faces, Philos. Trans. R. Soc. Lond. B Biol. Sci. 361 (1476) (2006) 2109–2128. [35] W.D. Killgore, D.A. Yurgelun-Todd, Activation of the amygdala and anterior cingulate during nonconscious processing of sad versus happy faces, Neuroimage 21 (4) (2004) 1215–1223. [36] C.D. Kilts, G. Egan, D.A. Gideon, T.D. Ely, J.M. Hoffman, Dissociable neural pathways are involved in the recognition of emotion in static and dynamic facial expressions, Neuroimage 18 (1) (2003) 156–168. [37] S.E. Kim, J.W. Kim, J.J. Kim, B.S. Jeong, E.A. Choi, Y.G. Jeong, J.H. Kim, J. Ku, S.W. Ki, The neural mechanism of imagining facial affective expression, Brain Res. 1145 (2007) 128–137.

264

E. Sabatini et al. / Brain Research Bulletin 79 (2009) 258–264

[38] K.S. LaBar, M.J. Crupain, J.T. Voyvodic, G. McCarthy, Dynamic perception of facial affect and identity in the human brain, Cereb Cortex 13 (10) (2003) 1023–1033. [39] K.S. LaBar, J.C. Gatenby, J.C. Gore, J.E. LeDoux, E.A. Phelps, Human amygdala activation during conditioned fear acquisition and extinction: a Mixed-Trial fMRI Study, Neuron 20 (1998) 937–945. [40] C. Lamm, H.C. Nusbaum, A.N. Meltzoff, J. Decety, What are you feeling? Using functional magnetic resonance imaging to assess the modulation of sensory and affective responses during empathy for pain, PLoS ONE 2 (12) (2007) 1292. [41] L. Mazzola, J. Isnard, F. Mauguière, Somatosensory and pain responses to stimulation of the second somatosensory area (SII) in humans. A comparison with SI and insular responses, Cereb. Cortex 16 (7) (2006) 960–968. [42] M.M. Mesulam, E.J. Mufson, Insula of the old world monkey. I. Architectonics in the insulo-orbito-temporal component of the paralimbic brain, J. Comp. Neurol. 212 (1) (1982) 1–22. [43] J.S. Morris, K.J. Friston, C. Büchel, C.D. Frith, A.W. Young, A.J. Calder, R.J. Dolan, A neuromodulatory role for the human amygdala in processing emotional facial expressions, Brain 121 (1) (1998) 47–57. [44] J.S. Morris, A. Öhman, R.J. Dolan, Conscious and unconscious emotional learning in the human amygdala, Nature 393 (1998) 467–470. [45] I.R. Olson, J.C. Gatenby, J.C. Gore, A comparison of bound and unbound audiovisual information processing in the human cerebral cortex, Brain Res. Cogn. Brain Res. 14 (1) (2002) 129–138. [46] K. Ostrowsky, M. Magnin, P. Ryvlin, J. Isnard, M. Guenot, F. Mauguière, Representation of pain and somatic sensation in the human insula: a study of responses to direct electrical cortical stimulation, Cereb. Cortex 12 (2002) 376–385. [47] D.N. Pandya, E.H. Yeterian, The anatomical substrates of emotional behaviour: the role of the cerebral cortex, in: G. Gainotti (Ed.), Handbook of Neuropsychology, vol. 5, Elsevier, Amsterdam, 2001, pp. 68–80. [48] M.P. Paulus, J.S. Feinstein, G. Castillo, A. Simmons, M.B. Stein, Dose-dependent decrease of activation in bilateral amygdala and insula by lorazepam during emotion processing, Arch. Gen. Psych. 62 (2005) 282–288. [49] W. Penfield, H. Jasper, Epilepsy and the Functional Anatomy of the Human Brain, Little Brown, Boston (MA), 1954. [50] R. Peyron, M. Frot, F. Schneider, L. Garcia-Larrea, P. Mertens, F.G. Barral, M. Sindou, B. Laurent, F. Mauguiére, Role of operculo-insular cortices in human pain processing: Converging evidence from PET, fMRI, dipole modelling, and intracerebral recordings of evoked potentials, NeuroImage 17 (2002) 1336– 1346. [51] E.A. Phelps, Emotion and cognition: Insights from studies of the human amygdala, Annu. Rev. Psychol. 57 (2006) 27–53. [52] L.M. Phillips, L.M. Williams, M. Heining, C.M. Herba, T. Russell, C. Andrew, E.T. Bullmore, M.J. Brammer, S.C. Williams, M. Morgan, A.W. Young, J.A. Gray, Differential neural responses to overt and covert presentations of facial expressions of fear and disgust, NeuroImage 21 (2004) 1484–1496. [53] D.A. Pizzagalli, L.L. Greischar, R.J. Davidson, Spatio-temporal dynamics of brain mechanisms in aversive classical conditioning: high-density event-related potential and brain electrical tomography analyses, Neuropsychologia 41 (2003) 184–194. [54] D.D. Price, Psychological and neural mechanisms of the affective dimension of pain, Science 288 (5472) (2000) 1769–1772. [55] G. Pourtois, B. de Gelder, A. Bol, M. Crommelinck, Perception of facial expressions and voices and of their combination in the human brain, Cortex 41 (1) (2005) 49–59. [56] P. Rainville, Brain mechanisms of pain affect and pain modulation, Curr. Opin. Neurobiol. 12 (2) (2002) 195–204. [57] S.L. Rauch, P.J. Whalen, L.M. Shin, S.C. McInerney, M.L. Macklin, N.B. Lasko, S.P. Orr, R.K. Pitman, Exaggerated amygdala response to masked facial stimuli in posttraumatic stress disorder: a functional MRI study, Biol. Psychiatry 47 (9) (2000) 769–776. [58] D.L. Robins, E. Hunyadi, R.T. Schultz, Superior temporal activation in response to dynamic audio-visual emotional cues, Brain Cogn. (2008).

[59] E.T. Rolls, J. O’Doherty, M.L. Kringelbach, S. Francis, R. Bowtell, R. McGlone, Representation of pleasant and painful touch in the human orbitofrontal and cingulate cortices, Cereb. Cortex 13 (2003) 308–317. [60] E.T. Rolls, M.J. Tovee, S. Panzeri, The neurophysiology of backward visual masking: information analysis, J. Cogn. Neurosci. 11 (3) (1999) 300–311. [61] D.N. Saito, K. Yoshimura, T. Kochiyama, T. Okada, M. Honda, N. Sadato, Crossmodal binding and activated attentional networks during audio-visual speech integration: a functional MRI study, Cereb. Cortex 15 (11) (2005) 1750–1760. [62] N. Sawamoto, M. Honda, T. Okada, T. Hanakawa, M. Kanda, H. Fukuyama, J. Konishi, H. Shibasaki, Expectation of pain enhanced responses to non-painful somatosensory stimulation in the anterior cingulate cortex and parietal operculum/posterior insula: An event related functional magnetic resonance imaging study, J. Neurosci. 20 (19) (2000) 7438–7445. [63] C.E. Schwartz, C.I. Wright, L.M. Shin, J. Kagan, P.J. Whalen, K.G. McMullin, S.L. Rauch, Differential amygdalar responses to novel versus familiar neutral faces: a functional MRI probe developed for studying inhibited temperament, Biol. Psychiatry 53 (2003) 854–862. [64] K. Sergerie, C. Chochol, J.L. Armony, The role of the amygdala in emotional processing: a quantitative meta-analysis of functional neuroimaging studies, Neurosci. Biobehav. Rev. 32 (4) (2008) 811–830. [65] K. Sergerie, M. Lepage, J.L. Armony, A face to remember: emotional expression modulates prefrontal activity during memory formation, Neuroimage 24 (2) (2005) 580–585. [66] P. Shaw, J. Bramham, E.J. Lawrence, R. Morris, S. Baron-Cohen, A.S. David, Differential effects of lesions of the amygdala and prefrontal cortex on recognizing facial expressions of complex emotions, J. Cogn. Neurosci. 17 (9) (2005) 1410–1419. [67] A. Simmons, C.M. Scott, B.S. Murray, P.P. Martin, Anticipation of emotionally aversive visual stimuli activates right insula, NeuroReport 15 (14) (2004) 2261–2265. [68] A. Simmons, I. Strigo, S.C. Matthems, M.P. Paulus, M.B. Stein, Anticipation of aversive visual stimuli is associated with increased insula activation in anxietyprone subjects, Biol. Psychiatry 60 (2006) 402–409. [69] L. Stefanacci, D.G. Amaral, Some observations on cortical inputs to the macaque monkey amygdala: an anterograde tracing study, J. Comp. Neurol. 451 (4) (2002) 301–323. [70] M.M. Strauss, N. Makris, I. Aharon, M.G. Vangel, J. Goodman, D.N. Kennedy, G.P. Gasic, H.C. Breiter, FMRI of sensitization to angry faces, NeuroImage 26 (2005) 389–413. [71] Y. Sugase, S. Yamane, S. Ueno, K. Kawano, Global and fine information coded by single neurons in the temporal visual cortex, Nature 400 (6747) (1999) 869–873. [72] J. Talairach, P. Tournoux, Coplanar Stereotactic Atlas of the Human Brain, Thieme, New York, 1988. [73] R.D. Treede, A.V. Apkarian, B. Bromm, J.D. Greenspan, F.A. Lenz, Cortical representation of pain: functional characterization of nociceptive areas near the lateral sulcus, Pain 87 (2000) 113–119. [74] P. Vuilleumier, J.L. Armony, J. Driver, R.J. Dolan, Distinct spatial frequency sensitivities for processing faces and emotional expressions, Nat. Neurosci. 6 (6) (2003) 624–631. [75] P.J. Whalen, S.L. Rauch, N.L. Etcoff, S.C. McInerney, M.B. Lee, M.A. Jenike, Masked presentations of emotional facial expressions modulate amygdala activity without explicit knowledge, J. Neurosci. 18 (1) (1998) 411–418. [76] M.A. Williams, A.P. Morris, F. McGlone, D.F. Abbott, J.B. Mattingley, Amygdala responses to fearful and happy facial expressions under conditions of binocular suppression, J. Neurosci. 24 (12) (2004) 2898–2904. [77] M.W. Woolrich, B.D. Ripley, M. Brady, S.M. Smith, Temporal autocorrelation in univariate linear modeling of FMRI data, NeuroImage 14 (2001) 1370–1386. [78] C.I. Wright, H. Fischer, P.J. Whalen, S.C. McInerney, L.M. Shin, S.L. Rauch, Differential prefrontal cortex and amygdala habituation to repeatedly presented emotional stimuli, Brain Imaging 12 (2) (2001) 379–383. [79] H. Yamasaki, K.S. LaBar, G. McCarthy, Dissociable prefrontal brain systems for attention and emotion, Proc. Natl. Acad. Sci. U.S.A. 99 (17) (2002) 11447–11451.