Social Perception in First-Episode Schizophrenia
Social Perception in First-Episode Schizophrenia
The current study aimed at investigating the neural mechanism underlying pre-reflective social perception in FES. In a previous neuroimaging study, it was shown that a neural network involved in first-person tactile experiences may underpin the sharing of others' bodily feelings on the one hand, and, on the other hand, self-related processes concerning multisensory integration and self-other distinction during social perception (Ebisch et al., 2011). Here, the same experimental paradigm was used to answer the question whether in FES, compared with HC participants, functional abnormalities during social perception of other individuals' affective tactile stimulation specifically concern the intersubjective domain or primarily have their roots in disturbances of self-experience.
With respect to brain regions showing overlapping activation for touch observation and experience, a significant difference between the HC and FES group was found for all the touch observation conditions in RH vPMC. BOLD response was significantly weaker in the FES group in all cases.
The coordinates of the vPMC cluster in the present study (50,7,32) are strikingly similar to the coordinates reported by other studies: 52,10,30 (Bremmer et al., 2001); 53,11,12 (Makin et al., 2007). This suggests that this region reflects the putative human homologue of monkey premotor area F4 (Bremmer et al., 2001; Buccino et al., 2001; Galati et al., 2001; Makin et al., 2007; Serino et al., 2011). Indeed, consistent with the multisensory properties of F4, in the present study this region responded to visual as well as somatosensory events in the HC group (Rizzolatti et al., 1981a, 1981b; Bremmer et al., 2001). It has been proposed that this region could be involved in the integration of multisensory information from vision, touch and proprioception onto the motor representations of different body parts (Fogassi et al., 1996a, 1996b; Graziano, 2001; Rizzolatti et al., 2002).
In the present study, activation in vPMC for the experience and observation of touch likely reflects a monitoring/integration of multisensory information, including proprioceptive, visual and tactile self-experiences, related to one's body in different situations. Adequate self-monitoring of multisensory information is crucial for the experience of a coherent sense of self and other (Parnas et al., 2002).
The lack of activation in RH vPMC in the FES group would suggest a disruption of an integrated multisensory representation of the bodily self. Interestingly, activation in this region in the FES group correlated negatively with BS severity: activation in vPMC decreased with augmented symptom severity. This relationship was consistently found between the social touch observation conditions and BS severity, but not with the degree of positive or negative symptoms. BS represent subjective experiential disturbances in the domains of cognition, perception, bodily experience, action and emotion (Huber, 1983; Klosterkotter, 1992). Therefore, the present results support a close link between impaired multisensory representations and a disrupted sense of a coherent self in everyday life (Parnas et al., 2002). A breakdown of self-monitoring has been suggested before in schizophrenia (Frith, 1987; Blakemore et al., 2000; Vinogradov et al., 2008) and vPMC could be a key structure underlying this link.
Such alteration could reasonably lead to the blurring of self boundaries and confusion in the inter-relationship with others (Sass and Parnas, 2003). Indeed, it was reported that schizophrenic patients with high self-monitoring skills had better social skills (Ihnen et al., 1998). This interpretation offers an intriguing hypothesis for further investigations regarding the multisensory properties of vPMC in relationship with self-experience disturbances. Moreover, given that BS remain stable during the entire disease progression, including the prodromal phase of schizophrenia, the study of their relationship with cortical processes and social deficits would be useful also from a clinical point of view, especially in the light of preventive approaches and an early diagnosis of schizophrenia (Addington et al., 2008a; Schultze-Lutter, 2009).
A second brain region of interest where differences were identified between the HC and FES group was pIC. In the HC group, BOLD response in pIC was positively modulated, compared with baseline, during first-person experience of touch, but negatively modulated (deactivated), compared with baseline, during the observation of touch in another individual. In contrast, differential activation for first-person touch experiences and the observation of touch in another individual was absent in the FES group; no deactivation was found in pIC during the observation of touch, though normal activation patterns were found in pIC for first-person touch experiences. Deactivation in pIC in the HC group was specific for the observation of social affective touch (Ebisch et al., 2011). Accordingly, significant differences in LH pIC between the HC and FES group were found specifically for the observation of affective social touch, either with a positive (caress) or with a negative (hit) valence in LH pIC. Significant differences in RH pIC between the HC and FES group were found specifically for the observation of a hit.
pIC is considered a central brain region for interoception (Craig, 2002). Thalamo-cortical pathways may provide a direct representation of homeostatic afferent information to pIC that interacts with limbic, somatosensory and motor regions, subserving the awareness of bodily feelings (Augustine, 1996; Saper, 2002; Critchley, 2005; Craig, 2009). Regarding the cutaneous senses, pIC could constitute the primary cortical locus of an interoceptive system regulating threatening (Craig, 2002) or comforting (Olausson et al., 2002; Loken et al., 2009) information from the skin.
In accordance with the idea of pIC as a central cortical node in a system constituting a neural representation of 'the material me' (Craig, 2002), empirical evidence also suggests that pIC is involved in self-awareness. A PET study in healthy adults showed a positive relationship between the subjective experience of the rubber hand illusion (RHI: a condition in which an observed rubber hand is subjectively experienced as if it actually were one's own hand; Botvinick and Cohen, 1998) and neural activation in pIC (Tsakiris et al., 2007). pIC has also been related to body part awareness in anosognosia patients with hemiplegia/hemiparesis (Karnath et al., 2005), and to the sense of agency (Farrer et al., 2003).
Finally, fMRI evidence suggests that pIC is involved in social perception. Opposite activation patterns for the experience and observation of touch suggest that pIC differentiates between self and other conditions during social perception when affective experiences are implicated (Ebisch et al., 2011). An inhibitory mechanism at the level of pIC during social perception may facilitate the observer to distinguish at the phenomenal level to whom the observed tactile experience belongs.
Hence, the reduced BOLD suppression in pIC reported here in FES could indicate an impaired differentiation between self and other conditions during social perception. We propose that the absence of deactivation in pIC during the observation of touch in another individual in FES reflects a deficit in the pre-reflective suppression of self-oriented affective arousal, which likely normally contributes to the differentiation between self and other conditions.
Indeed, previous studies in schizophrenia revealed confusion in the attribution of events in the external world to their origin (Bentall et al., 1991; Blakemore et al., 2000; Franck et al., 2001; Vinogradov et al., 2008; Voss et al., 2010). A relationship has been demonstrated between aberrant pIC activation and an impaired sense of agency in schizophrenia (Farrer et al., 2004). Furthermore, the RHI, associated with pIC functioning (Tsakiris et al., 2007), has been found enhanced in schizophrenic pathology (Peled et al., 2000; Morgan et al., 2011).
It remains unclear what may be the dysfunctional neural mechanism underlying the observed reduced suppression of BOLD response in pIC during affective touch observation in FES. Possibly, it could be based on altered connections involved in top–down control processes. Future (functional) connectivity studies are urged to elucidate this issue.
Some differences could be observed between HC and FES in brain regions not involved in first-person tactile experiences, too. Stronger activation in the HC group during the observation of others' tactile experiences, compared with the FES group, was found in higher order visual areas (ventral fusiform and OT cortex; Downing et al., 2006; Peelen and Downing, 2007) and regions underlying action–perception coupling (MCC/SMA; Vogt, 2005; Dayan et al., 2007). These results may suggest a reduced involvement of brain regions related to the visual and multimodal processing of bodies, animate objects and motion in FES. In contrast, increased activation in the FES group, compared with the HC group, in occipital cortex during the observation of the touch videos, may indicate a stronger involvement of low-level visual cortices in FES patients.
Finally, weaker deactivation was observed in vMPFC in the FES group, compared with the HC group, during touch observation. vMPFC has been associated with the coding of the self-relatedness of stimuli (Northoff, 2007), self-regulation (Heatherton, 2011) and dysfunctional self-processing in schizophrenia (van der Meer et al., 2010). However, since vMPFC often is characterized by task-negative activation patterns (Gusnard and Raichle, 2001), possibly indicating increased self-related processing during periods of rest, it remains to be established whether the observed decreased deactivation in vMPFC in FES may be caused by a reduced activity during periods of rest (fixation cross) or by a failure to suppress vMPFC activity during touch observation (Northoff, 2007). Whereas the former can be interpreted as dysfunctional self-related processes, the latter may suggest that FES patients fail to disengage self-related processes during social perception or to distinguish between self and other conditions.
In conclusion, the present study supports the hypothesis that social perception at a pre-reflective level in FES is primarily characterized by altered neural activation patterns underlying disturbances of self-experience and self-related processing, concerning both self-other distinction and a multimodal representation of the bodily self. No abnormalities were found in FES with respect to shared activation between experienced and observed touch in somatosensory cortices. Moreover, significant involvement of anterior insula for the hit and caress videos in the HC as well as in the FES group (see Results in Supplementary Data) could be indicative for distinguishing at the neural level the affective visual touch stimuli from the neutral stimuli in both groups (Menon and Uddin, 2010).
Nevertheless, self-reported social skills scores regarding an intuitive understanding of social situations were significantly reduced in the FES group and negatively correlated with BS, reflecting disturbed subjective self-experiences. BS also negatively correlated with neural activation in RH vPMC in the FES group. Therefore, self-experience disturbances in schizophrenia may extend to the social domain as well (Parnas et al., 2002; Gallese, 2003b; Sass and Parnas, 2003). For example, an incoherent sense of one's self-experiences accompanied by the loss of grip on the world may lead to a fading of the very distinction between self and other, and an incapacity to intuitively grasp the meaning of social situations.
Differently from previous studies (Kohler et al., 2000; Phillips et al., 2003; Marjoram et al., 2005; Shean et al., 2005), no relationship was detected between brain activation during social perception, social abilities and positive and negative symptoms. A possible explanation for this is that the included patients had a very recent illness onset and relatively low PANSS scores. Further studies will be needed to investigate patients with more pronounced positive and negative symptoms as well as chronic samples for a better understanding of the relationship between psychotic symptoms, altered social perception, and its progress over time.
There is growing evidence that social cognition deficits are related to social dysfunction in schizophrenia. The present results shed new light on the cortical basis of how self-experience disturbances in schizophrenia pervade the social domain at a relatively early stage. However, the present findings also depict dysfunctional social perception in schizophrenia as a complex impairment at multiple neural processing levels, rather than being confined to brain regions involved with first-person bodily (tactile) experiences.
Discussion
The current study aimed at investigating the neural mechanism underlying pre-reflective social perception in FES. In a previous neuroimaging study, it was shown that a neural network involved in first-person tactile experiences may underpin the sharing of others' bodily feelings on the one hand, and, on the other hand, self-related processes concerning multisensory integration and self-other distinction during social perception (Ebisch et al., 2011). Here, the same experimental paradigm was used to answer the question whether in FES, compared with HC participants, functional abnormalities during social perception of other individuals' affective tactile stimulation specifically concern the intersubjective domain or primarily have their roots in disturbances of self-experience.
vPMC and a Coherent Sense of Self
With respect to brain regions showing overlapping activation for touch observation and experience, a significant difference between the HC and FES group was found for all the touch observation conditions in RH vPMC. BOLD response was significantly weaker in the FES group in all cases.
The coordinates of the vPMC cluster in the present study (50,7,32) are strikingly similar to the coordinates reported by other studies: 52,10,30 (Bremmer et al., 2001); 53,11,12 (Makin et al., 2007). This suggests that this region reflects the putative human homologue of monkey premotor area F4 (Bremmer et al., 2001; Buccino et al., 2001; Galati et al., 2001; Makin et al., 2007; Serino et al., 2011). Indeed, consistent with the multisensory properties of F4, in the present study this region responded to visual as well as somatosensory events in the HC group (Rizzolatti et al., 1981a, 1981b; Bremmer et al., 2001). It has been proposed that this region could be involved in the integration of multisensory information from vision, touch and proprioception onto the motor representations of different body parts (Fogassi et al., 1996a, 1996b; Graziano, 2001; Rizzolatti et al., 2002).
In the present study, activation in vPMC for the experience and observation of touch likely reflects a monitoring/integration of multisensory information, including proprioceptive, visual and tactile self-experiences, related to one's body in different situations. Adequate self-monitoring of multisensory information is crucial for the experience of a coherent sense of self and other (Parnas et al., 2002).
The lack of activation in RH vPMC in the FES group would suggest a disruption of an integrated multisensory representation of the bodily self. Interestingly, activation in this region in the FES group correlated negatively with BS severity: activation in vPMC decreased with augmented symptom severity. This relationship was consistently found between the social touch observation conditions and BS severity, but not with the degree of positive or negative symptoms. BS represent subjective experiential disturbances in the domains of cognition, perception, bodily experience, action and emotion (Huber, 1983; Klosterkotter, 1992). Therefore, the present results support a close link between impaired multisensory representations and a disrupted sense of a coherent self in everyday life (Parnas et al., 2002). A breakdown of self-monitoring has been suggested before in schizophrenia (Frith, 1987; Blakemore et al., 2000; Vinogradov et al., 2008) and vPMC could be a key structure underlying this link.
Such alteration could reasonably lead to the blurring of self boundaries and confusion in the inter-relationship with others (Sass and Parnas, 2003). Indeed, it was reported that schizophrenic patients with high self-monitoring skills had better social skills (Ihnen et al., 1998). This interpretation offers an intriguing hypothesis for further investigations regarding the multisensory properties of vPMC in relationship with self-experience disturbances. Moreover, given that BS remain stable during the entire disease progression, including the prodromal phase of schizophrenia, the study of their relationship with cortical processes and social deficits would be useful also from a clinical point of view, especially in the light of preventive approaches and an early diagnosis of schizophrenia (Addington et al., 2008a; Schultze-Lutter, 2009).
pIC and Self-Other Distinction
A second brain region of interest where differences were identified between the HC and FES group was pIC. In the HC group, BOLD response in pIC was positively modulated, compared with baseline, during first-person experience of touch, but negatively modulated (deactivated), compared with baseline, during the observation of touch in another individual. In contrast, differential activation for first-person touch experiences and the observation of touch in another individual was absent in the FES group; no deactivation was found in pIC during the observation of touch, though normal activation patterns were found in pIC for first-person touch experiences. Deactivation in pIC in the HC group was specific for the observation of social affective touch (Ebisch et al., 2011). Accordingly, significant differences in LH pIC between the HC and FES group were found specifically for the observation of affective social touch, either with a positive (caress) or with a negative (hit) valence in LH pIC. Significant differences in RH pIC between the HC and FES group were found specifically for the observation of a hit.
pIC is considered a central brain region for interoception (Craig, 2002). Thalamo-cortical pathways may provide a direct representation of homeostatic afferent information to pIC that interacts with limbic, somatosensory and motor regions, subserving the awareness of bodily feelings (Augustine, 1996; Saper, 2002; Critchley, 2005; Craig, 2009). Regarding the cutaneous senses, pIC could constitute the primary cortical locus of an interoceptive system regulating threatening (Craig, 2002) or comforting (Olausson et al., 2002; Loken et al., 2009) information from the skin.
In accordance with the idea of pIC as a central cortical node in a system constituting a neural representation of 'the material me' (Craig, 2002), empirical evidence also suggests that pIC is involved in self-awareness. A PET study in healthy adults showed a positive relationship between the subjective experience of the rubber hand illusion (RHI: a condition in which an observed rubber hand is subjectively experienced as if it actually were one's own hand; Botvinick and Cohen, 1998) and neural activation in pIC (Tsakiris et al., 2007). pIC has also been related to body part awareness in anosognosia patients with hemiplegia/hemiparesis (Karnath et al., 2005), and to the sense of agency (Farrer et al., 2003).
Finally, fMRI evidence suggests that pIC is involved in social perception. Opposite activation patterns for the experience and observation of touch suggest that pIC differentiates between self and other conditions during social perception when affective experiences are implicated (Ebisch et al., 2011). An inhibitory mechanism at the level of pIC during social perception may facilitate the observer to distinguish at the phenomenal level to whom the observed tactile experience belongs.
Hence, the reduced BOLD suppression in pIC reported here in FES could indicate an impaired differentiation between self and other conditions during social perception. We propose that the absence of deactivation in pIC during the observation of touch in another individual in FES reflects a deficit in the pre-reflective suppression of self-oriented affective arousal, which likely normally contributes to the differentiation between self and other conditions.
Indeed, previous studies in schizophrenia revealed confusion in the attribution of events in the external world to their origin (Bentall et al., 1991; Blakemore et al., 2000; Franck et al., 2001; Vinogradov et al., 2008; Voss et al., 2010). A relationship has been demonstrated between aberrant pIC activation and an impaired sense of agency in schizophrenia (Farrer et al., 2004). Furthermore, the RHI, associated with pIC functioning (Tsakiris et al., 2007), has been found enhanced in schizophrenic pathology (Peled et al., 2000; Morgan et al., 2011).
It remains unclear what may be the dysfunctional neural mechanism underlying the observed reduced suppression of BOLD response in pIC during affective touch observation in FES. Possibly, it could be based on altered connections involved in top–down control processes. Future (functional) connectivity studies are urged to elucidate this issue.
Between-Group Differences in Other Brain Regions
Some differences could be observed between HC and FES in brain regions not involved in first-person tactile experiences, too. Stronger activation in the HC group during the observation of others' tactile experiences, compared with the FES group, was found in higher order visual areas (ventral fusiform and OT cortex; Downing et al., 2006; Peelen and Downing, 2007) and regions underlying action–perception coupling (MCC/SMA; Vogt, 2005; Dayan et al., 2007). These results may suggest a reduced involvement of brain regions related to the visual and multimodal processing of bodies, animate objects and motion in FES. In contrast, increased activation in the FES group, compared with the HC group, in occipital cortex during the observation of the touch videos, may indicate a stronger involvement of low-level visual cortices in FES patients.
Finally, weaker deactivation was observed in vMPFC in the FES group, compared with the HC group, during touch observation. vMPFC has been associated with the coding of the self-relatedness of stimuli (Northoff, 2007), self-regulation (Heatherton, 2011) and dysfunctional self-processing in schizophrenia (van der Meer et al., 2010). However, since vMPFC often is characterized by task-negative activation patterns (Gusnard and Raichle, 2001), possibly indicating increased self-related processing during periods of rest, it remains to be established whether the observed decreased deactivation in vMPFC in FES may be caused by a reduced activity during periods of rest (fixation cross) or by a failure to suppress vMPFC activity during touch observation (Northoff, 2007). Whereas the former can be interpreted as dysfunctional self-related processes, the latter may suggest that FES patients fail to disengage self-related processes during social perception or to distinguish between self and other conditions.
General Discussion and Conclusions
In conclusion, the present study supports the hypothesis that social perception at a pre-reflective level in FES is primarily characterized by altered neural activation patterns underlying disturbances of self-experience and self-related processing, concerning both self-other distinction and a multimodal representation of the bodily self. No abnormalities were found in FES with respect to shared activation between experienced and observed touch in somatosensory cortices. Moreover, significant involvement of anterior insula for the hit and caress videos in the HC as well as in the FES group (see Results in Supplementary Data) could be indicative for distinguishing at the neural level the affective visual touch stimuli from the neutral stimuli in both groups (Menon and Uddin, 2010).
Nevertheless, self-reported social skills scores regarding an intuitive understanding of social situations were significantly reduced in the FES group and negatively correlated with BS, reflecting disturbed subjective self-experiences. BS also negatively correlated with neural activation in RH vPMC in the FES group. Therefore, self-experience disturbances in schizophrenia may extend to the social domain as well (Parnas et al., 2002; Gallese, 2003b; Sass and Parnas, 2003). For example, an incoherent sense of one's self-experiences accompanied by the loss of grip on the world may lead to a fading of the very distinction between self and other, and an incapacity to intuitively grasp the meaning of social situations.
Differently from previous studies (Kohler et al., 2000; Phillips et al., 2003; Marjoram et al., 2005; Shean et al., 2005), no relationship was detected between brain activation during social perception, social abilities and positive and negative symptoms. A possible explanation for this is that the included patients had a very recent illness onset and relatively low PANSS scores. Further studies will be needed to investigate patients with more pronounced positive and negative symptoms as well as chronic samples for a better understanding of the relationship between psychotic symptoms, altered social perception, and its progress over time.
There is growing evidence that social cognition deficits are related to social dysfunction in schizophrenia. The present results shed new light on the cortical basis of how self-experience disturbances in schizophrenia pervade the social domain at a relatively early stage. However, the present findings also depict dysfunctional social perception in schizophrenia as a complex impairment at multiple neural processing levels, rather than being confined to brain regions involved with first-person bodily (tactile) experiences.