Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Psychiatry Research: Neuroimaging journal homepage: www.elsevier.com/locate/psychresns

Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study Jiacheng Liu a,b,n, Silvia Corbera a,c, Bruce Edward Wexler a a

Department of Psychiatry, Yale University School of Medicine, Connecticut Mental Health Center, 34 Park Street, CMHC 527, New Haven, CT 06519, USA Department of Radiology, Zhongda Hospital, Southeast University, 87 Dingjiaqiao, Nanjing, Jiangsu 210009, China c Olin Neuropsychiatry Research Center, Institute of Living, 400 Washington Street, Hartford, CT 06114, USA b

art ic l e i nf o

a b s t r a c t

Article history: Received 8 September 2013 Received in revised form 10 March 2014 Accepted 1 April 2014

Impairments in self-awareness contribute to disability in schizophrenia. Studies have revealed activation abnormalities in schizophrenia in cortical midline structures associated with self-reference. We used functional magnetic resonance imaging to compare activation throughout the brain in people with schizophrenia and healthy controls (Kelly et al., 2002) while they indicated whether trait adjectives described attributes of themselves, their mother or a former president of the United States. Blood oxygenation level dependent signal in each condition was compared to resting fixation. Patients were less likely and slower to endorse positive self-attributes, and more likely and quicker to endorse negative self-attributes than controls. Activation abnormalities reported previously in cortical midline structures were again noted. In addition, patients showed greater signal increases in frontal, temporal gyri and insula, and smaller signal decreases in posterior regions than healthy controls when thinking about themselves. Group differences were less evident when subjects were thinking about their mothers and tended to go in the opposite direction when thinking about a president. Many of the areas showing abnormality have been shown in other studies to differ between patients and controls in structure and with other activation paradigms. We suggest that general neuropathology in schizophrenia alters the neural system configurations associated with self-representation. & 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: Personal awareness Social cognition Mental disorder

1. Introduction Internal representation of self and others are essential parts of psychic structure and interpersonal function. Several studies report impairment in ability to recognize “self” in schizophrenia (Fleisher et al., 2009; Lysaker et al., 2005; Parnas et al., 2005). Social dysfunction is common in schizophrenia and close relationships with others are often lacking or problematic. Recent neuroimaging research in healthy individuals used a self-referential task to identify neural substrates related to self-representation. Stronger medial prefrontal cortex (MPFC) activation was linked to selfrelative to other-judgments (Craik et al., 1999; Kelley et al., 2002; Lieberman et al., 2004; Zhang et al., 2006), indicating that MPFC is engaged in representation of self-knowledge such as one's own personality traits. This area and other cortical midline structures

n Corresponding author at: Department of Psychiatry, Yale University School of Medicine, 34 Park St. CMHC 527, New Haven, CT 06519, USA. Tel.: þ 1 203 974 7339; fax: þ1 203 974 7881. E-mail addresses: [email protected] (J. Liu), [email protected] (S. Corbera), [email protected] (B. Edward Wexler).

(CMS) are also active during other aspects of social and emotional processing (Ochsner, 2008). In addition, other areas have also been shown to be associated with self- or others-awareness in healthy subjects, such as the inferior frontal gyrus (IFG), insula and temporal pole (TP) (Shad et al., 2012; van der Meer et al., 2010). Brain pathology has been noted in schizophrenia in many of these regions. Previous fMRI studies in schizophrenia using a character judgment paradigm to generate self-referential thinking found that, compared to healthy controls, patients had decreased ventral medial prefrontal cortex activation, increased middle- and posterior-dorsal cingulate cortex (m/pCC) activation (Holt et al., 2011; Shad et al., 2012), and increased temporal gyrus activation (Murphy et al., 2010). Other brain regions active during selfreferential tasks in healthy controls and known to show pathology in schizophrenia (Harrison, 1999; Harrison and Weinberger, 2005; Lewis, 2000; Lewis and Sweet, 2009; Powers, 1999; Selemon and Goldman-Rakic, 1999) received little attention in the three studies. Additionally, the studies in patients focused on patient-control differences in contrasts of the social versus a nonsocial cognitive task. Social tasks share cognitive processes with nonsocial tasks (Green and Horan, 2010), and the non-specific regions are

http://dx.doi.org/10.1016/j.pscychresns.2014.04.003 0925-4927/& 2014 Elsevier Ireland Ltd. All rights reserved.

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

necessary components of the activations associated with the social functions. In the present study we used functional magnetic resonance imaging (fMRI) to determine whether patient abnormalities when making character judgments about themselves extended beyond the CMS and were greater than when they made character judgments about others. We compared activations to a simple fixation baseline to provide a full view of activation differences.

2. Methods and subjects 2.1. Subjects Seventeen patients meeting DSM-IV criteria for schizophrenia or schizoaffective disorder were enrolled; two patients were excluded because of excessive movement during fMRI, (leaving male: 6; age: 50.0 7 7.4 years; range of age: 39–65 years; illness onset: 26.17 8.6; illness duration: 24.77 10.1). Fifteen healthy subjects (male: 8; age: 40.5 7 8.8; range of age: 31–58 years) participated. Patients were stable outpatients (no hospitalizations, homelessness or substance abuse in the last six months) and taking medication (chlorpromazine equivalent mean doses (Andreasen et al., 2010): 5337 576 mg/day). Symptoms were assessed with the Positive and Negative Syndrome Scale (PANSS) using 5 component scores: positive, negative, cognitive, hostility, and emotional discomfort (Bell et al., 1994). Mean total PANSS score was 60.8 7 15.3 (range 49–92). Positive symptom, negative symptom, and emotional discomfort scores were 15.7 7 7.0, 14.5 7 3.9 and 9.6 7 4.3, respectively. All subjects were without a history of neurological disease, brain injury, or developmental disability, and were right-handed (Wexler and Halwes, 1983) native English speakers with normal or corrected to normal vision. All gave written informed consent following procedures approved by the Yale Institutional Review Board and were drug free confirmed by urine toxicology before imaging. Patients and controls did not differ significantly in ethnicity or parental education (patients ¼ 12.4 7 5.4 yr, controls¼13.1 73.0 yr, P¼ 0.67 for paternal education; patients ¼ 12.5 7 3.6 yr, controls¼12.7 7 2.1 yr, P¼ 0.88 for maternal education), but patients had less education (patients ¼ 13.5 72.3 yr, controls¼ 16.0 7 2.1 yr, P¼ 0.004) and were significantly younger. 2.2. Paradigm for fMRI FMRI used a block design with the three referential tasks (Zhu et al., 2007). There were two blocks for each task: imagine “self”, “mother”, “president” and capitaljudgment. Prior to beginning, subjects selected a previous president with whom they were familiar. In each referential block ten adjectives were displayed one-by-one, and subjects indicated by button push whether each word described the person they were imagining, and in the capital-judgment block (half of the trials with capital and half of the trials with lower case), subjects indicated whether each adjective was in capital or lower case. Each adjective displayed for 3.5 s following a fixation cross at screen center for 1.0 s. Forty positive and 40 negative adjectives from normalized personality trait adjectives were used (Anderson, 1968). Five positive and five negative valence adjectives were chosen randomly for presentation in each block, no words were repeated and word length was balanced across blocks. A baseline block appeared before each task block, with a 3.5 second fixation cross alternating with a 1 s gray screen repeated four times. Subjects were trained on the task prior to imaging using adjectives different from those during scanning.

2.3. Image acquisitions MRI data were collected in the Yale MRI Research Center using a Siemens 3T scanner with twelve channel head coil. Axial slices were defined relative to the Anterior–Posterior commissure line. Blood oxygenation level dependent (BOLD) images were acquired during task performance. A T2n-weighted gradient echoplanar imaging sequence was used: 1500 ms repetition time, 30 ms echo time, 601 flip angle, 64  64 matrix, field of view of 220 mm  220 mm and 24 slices of 4.5 mm without gap. Prior to BOLD imaging, two-dimension axial T1-weighted spin-echo images with the same slices as BOLD images were acquired for anatomical reference. High-resolution saggittal T1-weighted spin-echo, images were acquired for coregistration with Montreal Neurological Institute (MNI) brain template.

2.4. Behavior data analysis R (http://www.r-project.org/) was employed to analyze behavioral data. ANOVA and post-hoc analysis were run with groups (patients vs. controls), tasks (self, mother, president), and valence (positive vs. negative attribute) as factors, response rates and times as dependent measures.

2.5. Image data analysis Following the preprocessing procedure at the Yale Magnetic Resonance Research Center, BOLD data were converted into nifty, timing and motion corrected using SPM algorithm (http://www.fil.ion.ucl.ac.uk/spm), registered to the MNI brain template, re-sampled into 3 mm  3 mm  3 mm in MNI brain space, and spatially filtered using a Gaussian filter with a 2 pixel full-width at half-maximum using the Bioimagesuite algorithm (http://bioimagesuite.yale.edu). All statistical analyses of image data were conducted using an in-house code written in R drawing on functions related to image analysis and statistical inference in neuroimaging (reviewed by Tabelow et al. (2011)). The quality attribution task was designed to produce extended internal representations of the self, mother and significant public figure in separate self, mother and president blocks. Possible signal drift over time during each block and the whole session, and other noise in BOLD data, were removed voxel-by-voxel for each subject individually by extracting signal over time in a large white matter area (bilateral half-ovary areas). Average percent change in BOLD was then defined by subtracting average signal in baseline blocks from average signal during each task and dividing the difference by the values during baseline blocks (Liu et al., 2012). The voxel-wise t test of “self þmotherþ president” versus capital-judgment was performed to identify regions related to the referential tasks in general. For each group, maps of brain activation changes during each referential task were created using t tests. A 2  3 ANOVA (group  referential task) with age as a covariate was run voxel-by-voxel. Post hoc tests were performed to identify group differences in each referential condition. In order to evaluate our question of primary apriori interest, interaction specific to the self condition, the contrast of patients(self–others)– controls(self–others) was tested as well. AlphaSim (Forman et al., 1995) was run to simulate the random field with noise at each given p level and give the combination of p-value and cluster-size to correct p-values for multiple comparisons in the whole brain to p-corrected r0.05 twotailed.

3. Results 3.1. Behavior results All subjects were more likely to endorse positive than negative attributes of self and others (Table 1, F(1,168) ¼534.5 Po0.00001). The main effect of group was not significant (F(1,168) ¼ 0.087, P¼0.77), but the two-way interaction of group by adjective valence (F(1,168) ¼ 7.71, P¼0.006) was significant with controls more likely to endorse positive attributes as applicable to self and others than were patients (P-corrected¼0.00024) while the groups did not differ in frequency of endorsing negative attributes. The main effect of referential task closely approached significance (F(2,168) ¼3.0, P¼0.053) and was qualified by a three-way interaction (F(2,168) ¼3.48, P¼0.033). Posthoc analysis revealed patients were less likely to endorse positive and deny negative attributes when referring to themselves than when referring to mother or president than were controls.: P-corrected¼0.022 for (patients–controls)  (self–mother)  (endorse positive–endorse negative) and p-corrected¼0.025 for (patients–controls)  (self–president)  (endorse positive–endorse negative). In regard to reaction times, endorsing positive attributes was faster than denying negative, denying positive and endorsing negative (P-corrected¼ 0.00072,o0.0001 and o0.0001, respectively). Denying positive and endorsing negative was slower than denying negative (P-corrected¼ 0.0052 and 0.0096, respectively).

Table 1 Response rates of endorsing. Tasks

Patients Positive adjectives (100%)

Self 0.7070.25 Mother 0.83 7 0.28 President 0.85 7 0.27

Controls Negative adjectives (100%)

Positive adjectives (100%)

Negative adjectives (100%)

0.26 70.17 0.317 0.22 0.217 0.20

0.86 7 0.24 0.81 7 0.24 0.90 7 0.18

0.09 7 0.09 0.25 7 0.21 0.20 7 0.25

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

The interactions of group by task (F(2, 1559) ¼4.08, P ¼0.017) and task by response type (F(6, 1559) ¼3.60, P ¼0.0015) were significant; the interaction of group by response type approached significance (F(3, 1559) ¼ 2.55, P ¼0.054); and the three-factor interaction (F(6, 1559) ¼2.36, P ¼0.028) was significant. Controls were faster than patients (P-corrected ¼0.028) in endorsing positive attributes of self but not in denying positive. Only controls were faster endorsing positive attributes than denying negative attributes of themselves (P-corrected¼ 0.0002). 3.2. fMRI results Activation maps in healthy controls show robust signal increases in MPFC (mainly left), left lateral IFG, and insula, and signal decreases in bilateral cuneus, in all three referential conditions (Fig. 1, rows (A)–(C), P o0.005, cluster4 38, whole brain P-correctedo0.05). These activation changes are highly consistent with those reported previously. Patients show generally similar activations (Fig. 1, rows (D)–(F), P o0.005, cluster438, whole brain P-correctedo 0.05), with some group differences of interest as noted below. Compared with capital-judgment, the all-subject group showed greater activation of the dorsal MPFC, left IFG, temporal gyri, anterior insula and bilateral basal ganglia, and lower activation of ventral prefrontal gyrus, bilateral fusiform areas (extending to middle and superior temporal gyri), TPJ, post-central gyri and visual processing areas in referential tasks (Fig. 1, row (G)). This is also consistent with previous research (reviewed by van der Meer et al. (2010)). ANOVA revealed significant main effects of referential task in the anterior-dorsal part of MPFC, bilateral lingual gyri and right posterior middle and superior temporal gyri. Post hoc tests showed self 4president in the MPFC and self o president in bilateral lingual gyri and right posterior middle and superior temporal gyri through all subjects. When “self” was compared to “others” (“mother” and “president”), self 4others signals were detected in middle-posterior dorsal MPFC and precuneus/cuneus and self o others signals in left occipital gyrus (P o0.005, cluster4 38, whole brain P-correctedo0.05) in addition to the regions with self versus president difference. Activation increases in the left IFG appear to be lower in the self-reference than in motherand president-reference in the activation maps but the self versus others contrast in the left IFG was significant only at P o0.005 and cluster size 410 and not in the whole brain correction map. ANOVA showed multiple significant group differences (Table 2, P o0.025, cluster 471, whole brain P-correctedo 0.05). These differences are most notable in the self and president conditions, with patients showing greater activation or less deactivation than healthy controls in multiple regions during the self condition (Fig. 2, row (A)), and generally less activation or smaller decreases than controls in the president condition (Fig. 2, row (C)). Post hoc comparisons showed that in the self condition, patients showed greater activation than controls in right lateral prefrontal gyri and m/pCC, and less deactivation than controls in bilateral precuneus/ cuneus, right posterior temporal gyri, occipital-parietal gyri and left lateral inferior frontal gyrus extending to superior temporal gyrus. The patients also had greater activations in bilateral extrastriate visual cortex (Z-score¼ 3.06 at peak of MNI:  20,  74, 6; and Z-score¼2.66 at peak of MNI: 18, 92,  14) in this condition, although the areas did not survive when the whole brain correction threshold was applied. There were no significant group differences during the mother condition (Fig. 2, row (B)). In the president condition, patients showed less activation than controls in the left IFG, and bilateral insula with possible extension to the basal ganglia bilaterally, on the left to the thalamus, and on the right to the temporal gyrus, and less deactivation than controls in the right occipital-parietal gyri and bilateral basal ganglia.

3

Although the 2-group by 3-task interactions in the ANOVA did not survive whole brain correction, focused comparison on the primary question of self versus others (motherþ president) showed significant interactions in the 2-group by 2-task ANOVA (P o0.025, cluster471, whole brain P-correctedo0.05) in ventral prefrontal gyri, bilateral insula, temporal gyri, basal ganglia and left IFG (Fig. 2, row (D)). Among these regions, patients had greater self-related activation in left IFG, left anterior and right posterior temporal gyri; less president-related activation of left IFG, insula, basal ganglia, left thalamus, anterior temporal gyri and right posterior temporal gyri, and less president-related deactivations of ventral prefrontal gyri. Patients also showed greater self–others activation differences than healthy controls in bilateral extrastriate visual cortex (Z-score¼2.73 at peak of MNI:  24,  60, 6; Z-score¼ 2.97 at peak of MNI: 22, 92,  14; and Z-score¼2.85 at peak of MNI: 4,  54, 6), although this group difference did not survive whole brain correction.

4. Discussion Patients and healthy controls were more likely to endorse positive and reject negative attributes relative to self, mother and presidents (higher rates and faster response). However, patients were less likely and slower to endorse positive selfattributes, and more likely and quicker to endorse negative selfattributes than controls. Regional activation changes in all three referential tasks included central midline structures as shown previously. Multiple other regions also showed activation increases during the referential tasks as compared to passive viewing of gray screens: left IFG, temporal pole, insula and basal ganglion, visual processing areas, left parietal gyrus, right lateral frontal areas and posterior middle and superior temporal gyri. These areas increase activation during a range of perceptual, working memory, language processing and other neurocognitive tasks as well as during social cognition (Green and Horan, 2010), indicating the complex and distributed nature of the neural resources required for the referential task. Additional areas decreased activity during the referential tasks including that of precuneus/cuneus, sub-regions of parietal gyri and temporal gyri. Many of these are part of the default network that is typically more active in resting or low demand baseline conditions than during cognitive processing. Comparison of the referential tasks to the capital judgment task showed extension of activation in dorsal MPFC, left IFG, temporal gyri, anterior insula and basal ganglia associated with the social tasks. Patients showed multiple areas of greater signal increases or smaller signal decreases than healthy controls when thinking about themselves. Group differences were less evident when subjects were thinking about their mothers and tended to go in the opposite direction when thinking about a president. This apparent specificity of group differences related to the self was supported statistically when the self condition was compared to the mother and president conditions. The fact that group differences varied as a function of what type of person the subjects were thinking about, despite similar cognitive and response requirements across conditions, provides some new and specific information relevant to self-concept and social function in people with schizophrenia. Of greatest interest in considering the present findings in relation to social dysfunctions in people with schizophrenia are the abnormalities in activation changes when patients think about their own qualities. When thinking about these abnormalities, one could start with the behavioral abnormalities in social cognition and social function, and argue that the abnormalities we observe

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

4

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 1. The z-score maps (in radiological view) of the regions with significantly referential task-related percentage signal changes compared with resting fixation baseline and capital-judgment (Po 0.005, cluster4 38, whole brain P-corrected o0.05). (A), Self-related in controls; (B), Self-related in controls; (C), Mother-related in controls; (D), Mother-related in patients; (E), President-related in patients; (F), President-related in patients; and (G), Three social tasks versus capital-judgment (social task – capitaljudgment). Healthy controls showed the referential-related activation of the dorsal-posterior medial frontal cortex, left inferior prefrontal gyri and basal ganglion, and deactivation of the precuneus/cuneus (A-C). Patients showed the pattern only in self- and mother-related processing (D, E), but president-related (F). The contrast between referential and non-social tasks also showed the dorsal-posterior medial frontal cortex and left inferior prefrontal gyri had greater activation and the precuneus/cuneus had lower activation in social tasks (G).

are neural correlates specific to them. However, the activation abnormalities we observed are in areas that change activation in response to a variety of cognitive demands and have been shown to show abnormal activation and/or structural abnormality in

other types of studies of patients. For example, patients had greater self-reference-related activation increases than controls in the left lateral IFG, visual cortex, m/pCC and left superior temporal gyrus. These regions have been identified as a network

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

5

Table 2 Group difference of each referential condition and self- versus others-reference. Regionsa

Volumeb (mm3)

MNIc

Z-score at peak

x

y

z

3672 4239 4482 3348 4401 2619 3942 2268

 54  18 30 54 24 6  18 30

6  75  27 27  72  15  48  78

3 6 18 15 36 42 48  36

3.11 3.06 3.78 3.58 3.28 3.32 3.02 3.15

2808 3267

 15 3

9  39

 30  36

3.77 3.58

President: patients 4controls R Occipital and parietal gyrus B Basal ganglion

2322 2295

27 9

 60 18

36 9

3.12 3.40

President: patients o controls L Pre-central and post-central gyrus L Inferior prefrontal gyri, superior temporal gyrus and temporal pole, insula, basal ganglion and thalamus R Temporal gyri, insula and basal ganglion R Posterior temporal gyri B Cerebellum

8559 2673 2268 4941

 30 30 57 9

9 3  42  48

3  15 9  15

3.87 3.41 3.40 3.44

self-others: patients 4controls L Lateral inferior prefrontal gyrus, insula, basal ganglion, superior temporal gyrus and temporal pole R Superior temporal gyrus, lateral inferior prefrontal gyrus, insula, and basal ganglion R Posterior temporal gyri

2727 5022 6399

 30 42 60

9 0  39

3  15 3

3.54 3.32 3.90

self-others: patients o controls L Inferior temporal gyrus

2916

 54

 42

 24

3.91

Self: L L R R R B B B

patients 4controls Inferior prefrontal gyrus extending to superior temporal gyrus and posterior insula Calcarine and lingual areas Posterior insula, posterior temporal gyri Lateral prefrontal gyrus Occipital and parietal gyri Middle- and posterior-dorsal cingulate cortex Precuneus Cerebellum

Self: patients ocontrols L Inferior temporal gyrus B Cerebellum Mother: patients 4controls NONE Mother: patients o controls NONE

a b c

L-left, R-right, B-bilateral. The volume of the region with the Po 0.025, cluster471, whole brain corrected Po 0.05. The coordinate of the voxel with maximal Z-score in the Montreal Neurological Institute brain template space.

for visual language processing with different components showing greater relative activation in different stages of processing (Buckner et al., 1999; Dhond et al., 2001; Marinkovic et al., 2003; Petersen et al., 1988; Poldrack et al., 1999). Multiple studies have shown attenuated as well as enhanced hemodynamic response abnormalities in patients with schizophrenia during language tasks in both left superior temporal gyrus and lateral IFG (Han et al., 2007; Kuperberg et al., 2007; Tagamets et al., 2013; Wilson et al., 2013). Thus, patient task-related activation abnormalities in these regions are not limited to social or self-referential tasks. As a second example, both patients and healthy controls showed a robust activation in dorsal MPFC during self and other referential processing reported by other investigators. However, this activation extended more posteriorly in patients than controls, into m/ pCC, which is consistent with a posterior shift in midline activation during self-referential thinking reported by other investigators (Holt et al., 2011; Shad et al., 2012). More posterior midline activation in patients has also been reported during other activation paradigms and even at rest (Bertolino et al., 2006; Callicott et al., 2003; Li et al., 2007; Tan et al., 2007). So again, the activation abnormalities associated with self-referential thinking are evident during other task demands as well. We suggest, therefore, that it might be more useful to consider how more general neuropathology in schizophrenia, conceived of in neural terms, impacts on the development of sense of self in people with schizophrenia, rather than thinking that our findings indicate a neural abnormality

specifically or primarily associated with abnormalities in selfrepresentation. Patient over activation in bilateral dorsal and inferior prefrontal gyrus, right insula and left anterior temporal gyri, and the failure to decrease activity in right posterior temporal gyri, were all significantly greater when patients were thinking about themselves than when thinking about their mothers or a president. What might it mean to have greater lateral frontal, left temporal and insula activation and at the same time failure to decrease activity in right posterior temporal regions when thinking about oneself? At the most general level, the sense of self, including associated affective valuation, rests upon a different pattern of regional activation in patients than in healthy controls. In a somewhat more specific sense, the neural activation pattern associated with activation of an image of the self in patients is greater anterior lateral activation and smaller decreases in posterior activation, leading to higher levels of activity throughout the brain. Behaviorally, during the self-referential processing, patients were less likely and slower to endorse positive selfattributes and more likely and relatively quicker to endorse negative self-attributes. How might the activation abnormalities during fMRI relate to these and other behavioral abnormalities associated with social function? We hypothesize that the general neuropathology in schizophrenia creates abnormal and less stable self-representation system configurations that are more likely to be associated with negative affect and influenced by changing visceral (insula) and external sensory (auditory processing in posterior temporal gyrus

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

Fig. 2. The maps (in radiological direction) of group difference (P o0.025, cluster4 71, whole brain P-correctedo 0.05); the warm color presented patient 4control and cool color presented patient o control. (A) Self-reference; (B) mother-reference; (C), president-reference; (D) Patients(self–others)–Controls(self–others).

and visual areas) input. We recognize the preliminary and incomplete nature of this conceptualization, but think it is better than suggesting a narrower, self-contained formulation based on localizationist models of normal or abnormal function, or phrenology-like identification of psychological processes with neural ones, that are false and misleading.

with self-referential thinking, the present findings provide information about how general neural pathology in schizophrenia might affect patient thinking about self.

4.1. Limitations

This study was supported by the National Institute of Health grants to Dr. Wexler (R01MH084079 and KO2 MH01296). This study was also supported by the Connecticut Mental Health Center and the Connecticut State Department of Mental Health and Addiction Services. The authors wish to thank Dr. Morris Bell and Dr. Tek Cenk in setting paradigm. The authors also would like to thank Ms. Karen Martin, Ms. Cheryl Lea McMurray and Ms. Hickey Terry in MRI data collection, and Ms. Cheryl Lacadie for giving the instruction to Bioimagesuite.

Patients were on average 10 years older than the healthy controls. Although we used age as a covariate in the analyses, this procedure does not rule out the possible contribution of age to the group differences in regional activation during the self-referential task. The relatively small sample size limits power to detect possible additional activation abnormalities or confirm that additional activation abnormalities are specifically linked to self rather than other reference. We do not have structural and functional connectivity data to further evaluate system configurations and function. There are many psychosocial factors that we have not evaluated and that could contribute to differences between patients and controls in the self-condition, including limited education, social isolation, unemployment, stigma and overall negative self-image. This study could not identify the relative contributions of these factors and the illness per se.

5. Conclusions Brain activation abnormalities are different in patients with schizophrenia when thinking about themselves than when thinking about others. Abnormalities include an elevated activation in m/pCC as reported by others, extend to other regions and include both areas of over activation and failure to decrease activation. Patients also show clinically relevant behavioral differences when thinking about themselves but not others. Since some of the brain activation abnormalities noted have also been reported with other activation paradigms and even at rest, and presumably therefore are not specifically associated

Acknowledgments

References Anderson, N.H., 1968. Likableness ratings of 555 personality-trait words. Journal of Personality and Social Psychology 9, 272–279. Andreasen, N.C., Pressler, M., Nopoulos, P., Miller, D., Ho, B.C., 2010. Antipsychotic dose equivalents and dose-years: a standardized method for comparing exposure to different drugs. Biological Psychiatry 67, 255–262. Bell, M.D., Lysaker, P.H., Beam-Goulet, J.L., Milstein, R.M., Lindenmayer, J.P., 1994. Five-component model of schizophrenia: assessing the factorial invariance of the positive and negative syndrome scale. Psychiatry Research 52, 295–303. Bertolino, A., Caforio, G., Petruzzella, V., Latorre, V., Rubino, V., Dimalta, S., Torraco, A., Blasi, G., Quartesan, R., Mattay, V.S., Callicott, J.H., Weinberger, D.R., Scarabino, T., 2006. Prefrontal dysfunction in schizophrenia controlling for COMT Val158Met genotype and working memory performance. Psychiatry Research 147, 221–226. Buckner, R.L., Kelley, W.M., Petersen, S.E., 1999. Frontal cortex contributes to human memory formation. Nature Neuroscience 2, 311–314. Callicott, J.H., Mattay, V.S., Verchinski, B.A., Marenco, S., Egan, M.F., Weinberger, D.R., 2003. Complexity of prefrontal cortical dysfunction in schizophrenia: more than up or down. American Journal of Psychiatry 160, 2209–2215. Craik, F.I.M., Moroz, T.M., Moscovitch, M., Stuss, D.T., Winocur, G., Tulving, E., Kapur, S., 1999. In search of the self: a positron emission tomography study. Psychological Science 10, 26–34. Dhond, R.P., Buckner, R.L., Dale, A.M., Marinkovic, K., Halgren, E., 2001. Spatiotemporal maps of brain activity underlying word generation and their modification during repetition priming. Journal of Neuroscience 21, 3564–3571.

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i

J. Liu et al. / Psychiatry Research: Neuroimaging ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Fleisher, A.S., Podraza, K.M., Bangen, K.J., Taylor, C., Sherzai, A., Sidhar, K., Liu, T.T., Dale, A.M., Buxton, R.B., 2009. Cerebral perfusion and oxygenation differences in Alzheimer”s disease risk. Neurobiology of Aging 30, 1737–1748. Forman, S.D., Cohen, J.D., Fitzgerald, M., Eddy, W.F., Mintun, M.A., Noll, D.C., 1995. Improved assessment of significant activation in functional magnetic resonance imaging (fMRI): use of a cluster-size threshold. Magnetic Resonance in Medicine 33, 636–647. Green, M.F., Horan, W.P., 2010. Social cognition in schizophrenia. Current Directions in Psychological Science 19, 243–248. Han, S.D., Nestor, P.G., Hale-Spencer, M., Cohen, A., Niznikiewicz, M., McCarley, R.W., Wible, C.G., 2007. Functional neuroimaging of word priming in males with chronic schizophrenia. Neuroimage 35, 273–282. Harrison, P.J., 1999. The neuropathology of schizophrenia. A critical review of the data and their interpretation. Brain 122, 593–624 (Pt 4). Harrison, P.J., Weinberger, D.R., 2005. Schizophrenia genes, gene expression, and neuropathology: on the matter of their convergence. Molecular Psychiatry 10, 40–68 (image 45). Holt, D.J., Cassidy, B.S., Andrews-Hanna, J.R., Lee, S.M., Coombs, G., Goff, D.C., Gabrieli, J.D., Moran, J.M., 2011. An anterior-to-posterior shift in midline cortical activity in schizophrenia during self-reflection. Biological Psychiatry 69, 415–423. Kelley, W.M., Macrae, C.N., Wyland, C.L., Caglar, S., Inati, S., Heatherton, T.F., 2002. Finding the self? An event-related fMRI study. Journal of Cognitive Neuroscience 14, 785–794. Kuperberg, G.R., Deckersbach, T., Holt, D.J., Goff, D., West, W.C., 2007. Increased temporal and prefrontal activity in response to semantic associations in schizophrenia. Archives of General Psychiatry 64, 138–151. Lewis, D.A., 2000. GABAergic local circuit neurons and prefrontal cortical dysfunction in schizophrenia. Brain Research Reviews 31, 270–276. Lewis, D.A., Sweet, R.A., 2009. Schizophrenia from a neural circuitry perspective: advancing toward rational pharmacological therapies. The Journal of Clinical Investigation 119, 706–716. Li, X., Branch, C.A., Ardekani, B.A., Bertisch, H., Hicks, C., DeLisi, L.E., 2007. fMRI study of language activation in schizophrenia, schizoaffective disorder and in individuals genetically at high risk. Schizophrenia Research 96, 14–24. Lieberman, M.D., Jarcho, J.M., Satpute, A.B., 2004. Evidence-based and intuitionbased self-knowledge: an FMRI study. Journal of Personality and Socical Psychology 87, 421–435. Liu, J., Qiu, M., Constable, R.T., Wexler, B.E., 2012. Does baseline cerebral blood flow affect task-related blood oxygenation level dependent response in schizophrenia? Schizophrenia Research 140, 143–148. Lysaker, P.H., Carcione, A., Dimaggio, G., Johannesen, J.K., Nicolo, G., Procacci, M., Semerari, A., 2005. Metacognition amidst narratives of self and illness in schizophrenia: associations with neurocognition, symptoms, insight and quality of life. Acta Psychiatrica Scandinavica 112, 64–71. Marinkovic, K., Dhond, R.P., Dale, A.M., Glessner, M., Carr, V., Halgren, E., 2003. Spatiotemporal dynamics of modality-specific and supramodal word processing. Neuron 38, 487–497.

7

Murphy, E.R., Brent, B.K., Benton, M., Pruitt, P., Diwadkar, V., Rajarethinam, R.P., Keshavan, M.S., 2010. Differential processing of metacognitive evaluation and the neural circuitry of the self and others in schizophrenia: a pilot study. Schizophrenia Research 116, 252–258. Ochsner, K.N., 2008. The social-emotional processing stream: five core constructs and their translational potential for schizophrenia and beyond. Biological Psychiatry 64, 48–61. Parnas, J., Handest, P., Jansson, L., Saebye, D., 2005. Anomalous subjective experience among first-admitted schizophrenia spectrum patients: empirical investigation. Psychopathology 38, 259–267. Petersen, S.E., Fox, P.T., Posner, M.I., Mintun, M., Raichle, M.E., 1988. Positron emission tomographic studies of the cortical anatomy of single-word processing. Nature 331, 585–589. Poldrack, R.A., Wagner, A.D., Prull, M.W., Desmond, J.E., Glover, G.H., Gabrieli, J.D., 1999. Functional specialization for semantic and phonological processing in the left inferior prefrontal cortex. NeuroImage 10, 15–35. Powers, R.E., 1999. The neuropathology of schizophrenia. Journal of Neuropathology and Experimental Neurology 58, 679–690. Selemon, L.D., Goldman-Rakic, P.S., 1999. The reduced neuropil hypothesis: a circuit based model of schizophrenia. Biological Psychiatry 45, 17–25. Shad, M.U., Keshavan, M.S., Steinberg, J.L., Mihalakos, P., Thomas, B.P., Motes, M.A., Soares, J.C., Tamminga, C.A., 2012. Neurobiology of self-awareness in schizophrenia: an fMRI study. Schizophrenia Research 138, 113–119. Tabelow, K., Clayden, J.D., de Micheaux, P.L., Polzehl, J., Schmid, V.J., Whitcher, B, 2011. Image analysis and statistical inference in neuroimaging with R. Neuroimage 55, 1686–1693. Tagamets, M.A., Cortes, C.R., Griego, J.A., Elvevag, B., 2013. Neural correlates of the relationship between discourse coherence and sensory monitoring in schizophrenia. Cortex. Tan, H.Y., Callicott, J.H., Weinberger, D.R., 2007. Dysfunctional and compensatory prefrontal cortical systems, genes and the pathogenesis of schizophrenia. Cerebral Cortex 17 (Suppl 1), i171–181. van der Meer, L., Costafreda, S., Aleman, A., David, A.S., 2010. Self-reflection and the brain: a theoretical review and meta-analysis of neuroimaging studies with implications for schizophrenia. Neuroscience and Biobehavioural Reviews 34, 935–946. Wexler, B.E., Halwes, T., 1983. Increasing the power of dichotic methods: the fused rhymed words test. Neuropsychologia 21, 59–66. Wilson, L.B., Rojas, D.C., Shatti, S., Tregellas, J.R., 2013. Greater neuronal responses during automatic semantic processing in schizophrenia. NeuroReport 24, 212–216. Zhang, L., Zhou, T.G., Zhang, J., Liu, Z.X., Fan, J., Zhu, Y., 2006. In search of the Chinese self: An fMRI study. Science in China Series C 49, 89–96. Zhu, Y., Zhang, L., Fan, J., Han, S., 2007. Neural basis of cultural influence on selfrepresentation. Neuroimage 34, 1310–1316.

Please cite this article as: Liu, J., et al., Neural activation abnormalities during self-referential processing in schizophrenia: An fMRI study. Psychiatry Research: Neuroimaging (2014), http://dx.doi.org/10.1016/j.pscychresns.2014.04.003i