افزایش فعالیت لوب تمپورال داخلی در طول مشاهده منفعل حالات چهره عاطفی و خنثی در اسکیزوفرنی

Abstract Introduction Patients with schizophrenia show deficits in facial affect and facial identity recognition and exhibit structural and neurophysiological abnormalities in brain regions known to mediate these processes. Functional neuroimaging studies of neural responses to emotional facial expressions in schizophrenia have reported both increases and decreases in medial temporal lobe (MTL) activity in schizophrenia. Some of this variability may be related to the tasks performed and the baseline conditions used. Here we tested whether MTL responses to human faces in schizophrenia are abnormal when unconstrained by a cognitive task and measured relative to a low-level baseline (fixation) condition. Methods 15 patients with schizophrenia and 16 healthy control subjects underwent functional magnetic resonance imaging (fMRI) while passively viewing human faces displaying fearful, happy, and neutral emotional expressions. Results Relative to control subjects, the patients demonstrated (1) significantly greater activation of the left hippocampus while viewing all three facial expressions and (2) increased right amygdala activation during the initial presentation of fearful and neutral facial expressions. Conclusions In schizophrenia, hippocampal and amygdala activity is elevated during the passive viewing of human faces. Keywords Schizophrenia; Functional magnetic resonance imaging; Amygdala; Hippocampus; Emotion; Face

. Introduction Over the past several decades, there has been increasing interest in understanding the role of affect processing abnormalities in the acute symptoms and the functional impairment observed in schizophrenia. Severity of emotional perception deficits (Kee et al., 2003) and negative symptoms, which include affective flattening, apathy and anhedonia (Milev et al., 2005), predict poor psychosocial outcome in schizophrenia. Also, it has been hypothesized that positive symptoms (delusions and hallucinations) arise from fundamental abnormalities in emotional perception and social cognition (Bentall et al., 2001 and Phillips et al., 2003). One basic component of social cognitive functioning is facial affect recognition; recognizing and discriminating among distinct facial emotional expressions allows one to infer the state of mind of others. Both positive (Hall et al., 2004) and negative (Kohler et al., 2000, Martin et al., 2005 and Sachs et al., 2004) symptoms have been correlated with poor facial affect recognition in schizophrenia. However, there is also evidence that these facial affect recognition deficits may be at least partially attributable to a number of factors unrelated to affect processing including a primary deficit in face perception (Onitsuka et al., 2003 and Quintana et al., 2003), the overall neurocognitive impairments associated with the disorder (Addington and Addington, 1998, Bozikas et al., 2004, Kohler et al., 2000, Sachs et al., 2004 and Whittaker et al., 2001), illness chronicity (Mueser et al., 1997 and Penn et al., 2000) and medication effects (Whittaker et al., 2001). Impaired facial affect processing in schizophrenia may be related to dysfunction of the amygdala and hippocampus. The amygdala is known to be crucial for the accurate recognition of facial expressions, particularly fear (Adolphs et al., 1994). Recently, it has become clear that the hippocampus, primarily known for its role in declarative memory, is also involved in the formation of emotional memories and associations (Buchel et al., 1999, Phelps, 2004 and Sanders et al., 2003). During affect processing, the hippocampus and amygdala may influence one another via bi-directional projections (Krettek and Price, 1977 and Pitkanen et al., 2000) and projections to common efferent targets (Canteras and Swanson, 1992, Nauta, 1986 and Ragsdale and Graybiel, 1988). Evidence for amygdala–hippocampal interactions during affect processing has been provided by human neuroimaging studies which have shown concurrent, and in some cases correlated (Dolcos et al., 2004 and Kensinger and Corkin, 2004), activation of the amygdala and hippocampus during the successful encoding of emotional information (Dolcos et al., 2004, Maratos et al., 2001 and Smith et al., 2004), viewing of emotional facial expressions (Gur et al., 2002a and Williams et al., 2001), and acquisition of conditioned responses to aversive stimuli (Buchel et al., 1999). Because of these studies and evidence from post-mortem and morphometric MRI studies for structural changes in the hippocampus and amygdala (Heckers and Konradi, 2002, Nelson et al., 1998 and Wright et al., 2000), we hypothesized that facial affect processing in schizophrenia is associated with hippocampal and amygdala dysfunction. Previous functional neuroimaging studies of neural responses to emotional facial expressions in schizophrenia have demonstrated decreases in the amygdala (Gur et al., 2002b, Hempel et al., 2003, Phillips et al., 1999, Schneider et al., 1998 and Williams et al., 2004) and hippocampus (Gur et al., 2002b and Hempel et al., 2003). However, one functional magnetic resonance imaging (fMRI) study showed an increased response of the right amygdala to happy faces (Kosaka et al., 2002), and another demonstrated a sustained response of the right anterior hippocampus to fearful faces (Holt et al., 2005) in patients with schizophrenia. Differences among these studies in the baseline condition used or in the cognitive task performed by the subjects may account for some of the discrepancies among these results. In contrast to the studies which found increases (Holt et al., 2005 and Kosaka et al., 2002), several studies which found abnormally reduced amygdala and/or hippocampal activation in schizophrenia used neutral (Phillips et al., 1999 and Williams et al., 2004) or inverted neutral (Hempel et al., 2003) faces as the baseline condition. Amygdala activation occurs in response to neutral faces (Schwartz et al., 2003 and Somerville et al., 2004) and patients with schizophrenia frequently respond to neutral stimuli–faces and words rated as “neutral” by healthy volunteers–as if they were emotional or arousing (Holt, submitted for publication, Kohler et al., 2003 and Williams et al., 2004). Thus, studies which measure neural responses to emotional facial expressions relative to responses to neutral expressions may underestimate the magnitude of hippocampal or amygdala activation. Also, other than one study which used an affect induction paradigm (Schneider et al., 1998), in these previous studies subjects performed a cognitive task such as gender discrimination (Phillips et al., 1999 and Williams et al., 2004), facial affect discrimination (Gur et al., 2002b, Hempel et al., 2003 and Kosaka et al., 2002) or facial affect labeling (Hempel et al., 2003). Performance of a cognitive task can reduce MTL responses to affective stimuli in healthy subjects (Lange et al., 2003 and Taylor et al., 2003), possibly due to an inverse relationship between activity in prefrontal and medial temporal cortices (Kim et al., 2003 and Pezawas et al., 2005). Recently, it has been shown that coupling of prefrontal and medial temporal activity observed in healthy subjects may be augmented or abnormally persistent in patients with schizophrenia, particularly during conditions of increased cognitive load (Meyer-Lindenberg et al., 2005). Thus, task performance may affect magnitudes of MTL responses to emotional stimuli, as well as to non-emotional stimuli, to an extent which may vary with psychopathology, cognitive capacity, and task difficulty. Here we conducted a block-design fMRI study in which patients with schizophrenia and healthy control subjects passively viewed fearful, happy, and neutral facial expressions. Hemodynamic responses in the amygdala and hippocampus to fearful, happy, and neutral facial expressions were compared with responses to a low-level baseline condition (viewing of a fixation cross). As a control, activation in the fusiform face area (FFA) to each face category was also measured and compared between groups. In addition, because MTL activity habituates rapidly with stimulus repetition (Breiter et al., 1996 and Wright et al., 2001) and a block-design study is limited in its capacity to detect transient changes in neural function, we measured amygdala and hippocampal activation during the first block of each facial expression category.

3. Results 3.1. Post-scanning debriefing results A repeated-measures ANOVA on the post-scanning debriefing data revealed no group by emotion interaction (F = 2.3, df = 2, p = .11); mean post-scan ratings of the intensity of the fearful, happy, and neutral emotional expressions of the faces were not significantly different between the two groups ( Table 2). Table 2. Post-scanning debriefing results: affect intensity ratings (from 1 to 6; 6=highest intensity) for the faces viewed Face Category Schizophrenia (N = 15) Controls (N = 16) p Fearful 4.88 ± 1.03 5.20 ± 0.78 0.33 Happy 5.25 ± 0.68 4.80 ± 1.21 0.21 Neutral 4.81 ± 1.22 4.13 ± 1.30 0.15 Values are presented as means ± standard deviations. Table options 3.2. fMRI results 3.2.1. Activation in the MTL and FFA to fearful, happy, and neutral facial expressions For the Fearful vs. Fixation, Happy vs. Fixation, and Neutral vs. Fixation contrasts, both the control subjects and the patients demonstrated bilateral MTL activation (Table 3). Comparisons between the two groups revealed greater left hippocampal activation in the patients with schizophrenia relative to the controls for all three contrasts (Table 3; Fig. 1). This pattern of elevated left hippocampal activation in the patients to fearful and happy faces was not found when neutral faces were used as the baseline condition; there were no between-group differences in MTL activation for the Fearful vs. Neutral and Happy vs. Neutral contrasts. No regions of the MTL showed greater activation in the control group relative to the patients. Table 3. Anatomical locations, Talairach coordinates, and z scores for peak MTL and FFA activations to fearful, happy, and neutral faces in the controls (N = 16), patients with schizophrenia (N = 15), and the between-group differences Contrast Region Control (N = 16) Schizophrenia (N = 15) Schizophrenia > Control x,y,z z score x,y,z z score x,y,z z score Fearful vs. Fixation Left amygdala − 22, − 8, − 14 3.22 Left hippocampus − 34,− 18,− 22 3.03 − 22,− 24,− 14 4.66 − 22,− 24,− 14 3.42 Right amygdala and hippocampus⁎ 24,− 8,− 20 3.04 26,− 6,− 16 3.77 Left fusiform gyrus − 38,− 47,− 15 4.33 − 36,− 36,− 17 5.3 Right fusiform gyrus 42,− 42,− 20 4.48 42,− 42,− 16 5.7 Happy vs. Fixation Left amygdala and hippocampus − 24,− 12,− 16 3.13 − 20,− 8,− 16 2.95 Left hippocampus − 22,− 16,− 14 2.95 − 22,− 24,− 14 2.95 Right amygdala 22,− 8,− 20 4.04 Right amygdala and hippocampus⁎ 26,− 6,− 16 3.22 Right hippocampus 22,− 20,− 12 3.44 28,− 16,− 12 3.55 Left fusiform gyrus − 40,− 46,− 20 4.93 − 40,− 42,− 20 5.26 Right fusiform gyrus 40,− 46,− 16 4.18 40,− 42,− 16 4.76 Neutral vs. Fixation Left amygdala − 28,− 10,− 18 3.41 − 16,− 4,− 18 2.71 Left hippocampus − 32,− 20,− 10 3.43 − 22,− 20,− 10 3.92 − 22,− 22,− 12 2.91 Right amygdala 28,− 8,− 20 3.44 30,− 4,− 16 2.9 Right hippocampus 24,− 22,− 10 3.03 24,− 18,− 12 3.30 Left fusiform gyrus − 40,− 44,− 20 4.67 − 42,− 44,− 18 5.89 Right fusiform gyrus 38,− 38,− 17 4.2 40,− 40,− 20 5.03 First Block Analysis Fearful vs. Fixation Right amygdala 30,− 4,− 26 3.66 28,0,− 24 3.30 Happy vs. Fixation Left amygdala − 18,2,− 18 2.89 − 30,− 4,− 26 2.67 Right amygdala 24,0,− 18 3.14 Neutral vs. Fixation Right amygdala 28,2,− 22 2.68 26,2,− 24 2.94 Those that show significantly greater activation in the schizophrenia vs. the control group are displayed in bold. ⁎ Although the voxel with the peak activation is in the amygdala, this cluster extends into the hippocampus. Table options Increased left hippocampus activation in the patients with schizophrenia (n=15) ... Fig. 1. Increased left hippocampus activation in the patients with schizophrenia (n = 15) relative to controls (n = 16) during the passive viewing of (A) fearful faces (Talairach coordinates (x,y,z) and z score of the voxel with the maximum activation difference: − 22, − 24,− 14, z = 3.42), (B) happy faces (− 22, − 24, − 14, z = 2.95), and (C) neutral faces (− 22, − 24,− 12, z = 2.91). Significance and display threshold = p < 0.005. Figure options For each of these three contrasts, both the control and schizophrenia group demonstrated bilateral activation of the FFA. Direct comparisons of the two groups showed that there was no between-group differences in the FFA activation (Table 3). 3.2.2. The First Block analysis We measured MTL activation during the first block of each category of facial expression. For the Fearful vs. Fixation and Neutral vs. Fixation contrasts, the patients exhibited right amygdala activation while the control subjects did not exhibit any significant MTL activation (Table 3). Direct comparisons between the two groups confirmed the patterns observed for the within-group analyses; greater right amygdala activation in the schizophrenia group relative to the controls was found for the Fearful vs. Fixation and Neutral vs. Fixation contrasts (Table 3; Fig. 2). Increased right amygdala activation in the patients with schizophrenia (n=15) ... Fig. 2. Increased right amygdala activation in the patients with schizophrenia (n = 15) relative to controls (n = 16) during the first presentation of (A) fearful faces (Talairach coordinates (x,y,z) and z score of the voxel with the maximum activation difference: 28, 0, − 24; z = 3.30) and (B) neutral faces (26, 2, − 24; z = 2.94). Significance and display threshold = p < 0.005. Figure options For the Happy vs. Fixation contrast, both subject groups demonstrated amygdala activation and direct comparisons revealed no between-group differences (Table 3). Finally, no MTL region showed greater activation in the control relative to the schizophrenia group.