حافظه کاذب به محرک های عاطفی در بیماران سکته مغزی آسیب دیده در نیمکره های راست ـ چپ مغز به یک اندازه تحت تاثیر قرار نمی گیرند

Previous research has attributed to the right hemisphere (RH) a key role in eliciting false memories to visual emotional stimuli. These results have been explained in terms of two right-hemisphere properties: (i) that emotional stimuli are preferentially processed in the RH and (ii) that visual stimuli are represented more coarsely in the RH. According to this account, false emotional memories are preferentially produced in the RH because emotional stimuli are both more strongly and more diffusely activated during encoding, leaving a memory trace that can be erroneously reactivated by similar but unstudied emotional items at test. If this right-hemisphere hypothesis is correct, then RH damage should result in a reduction in false memories to emotional stimuli relative to left-hemisphere lesions. To investigate this possibility, groups of right-brain-damaged (RBD, N = 15), left-brain-damaged (LBD, N = 15) and healthy (HC, N = 30) participants took part in a recognition memory experiment with emotional (negative and positive) and non-emotional pictures. False memories were operationalized as incorrect responses to unstudied pictures that were similar to studied ones. Both RBD and LBD participants showed similar reductions in false memories for negative pictures relative to controls. For positive pictures, however, false memories were reduced only in RBD patients. The results provide only partial support for the right-hemisphere hypothesis and suggest that inter-hemispheric cooperation models may be necessary to fully account for false emotional memories.

Emotional stimuli are remembered better and more vividly than non-emotional stimuli (Hamann, 2001 and Kensinger, 2004). This phenomenon, known as the emotional enhancement of memory, has been replicated across a range of paradigms and stimulus types (e.g., Borsutzky et al., 2010, Bradley et al., 1992, Kensinger and Corkin, 2004, Nagae and Moscovitch, 2002 and Talmi et al., 2008) and is particularly robust for arousing events (e.g., Anderson et al., 2006, Christman et al., 2004, Ochsner, 2000 and Schaefer et al., 2011). However, not all aspects of memory are enhanced by emotional stimuli (Bradley et al., 1992). Peripheral features of visual scenes are remembered less well when an emotional item is present in the scene than when only non-emotional items are present (Kensinger et al., 2007a and Talmi et al., 2008). In addition, memory for scene details can be impaired by emotionality, even when these details belong to a central element of the scene (Adolphs, Denburg, et al., 2001 and Denburg et al., 2003). For example, participants may remember well a picture of a dead body compared to a picture of a living person (gist memory), but they may remember less well the spatial orientation of the body than the orientation of the living person (memory for scene details) ( Adolphs, Denburg, et al., 2001; but see Kensinger, 2009, for a different perspective). Perhaps more surprisingly, emotional stimuli can also induce more false memories than non-emotional stimuli (Dehon et al., 2010 and Porter et al., 2003). For example, Porter et al. (2003) showed negative, positive or neutral pictures to different groups of participants and asked them a few questions, some of which contained misleading information about the pictures (e.g., that there was a large animal). When asked to recall the pictures 1 h later, twice as many participants in the negative group falsely recalled the misleading information compared to participants in the positive and neutral groups. This apparent paradox – that negative emotion can simultaneously improve and impair memory – has been repeatedly found in recognition memory experiments with word stimuli (Brainerd et al., 2008, Grassi-Oliveira et al., 2011 and Maratos et al., 2000), even when potential confounds, such as concreteness or semantic cohesiveness, are taken into account (Dehon et al., 2010 and McNeely et al., 2004). Thus, results from studies using words and complex scenes suggest that highly arousing negative stimuli can increase both true and false memories. Research into the neural correlates of emotional memories (LaBar & Cabeza, 2006) and false memories (Schacter & Slotnick, 2004) started uncovering a network of brain structures that are commonly involved in these phenomena, including amygdala, hippocampus, pre-frontal, orbitofrontal and parietal cortices. Less is known, however, about the neural structures underlying false emotional memories and whether or not, like emotional processing, these networks show some degree of lateralization. In the following, we briefly review the literature implicating right-hemisphere structures in both emotional processing and false memories and outline the main goals and hypotheses of the present study. 1.1. Right hemisphere and emotional processing The right brain hemisphere (RH) has been consistently linked to a preferential processing of emotional stimuli in comparison to the left hemisphere (LH) (Abbott et al., 2013, Borod et al., 2002, Demaree et al., 2005, Kucharska-Pietura, 2006 and Witteman et al., 2011). However, there is still debate about which processes (expression vs. perception) and types (negative vs. positive) of emotion are best supported by right-hemisphere networks. For emotional perception, two main hypotheses have been put forward. The right-hemisphere hypothesis posits that the RH specializes in processing both positive and negative emotions (e.g., Borod et al., 2002). The valence-specific hypothesis, on the other hand, posits that the RH specializes in negative emotions, whereas the LH specializes in positive emotions (e.g., Davidson, 1992). Consistent with the right-hemisphere hypothesis, Borod et al. (1998) found that perception of emotional faces, prosody and written words was impaired in right-brain-damaged patients compared to left-brain-damaged patients and healthy controls, who did not differ from each other. Similarly, Alves, Aznar-Casanova, and Fukusima (2009) showed that perception of emotional faces in healthy participants was faster when the faces were presented to participants’ RH (via their left visual field) than when the faces were presented to participants’ LH (via their right visual field), suggesting that emotional stimuli are preferentially processed in the RH. By contrast, Natale, Gur, and Gur (1983) showed that participants judged faces with negative expressions as more negative when they were presented to the RH than when they were presented to the LH. Conversely, participants judged faces with positive expressions as more positive when they were presented to the LH than when they were presented to the RH. These results directly supported the valence-specific hypothesis. Additional evidence for the valence-specific hypothesis came from neuroimaging and electrophysiological studies. Canli, Desmond, Zhao, Glover, and Gabrieli (1998) found in a functional magnetic resonance imaging (fMRI) study that brain activation was stronger in the RH when participants saw a sequence of negative pictures and stronger in the LH when they saw a sequence of positive pictures. Likewise, Davidson (1992) found in several electroencephalogram (EEG) studies that brain activity was higher in right frontal electrodes when participants reacted to negative film clips and higher in left frontal electrodes when they reacted to positive clips. More recent results, however, suggest that these hypotheses may not capture the complexity of emotional processing. In an fMRI study, Killgore and Yurgelun-Todd (2007) presented sad and happy chimeric faces very briefly to normal participants who were only required to determine the sex of the face (but were not asked to make any overt emotional judgement). The pattern of brain activations, which was linked to the non-conscious perception of affective faces presented to either hemisphere, showed that not only the RH was more responsive than the LH to both face types (a result consistent with the right-hemisphere hypotheses) but also that the LH was more responsive to sad faces than to happy faces (a result inconsistent with both the right-hemisphere and the valence-specific hypotheses). More surprisingly, Paradiso, Anderson, Ponto, Tran, and Robinson (2011) reported a group of patients with stable right-hemisphere lesions who showed an impairment relative to healthy controls in their ability to judge the emotionality of positive pictures but no impairment in their ability to judge negative pictures, a result that supports only partially the right-hemisphere hypothesis and directly contradicts the valence-specific hypothesis. Taken together, these results indicate a lack of consensus regarding laterality and emotional processing, which might be a result of different experimental paradigms, sample characteristics and stimulus types across studies. However, as most evidence supports a relative specialization of the RH towards emotion perception ( Abbott et al., 2013, Adolphs, Jansari, 2001, Borod et al., 2002, Charbonneau et al., 2003, Kucharska-Pietura et al., 2003 and Nijboer and Jellema, 2012), we adopt this view to derive our predictions. 1.2. Right hemisphere and false memories In addition to its role in emotional processing, the RH has also been implicated in the production of false memories (e.g., Marchewka et al., 2009 and Westerberg and Marsolek, 2003). Patients undergoing the intracarotid amobarbital sodium procedure, which selectively anesthetizes only one hemisphere at a time, show a marked increase in false alarms during recognition memory tests following LH injection (Loring, Lee, & Meador, 1989). That is, patients incorrectly say more often that an unstudied test item has been studied when the RH is operational and the left is anesthetized than when the LH is operational and the right is anesthetized, suggesting that false alarms are generated by processes at play in the RH. Patients with RH lesions, particularly in frontal regions, have also been shown to produce more false memories than controls in studies using words (Delbecq-Derouesne, Beauvois, & Shallice, 1990), faces (Rapcsak, Polster, Glisky, & Comer, 1996), and pictures (Schacter, Curran, Galluccio, Milberg, & Bates, 1996). Consistent with these results, a structural neuroimaging study has shown that healthy participants who generated the highest levels of false memories in a recognition memory test using pictures also possessed the lowest densities of gray matter in their right frontal gyrus (Marchewka et al., 2009). In healthy participants, most evidence that the RH is preferentially involved in producing false memories comes from studies combining the divided visual field technique (Bourne, 2006) with the DRM paradigm (Roediger & McDermott, 1995). In the DRM paradigm, participants study lists of words (e.g., candy, sugar, tooth) that are all related to a single unstudied word (e.g., sweet). In a subsequent recognition test, participants often incorrectly believe that the related word was present in the study list. The DRM paradigm has been widely used to investigate false memories in normal and patient groups alike ( Gallo, 2010). When the DRM paradigm is coupled with the divided field technique, the common finding is that false memories are higher when test words are presented to the left visual field (right hemisphere) than when they are presented to the right visual field (left hemisphere) ( Ben-Artzi et al., 2009, Faust et al., 2008, Giammattei and Arndt, 2012, Marchewka et al., 2009, Schmitz et al., 2013 and Westerberg and Marsolek, 2003). These results have been interpreted in the context of fine-coarse coding theory ( Beeman et al., 1994 and Jung-Beeman, 2005). According to this theory, input stimuli activate semantic networks that have different structural and functional properties in the left and right hemispheres. In the LH, stimuli strongly activate small semantic networks that represent the dominant meaning of the input. In the RH, stimuli weakly activate large networks that represent the meanings of the input and of its semantic associates. In this view, stimulus representation is fine in the LH and coarse in the RH (but see Marsolek, 1999 and Marsolek and Burgund, 2008, for a different account). Although the theory was originally developed in the context of language comprehension, recent evidence from divided visual field studies has extended its applicability to verbal memory ( Ben-Artzi et al., 2009 and Faust et al., 2008) and picture comprehension ( Lovseth & Atchley, 2010). Fine-coarse coding theory can account for false memories in the RH because studied stimuli, words or pictures, are more likely to activate the representations of related stimuli in the (gist-based) right hemisphere than in the (veridical) left hemisphere. The activation of the unstudied, related stimuli in the RH can then increase the likelihood of participants falsely accepting them as if they had been previously studied. Thus, a number of studies in both healthy participants and brain damaged patients indicate that the RH plays a prominent role in the creation of false memories, a phenomenon that can be accounted for by the right-hemisphere’s coarser representation of studied stimuli. 1.3. Present study Despite the wealth of evidence implicating the RH in both emotional processing and false memories, there is surprisingly little research into the role of the RH in false emotional memories. In the few studies that have investigated the lateralization of emotional memories, only data for true memories was reported (Mneimne et al., 2010 and Nagae and Moscovitch, 2002). To our knowledge, only one study has directly examined the lateralization of false emotional memories (Marchewka et al., 2008). In Marchewka et al.’s (2008) fMRI study, participants saw a mixture of negative and neutral complex scenes taken from the International Affective Picture System (IAPS; Lang, Bradley, & Cuthbert, 2008) and then completed a recognition memory test. The pictures were briefly presented (400 ms) to participants’ left or right visual fields at both encoding and retrieval stages, and fMRI scans were obtained during the retrieval stage. Two interesting findings emerged from that study. The first was that false-alarm rates were equivalent between the hemispheres when the pictures were neutral, but were higher in the RH when the pictures were negative. The second result was that false-alarm rates were associated with stronger activation in the right pre-frontal cortex relative to correct rejections (when participants answer “no” for a picture that has not been previously presented), and that this activation was also stronger for negative than for neutral pictures. Marchewka et al.’s (2008) results are consistent with both the right-hemisphere hypothesis and valence-specific hypothesis of hemispheric specialization of emotional processing. Their results are also consistent with the fine-coarse coding account of false memories. In the present study, we investigate hemispheric asymmetries in emotional memories by testing groups of right-brain-damaged (RBD), left-brain-damaged (LBD) and healthy control (HC) participants in a recognition memory task. False memories were operationalized as the false-alarm rate to related, unstudied pictures. This definition is commonly used in memory research (for reviews, see Schacter et al., 1998 and Schacter and Slotnick, 2004) and represents the proportion of novel items that are wrongly interpreted as “old” by participants. By construction, novel items are related to studied items and, consequently, share with them semantic and/or perceptual features. For example, participants may see at study a picture of an angry dog and at test a picture of a different angry dog. Because studied and tested pictures share conceptual and perceptual characteristics, participants may incorrectly think at test that the picture was previously seen at study. To the extent that novel, related test items are mistaken for previously studied items, they can be used to gauge false memories. This type of false memory, also called gist-based false recognition ( Garoff-Eaton et al., 2007 and Gutchess and Schacter, 2012), has been shown to increase with age ( Koutstaal & Schacter, 1997) and as a result of frontal-lobe lesions ( Curran et al., 1997 and Rapcsak et al., 1996) and represents only one among several manifestations of false-memory phenomena ( Brainerd and Reyna, 2002, Gallo, 2010, Kopelman, 1999 and Schacter and Slotnick, 2004). This study extends Marchewka et al.’s (2008) work by including both negative and positive emotional pictures, which allows us to contrast directly the right-hemisphere hypothesis with the valence-specific hypothesis. Moreover, we report and analyze not only data for false memories but also data for hits (true memories) and for false alarms to unrelated, unstudied pictures (a measure of response bias). Both, true memories and response bias are important in helping to constrain potential accounts of false-memory data (Wixted & Stretch, 2000). If the right-hemisphere hypothesis is correct, then RBD patients should produce fewer false memories to emotional stimuli than both LBD patients and healthy controls. The prediction follows from the assumption that the RH (i) specializes in processing emotions in general (both negative and positive) and (ii) specializes in representing stimuli in a coarse manner. Lesions to the RH should thus reduce false memories more in the case of emotional pictures (negative and positive) than in the case of neutral pictures, because the stronger but coarser representations of emotional stimuli in the RH, which enhance false memories, are more degraded in RBD than in LBD patients. If, on the other hand, the valence-specific hypothesis is correct, then RBD patients should produce fewer false memories to negative stimuli than both LBD and healthy controls, whereas LBD patients should produce fewer false memories to positive stimuli than RBD patients and controls. These predictions follow from the assumption that the RH specializes in processing negative emotions and the LH specializes in processing positive emotions. Lesions to the RH should thus reduce false memories to negative pictures, because the representations of negative stimuli are degraded. By contrast, lesions to the LH should reduce false memories to positive pictures, because the representations of positive stimuli are degraded. Finally, following fine-coarse coding theory, it is predicted that false memories to neutral items should be lower in RBD than in LBD patients, because the more veridical LH is intact in RBD patients, which should contribute to reduce false memories, but is degraded in LBD patients, which should contribute to increase false memories.