Eingang zum Volltext in OPUS
Hinweis zum Urheberrecht
Dissertation zugänglich unter
Attentional and Neural Mechanisms Underlying the Trade-Off Effects of Negative Emotion on Memory
Der Trade-Off-Effekt von negativen Emotionen auf die Erinnerungsleistung : zugrunde liegende Aufmerksamkeitsprozesse und neuronale Korrelate
Kim, Johann Sung-Cheul
Dokument 1.pdf (1.869 KB)
Freie Schlagwörter (Deutsch):
Emotion , Gedächtnis , Aufmerksamkeit , fMRI , Eye-Tracking
Freie Schlagwörter (Englisch):
Emotion , Memory , Attention , fMRI, Eye-Tracking
77.37 , 77.31 , 77.62 , 77.46
Gamer, Matthias (Prof. Dr.)
Tag der mündlichen Prüfung:
Kurzfassung auf Deutsch:
Studien zum emotionalen Gedächtnis haben gezeigt, dass bei negativem Kontext differnzierte Effekte auf die Gedächtnisleistung für Details eines Erlebnisses festzustellen sind. Bei einem negativem Kontext werden Details besser erinnert, wenn sie von zentraler Relevanz sind, während periphere Details schlechter erinnert werden (für Reviews, siehe Christianson, 1992; Kensinger, 2009; Levine & Edelstein, 2009; Mather & Sutherland, 2011; Reisberg & Heuer, 2004). Der emotionale trade-off Effekt wird in verhaltensbasierten Studien und in Studien zu neuronalen Korrelaten gemeinhin mit Unterschieden in der Aufmerksamkeit begründet (Christianson, 1992; Adolphs et al., 2005; Kensinger et al., 2007; Kensinger & Schacter, 2006; Kensinger, 2009; LaBar & Cabeza, 2006; Mather et al., 2006; Phelps, 2006). Bei negativem Kontext, so die verbreitete Begründung, fokussiere sich die Aufmerksamkeit verstärkt auf zentrale Informationen, während periphere Informationen weniger fokussiert würden und dies bewirke den emotionalen trade-off bei der Erinnerungsleistung.
Unser Interesse galt der Untersuchung der trade-off Effekte und möglichen Erklärungen dieser Effekte durch Aufmerksamkeitsprozesse und neuronale Korrelate. Dabei war ein zentrales Anliegen die Generalisierbarkeit von experimentellen Befunden auf lebensechte Ereignisse. Daher haben wir 13 negative und 13 neutrale Bildergeschichten entwickelt und Probanden zur inzidenziellen Enkodierung dargeboten. Blickbewegungen wurden mit einem Eye-Tracker aufgezeichnet und neuronale Aktivität wurde ebenfalls erfasst. Erinnerungsleistung wurde anhand von zentralen und peripheren Objekten aus den Bildergeschichten getestet und zudem wurden Nacherzählungen der Bildergeschichten ausgewertet. Erwartungsgemäß wurden negative Bildergeschichten als erregender und negativer beurteilt. Ebenso zeigte sich mit einer signifikanten Interaktion zwischen Emotion (negativ vs. neutral) und Relevanz (zentral vs. peripher) der emotionale trade-off Effekt in der Gedachtnisleistung. Entgegen der allgemeinen Annahme scheint der emotionale trade-off Effekt jedoch nicht durch Aufmerksamkeitsprozesse mediiert zu sein. Analysen der Blickbewegungen legen dagegen folgenden Befund nahe: der prädiktive Zusammenhang von Aufmerksamkeitsparametern (während der Enkodierung der Bildergeschichten) für die spätere Erinnerungsleistung von zentralen und peripheren Details nimmt ab, je negativer Bildergeschichten erlebt werden. Im Gegensatz zur Mediationshypothese deuten unsere Ergebnisse darauf hin, dass die trade-off Effekte in neuronalen Unterschieden während der inzidenziellen Enkodierung begründet sind. Diese Unterschiede scheinen primär mit der Verarbeitung von sozialen Kognitionen zusammenzuhängen und verweisen auf die Bedeutung von öklogisch validen Experimenten.
Kurzfassung auf Englisch:
In our studies we aimed to replicate enhancing effects negative emotion has on memory for an experience per se, for plot-relevant information and for central details, as well, as for diminishing effects negative emotion has on memory for details irrelevant to the plot of picture stories (cf., Christianson, 1992; Kensinger, 2009; Levine & Edelstein, 2009; Mather & Sutherland, 2011; Reisberg & Heuer, 2004). We furthermore aimed to investigate the role of attentional processing and the neural bases of these effects. Not many previous experimental studies have used narrative stimuli to explicitly test effects of negative emotion on memory for different kind of information (Adolphs et al., 2005; Burke et al., 1992; Christianson, 1984; Christianson & Loftus, 1987; Heuer & Reisberg, 1990; Laney, Campbell, Heuer & Reisberg, 2004; Loftus et al., 1987; Wessel, van der Kooy & Merckelbach, 2000) and some of these studies did not confirm both effects (Laney et al., 2004; Loftus et al., 1987; Wessel et al., 2000). Moreover, the neural basis of these effects is largely unexplored especially regarding ecologically more complex stimuli with a narrative structure. We developed a set of divers, thematically driven picture stories with relevant and irrelevant test details appearing naturally within these stimuli and analyzed affective ratings, physiological responses, eye movements, recognition memory, free recall memory and neural activity of incidental encoding these picture stories, as well, as neural activity during the recognition memory test.
13 Negative and 13 neutral picture stories differed distinctly in participants' ratings regarding valence and arousal and explicit affective ratings corresponded to pronounced differences in neural activity. More activity for negative picture stories was found in emotion processing areas and we found increased allocation of resources to areas known to be involved in bottom-up and top-down driven allocation of attention, processing of faces and bodies, interpreting goal-directed movements, action observation and in areas known to be involved in processing higher order social cognitions, like empathy, theory of mind and moral judgments. In light of the notion, that narrative coherent structures of experiences and especially the dimension "meaning" strongly predict whether a real-life experience will be remembered after many years (cf., Brown & Kulik, 1977; Kızılöz & Tekcan, 2013; Peterson et al., 2014; Reese et al., 2011), our results in neural activity of viewing picture stories suggest that the differences substantially reflect enhanced allocation of resources "to make sense" of given visual input with regard to those aspects or dimensions. Affective ratings and the complex differences in neural processing of social information were correspondingly related to findings in free recall data. We could show that stimuli per se were more probably remembered after a retention delay of one day, when they depicted emotionally negative narrations. Free recall memory data moreover revealed that negative picture stories were reported with more information regarding persons and their actions but less information were reported regarding "any other" information. Additionally memories of negative picture stories were rated to be more detailed and more vivid.
We furthermore investigated differential effects of negative emotion on memory for specific details of differing relevance and the assumed mediative role of attentional processing. By measuring eye-movements we could analyze the impact of relevance on attentinoal processing of narrative stimuli and we could show that the above described differences between viewing negative vs. neutral picture stories (in affective ratings, neural processing and free recall memory) were related to emotion specific pattern in attentional processing of stimuli. Participants spent more than 30% of the duration looking at the persons of the picture stories but participants showed more proportional viewing time of people from negative stories. However, this difference amounted only to about 4% of the total proportional viewing time. Interesting to memory for specific details, attentional processing of relevant vs. irrelevant objects varied as a function of centrality and of emotion × centrality. Objects of central relevance were fixated for much longer, than comparable objects that were not relevant for the stories. About 20% of proportional viewing time was given to attentional processing of the most important object of the picture stories, while only 5% of the duration was given to the irrelevant objects. Emotional context was additionally affecting processing of central items. While 24% of the duration was given to the relevant objects when they were part of a neutral picture-story (e.g., a cup, an apple, a tennis-racket), were plot-relevant item of negative stories (e.g., a belt, a set of keys, a skate- roller) attended to for a much shorter duration (16%). But, although central relevant items of negative contexts were less looked at, they were nevertheless remembered equally well in the recognition task. And with smaller effects, we found that although peripheral objects of negative and neutral stories were attended to for about the same time, negative peripheral objects were remembered worse. Thus, we found that the pattern of results regarding attentional processing did not correspond to the pattern of results in recognition memory. Trail based hierarchical regression analyses additionally suggested that the amount of attentional processing was not mediating the differential effects of negative emotion on memory. Regression analyses instead suggested the following effects of negative emotion and of relevance on the relationship of attention and later memory of naturally embedded objects derived from visual narratives: the role of attentional processing for later memory depends on relevance and on emotional context but not on their interaction. Memory for non-human but very relevant information seems to depend less on attentional processing at incidental encoding compared to more attention- dependent memory regarding irrelevant information. And driven by effects of negative emotion, recognition memory of relevant objects is ensured well and recognition memory of irrelevant objects is diminished, both more independent of proportional viewing time at incidental encoding. These assumptions about the role of attention for memory as a function of relevance and of emotion were additionally supported by corresponding differences in neural processing at retrieval. We obtained significantly enhanced bilateral hippocampal activity at retrieval related to more attention-independent successful incidental encoding of details derived from negative picture stories and related to attentionally more dependent successful incidental encoding of details derived from neutral picture stories.
That attention is driving the effects of negative emotion on memory for relevant vs. irrelevant information has been a prominent explanation in discussions about differential effects of emotion on memory and was initially assumed regarding findings on eyewitness testimonies and on the so called “weapon-focus” effect (Loftus et al., 1987). Research on sensory processing as an interaction of the environment and the observer has accumulated detailed understanding about how we process information given limited capacity to deal with the amount of permanently emerging information. Attention is a primary characteristic within the competitive process to select the most relevant information at any point in time (Desimone & Duncan, 1995). This results in the familiar experience of daily life, dealing with information in a serial nature, consciously processing only a restricted fraction of available information. In attention research the serial nature of visual scene analysis is often likened by a “spotlight” metaphor (A. Treisman G. & Gelade, 1980). This behaviorally crucial role of attention to selectively allocate cognitive resources to be consciously accessed is generally understood to be driven by a top-down system reflecting executive control over deployment of attention and a bottom-up system compressing and filtering sensory information for saliency (Itti & Koch, 2001). Salient stimuli are infrequent or of instinctive or learned importance. Within this framework an understanding has been established regarding the role of emotion in biasing the processing of incoming sensory information (e.g., Pourtois, Schettino, & Vuilleumier, 2013). Based on more specific research on the differential effects of negative emotion regarding memory for details, some theories (Kensinger, 2009; Levine & Edelstein, 2009; Mather & Sutherland, 2011) have been proposed to integrate findings of enhanced memory for central details and reduced memory for peripheral information, when the context of incidental encoding is negative. In the ABC theory, a framework about "arousal-biased competition in perception and memory", Mather & Sutherland (2011) explain divergent findings of enhanced memory for central details and reduced memory for peripheral information into fundamental assumptions of basic research. Arousal is understood as a factor increasing both, top-down and bottom-up processes to selectively allocate cognitive resources to (internal and external) information of relevance. Thus arousal affects the competition between different information by increased processing of high priority information and decreased processing of low priority information. Importantly however, distinctions have to be drawn between attentional processes and further processes when explaining arousal related increased prioritizing of (relevant vs. irrelevant) information and the effects on subsequent memory. We could show that neural processes linked to processing information of social relevance – and not proportional viewing time – were involved in arousal induced increased prioritization regarding later memory. Results of previous brain lesion studies (Adolphs et al., 2005; Adolphs et al., 2001) suggested the involvement of the amygdala in driving the trade-off effects of negative emotion on memory for details of picture-stories. Imaging studies using more artificial stimuli also suggested differences in neural processes of the amygdala and furthermore differences in neural processes related to regions for focused attention and visual processing (Kensinger et al., 2007; Kensinger & Schacter, 2006; Mather et al., 2006; Waring & Kensinger, 2011). Experimental research may consider more the ecological validity of findings on emotional memory by inducing experiences based on stimuli where social information is embedded in a complex but coherently related narrative structure. With reference to findings based on narrative stimuli, Mather & Sutherland (2011) point out, that events which "have an overarching theme or narrative, the gist or schema is likely to be more salient and have more relevance than the details" (p. 124), thus strengthening representation of those information which are more related to a narration including enhanced memory for details of central relevance and reduced memory for peripheral information. In light of these assumptions, it seems essential to determine what the relevance of stimuli is and to consider differences between controlled stimuli and real-life events (cf., Levine & Edelstein, 2009). Our findings in brain imaging data related to processing negative vs. neutral, thematically driven picture stories suggest, that ecologically valid reasons for the differential effects of negative emotion on memory for relevant vs. irrelevant details are rather to be found in enhanced allocation of resources to process higher-order social cognitions.