Memory for emotional events: The

The emotional intensity of an event is a significant predictor for vividness of event memory. Nevertheless, during the last few decades, there has been some confusion in literature as to whether emotional events are poorly or well retained. It is important to consider that not all details of emotional events are equally remembered: Memory for the central details seems to be relatively good, whereas memory for peripheral details appears to be relatively poor. The aim of the present study was to investigate the accuracy of central vs. peripheral details of an emotional event in a natural but controlled context: the emotional event is a simulated life event, the central and peripheral details of the emotional event were controlled. Indeed previous research work was simply based on the induction of an emotional state in an experimental context and subsequent assessment of a performance memory task. Results showed that, following an emotional event, individuals provided a vivid and accurate recollection not only of the central gist of the event, but also of the context and peripheral details. Implications for literature on emotional autobiographical memories were discussed.

unknown.A high memory performance for emotionally arousing events has been demonstrated (Deffenbacher, 1983;Deffenbacher, Bornstein, Penrod, & McGorty, 2004).A number of studies have indicated that the emotional intensity of an event is a significant predictor for how vividly the event is recalled (e.g.Reisberg & Heuer, Memory for emotional events 324 2004; Rubin & Kozin, 1984;Talarico, Bernsten, & Rubin, 2009).What these studies have shown us is that emotional events are remembered with great richness of details, although no information has been provided on the accuracy of these remembered details.Indeed, during the last few decades, there has been some confusion in literature as to whether emotional events are poorly or well retained.Whereas laboratory research has shown mixed results concerning memory for emotional events (Christianson, 1984;Heuer & Reisberg, 1990;Deffenbacher, 1983;Loftus & Burns, 1982), research on real-world events has usually demonstrated that details of emotional events are relatively well retained in memory (Bohannon, 1988;Brown & kulik, 1977;Rubin & Kozin, 1984).For a clear investigation of the link between memory and emotion, it is important to consider that not all details of emotional events are remembered equally well.More specifically, it is thought that memory for the gist (central details) of an emotional event is well retained, whereas memory for irrelevant information (peripheral details) is poorly recalled (see, for overviews, Christianson, 1992;Heuer & Reisberg, 1992).Some studies have found that, whereas memory for peripheral details seems to be diminished by high levels of arousal, memory for central details (emotion-related and plot-relevant details) appears to be facilitated (e.g., Burke, Heuer, & Reisberg, 1992;Christianson & Loftus, 1991;Christianson, Loftus, Hoffman, & Loftus, 1991;see Christianson, 1992, for a review).The idea that high arousal and negative affect enhance recall of central aspects of events has been supported in several research fields: autobiographical memory (Berntsen, 2002;Christianson & Loftus, 1990;Strube & Neubauer, 1988;Wessel & Merckelbach, 1994), eyewitness memory (Steblay, 1992;Yuille & Cutshall, 1986), event memory (Christianson & Loftus, 1991;Reisberg & Heuer, 2004), episodic memory (Kensinger & Corkin, 2003;MacKay et al., 2004), animal learning (Easterbrook, 1959) and perception (O¨hman, Flykt, & Esteves, 2001).However, a number of laboratory studies on episodic memory have recently shown that also contextual and peripheral information are enhanced for emotional stimuli, such as colour (Kensinger & Corkin, 2003;MacKay et al., 2004), spatial location (MacKay & Ahmetzanov, 2005), or temporal context (D'Argembeau & van der Linden, 2005).
What exactly are "central" or "peripheral" details?Where is the boundary between the center of an event and its periphery?Burke Heuer, and Reisberg (1992) proposed the categorization of to-be-remembered material in four categories of information.The first two are commonly considered as central details and contain a) details pertaining to the gist or plot of the emotional event and b) the materials visually central to the event.The second two categories are designed to categorize what have been considered peripheral details and include c) details that are attached to the visually central materials and d) details from the background and the context of the event.An alternative to Burke and colleagues' categorization (Burke, Heuer, & Reisberg, 1992) was proposed by Christianson and Loftus (1991).
According to the authors, central characters could be identified in terms of their centrality to the subject's attention, rather than relevance to the plot.In this definition, "central" details would be those details associated with material central to attention, independent of whether they are also associated with material central to the event's plot (Christianson & Loftus, 1991).For the purpose of the present study, we considered the Christianson and Loftus's categorization: The centrality of details refers to the centrality for the subject's attention, rather than to the relevance to the plot.A common explanation of the link between emotion and the type of remembered details is based on the attentional narrowing hypothesis (Christianson, 1992), according to which physiological arousal results in attention being directed towards central rather than peripheral characteristics of the situation.Consequently, memory for central details would be relatively good, whereas memory for peripheral information would be impaired.There is empirical evidence to support Christianson's attentional narrowing hypothesis.Field studies examining the memory of witnesses for robberies (Christianson & Hübinette, 1993), or that of college students for emotional events (Christianson & Loftus, 1990;Wessel & Merckelbach, 1994) suggest that central information is, indeed, better remembered than peripheral information.
However, a limit of this attentional narrowing hypothesis concerns the ecological validity of the procedures employed to assess emotional memory.A general criticism of laboratory studies on emotion and memory is that they are too far removed from real life and that generalising experimental data to emotional life events (e.g.trauma) is not acceptable (e.g.Terr, 1994;Yuille & Tollestrup, 1992).In other words, this implies that emotional memory can only be examined in field studies.However, on the other hand, field studies may suffer from problems such as retrospective bias due to post-encoding categorisation of central and peripheral information (Christianson & Loftus, 1990;Wessel & Merckelbach, 1994).In sum, field studies on emotional memory may suffer from retrospective and report biases, whereas experimental studies may not be generalized to real-life situations.Because these problems can be circumvented in controlled experiments, it seems worthwhile to find approaches that increase the ecological validity of laboratory studies on emotion and memory.

Overview and hypotheses
As can be seen from the aforementioned, empirical research concerning the accuracy of emotional memory details has reached controversial and debatable results.The aim of the present study was to investigate the accuracy of remembered central and peripheral details of an emotional event; more specifically we considered memory for the gist of the event itself as memory for the central information, and memory for the surrounding context as memory for peripheral details (Brown & Kulik, 1997).The strength of the present study is to explore the accuracy of emotional memory details in a natural but controlled context, whereas previous reserch work has mainly focused on the induction of an emotional state in an experimental setting and, then on the assessment of a performance memory tasks (Reisberg & Huer, 1992;Christianson, 1992).The novelty of this research is the attempt to overcome, at the same time, the limit of the experimental studies -by choosing an emotional life event -and the limit of the field studies -by adopting a controlled procedure -.Indeed, in the present study, the emotional event is a real-life emotional event simulated with the help of an accomplice, and central details of the emotional event and peripheral details of the context of the event were controlled.
The main characteristics of this study were the following: a) it concerned a negative emotional event that had been simulated with the help of an accomplice; b) participants' memories for both central and peripheral details were collected (Burke, Heuer, & Reisberg, 1992;Christianson & Loftus, 1991;Reisberg & Heuer, 1992); c) participants' memories for the emotional event were assessed in both a free and probed recall procedure (Bohannon, 1988;Bohannon & Symons, 1992;Brown & Kulik, 1977), and d) participants' memories for the emotional event were assessed both immediately after the event (Time 1), and a 5-month delay (Time 2) (Christianson, 1984;Levonian, 1967).
Participants were expected to remember not only the gist of the emotional event (central details) but also the contextual details (peripheral details) (Brown and Kulik, 1977;D'Argembeau & van der Linden, 2005;Kensinger & Corkin, 2003;MacKay & Ahmetzanov, 2005;MacKay et al., 2004).Both central and peripheral details would be subject to the effect of time (Heuer and Reisberg, 1992).In line with the emotional memory studies, peripheral details were expected to be better stored through probed recall than free recall, since individuals remember more accurately contextual details when they are asked to do this (Bohannon & Symons, 1992;Nachson & Zelig, 2003).

Method Design
The present study used a 2x2 repeated-measure design with Retention Interval (Time 1 vs.Time 2) and Type of Details (Central vs. Peripheral Details) as within subjects factors.Dependent variables were Accuracy of remembered details score in the free recall task, and Accuracy of remembered details score in the probed recall task.

Measures and Procedure
Students were recruited during a psychology class and requested to write a diary and answer several questions for seven days as soon as an emotional event (positive or negative) had happened in their life in the days following the delivery of the diary (cover story).Immediately after this delivery, a state of emotional stress and alarm was simulated in the class: An accomplice suddenly came into the room, shouting nonsensically and moving around the class for a few minutes provoking panic among the students.As soon as he went out, all students agreed to write in their diaries about this negative event.This situation was planned in order to have a standard stimulus event to which all students might have responded.The event has been chosen on the basis of features of displeasure, surprise, personal importance, and negative emotional intensity, as so as a real life event.In order to check the accuracy of peripheral details, many details of the context were controlled: the position of the teacher and her assistants, their clothing, the position of particular objects in the classroom, and a specific sound.
The questionnaire used for the present study was composed of several sections: 1) Free recall of the event, 2) Probed recall for central details, 3) Probed recall for peripheral details, 4) Emotional feeling state, 5) Novelty, 6) Importance.
Free recall.Following the simulated emotional event, participants were requested to write a diary for seven days about it.The written accounts were submitted to a content analysis in order to estimate the occurrence frequency of remembered central vs. peripheral details.The content analysis entailed independent coding of narratives by two different judges (95% of inter-judge agreement).Details referring to the protagonist of the event were considered as central: a) his physical  Type of Details (Central vs. Peripheral) as within subjects factors.A significant main effect of Retention Interval was found to be significant (F1,94= 322.13, p < .001), in that the accuracy of remembered details declined over time (see figure 2).A significant main effect of Type of Details was also found to be significant (F1,94= 852.16, p < .001),with peripheral details more accurately remembered than central details (see figure 2).Finally, a significant interaction effect of Retention Interval by Type of Details was found on the number of remembered details (F1,93= 150.06, p < .001)since memory decline was found to be more evident on the average amount of recalled peripheral details, compared to central details (see figure 2).
characteristics, b) his clothing, c) what he did, and d) what he said.The proportion given by the number of remembered central details divided by the number of total remembered details was considered in data analysis as the Accuracy for central details score in the free recall task scale, ranging from 0 to 1.

Figure 1 :
Figure 1: ANOVA Retention Interval by Type of Details on the Accuracy of remembered details score in the free recall task.

Figure 2 :
Figure 2: ANOVA Retention Interval by Type of Details on the Accuracy of remembered details score in the probed recall task.