Age-related differences in the phenomenal characteristics of long-term memories of March 11, 2004 terrorist attack

There is a great debate about whether the formation of FBMs follows different processes than the formation of other types of memories. Tinti, Schmidt, Testa, and Levine (2014) conducted a study to try to delve into this topic. They proposed to test two hypotheses: a) that event memory and FBM have different determinants and b) that event memory is not necessarily a direct causal determinant of FBM. For this matter, they took as reference the memory of an important and positive event, the Italian victory in the 2006 World Cup Football Championship. The results showed significant differences in both types of memory, so they concluded that each type of memory is originated following different processes. This finding also suggests that the processes involved in maintaining memory also differ: event memory was maintained by mere exposure to the media, whereas FBMs could only be maintained after thinking and discussing with others about the event experienced. Therefore, these authors showed that both types of memory did not have to be related. Furthermore, according to their model, the only direct determinant in maintaining FBMs was the recovery of the personal circumstances surrounding the event, while maintaining the event memory depended on prior knowledge and exposure to the media.

In contrast, Kraha, Talarico, and Boals (2014) , taking as reference a positive event for the Americans (the murder of Osama Bin Laden), carried out a study whose results showed that the memory of a positive event is not so vivid, nor is it remembered with as much confidence as a negative event. According to these authors, these results provide further evidence against the existence of a special memory mechanism involved in the formation of FBMs.

Also, Curci et al. (2015) , taking as reference the resignation of Pope Benedict XVI, d the memory of three groups with different levels of religious involvement (practicing Catholics, non-practicing Catholics and Evangelists) and concluded that there was a significant difference in the processes that are activated when a person is asked to remember the original news or the context in which he/she received the news.

Memory of a complex event is affected by many variables. These variables can be grouped on the basis of their influence on information encoding, retention, and recovery. Two important aspects of autobiographical memory are the qualitative (or phenomenal) aspects of memory and the subjective point of view that leads to remembering ( Argembeau, Comblain, & Van der Linden, 2003; Larsen, 1998 ). The phenomenal characteristics (amount of sensory detail or clarity of location, for example) give the individual the feeling that a particular mental representation is a memory of a real event, as opposed to an imagined event or any other kind of mental representation ( Johnson, Hashtroudi, & Lindsay, 1993 ). In addition, the viewpoint of recall ( Crawley & French, 2005; Robinson & Swanson, 1993 ) is one of the characteristics that indicate the degree of involvement of the subject in the event, as well as the deterioration of the quality of the memory: greater involvement and time delay tend to produce memories from an observer perspective rather than from a field perspective.

To examine the phenomenal characteristics of emotional memories, several studies have used an array of questionnaires that enable specific memory features to be explored (for a review of these studies, see Manzanero & Recio, 2012 ). One of the earliest studies looked that has at the phenomenal characteristics of memories of negative events ( Tromp, Koss, Figueredo, & Tharan, 1995 ) concluded that, compared to other types of memory, these memories were less clear and vivid, less well remembered, thought about and talked about more, and had less visual detail. In contrast, another study ( Byrne, Hyman, & Scott, 2001 ) found that memories of negative events differed only in the sense that they were reported as having less sensory information.

In any case, the characteristics of memories of traumatic events may depend on cultural and/or resilience factors, as shown in a recent study assessing positive and negative memories amongst the people of the Gaza Strip ( Manzanero, López, Aróztegui, & El-Astal, 2015 ). Thus, the same negative event could be remembered in a vivid and consistent manner or, conversely, in a fragmented manner or perhaps not at all.

Only a few studies have examined the phenomenology of negative memories in children and adolescents. The few existing related to phenomenal qualities and ageing have resulted in mixed findings: some of them have found age-related effects in the vividness or specificity of the details remembered ( Denver, Lane, & Cherry, 2010; Kensinger, Krendl, & Corkin, 2006 ) and others have not found any age-related differences ( Conway et al., 2009; Kvavilashvili et al., 2010 ).

In a study of the developmental aspects of FBMs, Winograd and Killinger (1983) asked 338 high school and college students to state their personal memories surrounding 7 major events. These subjects were 1-7 years old at the time of the event and, for the most well-known events, recall was shown to improve steadily with increasing age. Winograd and Killinger argued that the differences found between the college students and the high school students could be attributed to their neurological immaturity or their inability either to perceive the importance of the event or to be surprised by it.

Warren and Swartwood (1992) found that, among children who were in kindergarten through eighth grade at the time of the space shuttle Challenger explosion, only those who reported a high emotional response to that event could remember a lot about it two years later. Using data gathered at two weeks, two months, and two years and taking all the children into account, with the varied intensity of their emotional responses, they found that a great deal had been forgotten or distorted over time.

On the other hand, there is some variability in the phenomenology of children's memories of negative events, as shown in some studies of sexually abused children ( Burgess, Hartman, & Baker, 1995 ) and of young children who had experienced an earthquake ( Azarian, Lipsitt, Miller, & Skriptchenko-Gregorian, 1999 ). Kensinger, O’Brien, Swanberg, Garoff-Eaton, and Schacter (2007) suggested that there is reason to believe that young adults retain more vivid and detailed memories of a negative experience than of a positive experience. Other studies have found that older adults tend to process information with a particular focus on their affective response to the information (Comblain, D’Argembeau, & Van der Linden, 2005), which could favor the preservation of memory with emotional relevance, specifically.

Cohen, Conway, and Maylor (1994) and Kensinger et al. (2006) found significant age effects for memories of the resignation of Margaret Thatcher and of the space shuttle Columbia explosion, respectively.

The present work was designed with the objective of exploring age-related effects on the phenomenal characteristics of long-term memories. It was hypothesized that people who at the time of occurrence of the events were of an age that did not allow them to consider the transcendence of the event would generate a memory of that event that would not comply with the characteristics of a FBM.

Table 1 shows mean scores for each memory variable as a of age. A t -test (2-tailed) was conducted to assess the effects of age on each memory characteristic. This analysis revealed a number of significant differences in the dependent variables d, for many of which the scores obtained by the adult group were higher than the scores obtained by the young group. These findings are consistent with the idea that there is a relationship between an individual's memory of an event and his/her age when the event took place. The results that the characteristics of memories for flashbulb events vary depending on the individual's age at the time of the event.

The variables that appeared to be affected were the following:

Memory quality: definition, vividness, detail, sensory and contextual information, fragmentation, comprehension, complexity, doubts, and recall perspective were evaluated in studying this variable. Results that young memories of the attacks had less definition, less vividness, fewer details, less fragmentation, less visual, auditory and odor information, and were less complex and understandable. They also reported a shorter duration for the event and had worse memory of the place where the event happened and the exact year and hour when it occurred. When shift in perspective is considered, adults tend to remember the event in an observed perspective more than younger people. No effects on sensorial characteristics about color, haptic, and taste were found because of floor effects. The only variable in which the young group has obtained a higher score is in doubts about their own memory.

Associated emotion: in evaluating associated emotions, the variables taken into account were implications, valence, intensity of feelings, relevance, feelings during the event and now, and thoughts at the time of the event. The young group gave the same implications to the event as the adults but had less negative valence, intense feelings, and less relevance associated with the event. Compared to the adult group, they also tended to have less memory of their thoughts and feelings at the moment the event took place, as well as at the present time.

Accessibility: data showed that the young group would tended to talk and think about the event on fewer occasions and made less effort at retrieval. Also, young people remembered worse the related events that took place before and after the attacks. No differences were found between the two groups regarding problems to talk about the event.

HDV Graphs

HDV graphs facilitate the visualization of complex data. This technique displays all the data at once, enabling researchers to graphically explore in search of data distribution patterns ( Manzanero, El-Astal, & Aróztegui, 2009; Manzanero, López, & Aróztegui, 2016 ). The graphs are similar to scatter plots. The different variables corresponding to a subject's responses on questionnaire items are represented as a point in a high-dimensional space. When there are more than three variables, as in this study, mathematical dimensionality reduction techniques are used to build a 3D graph ( Buja et al., 2008; Cox & Cox, 2001 ). Each point in the hyperspace has a distance to each of the other points. Multidimensional scaling will search 3D points, preserving the distances between points as much as possible ( Barton & Valdés, 2008). Sammon's error (Barton & Valdés, 2008 ) is used to calculate the 3D transformation error.

3D points are represented using Virtual Reality Modeling Language (VRML). VRML files allow graphical rotation and exploration to facilitate graphical data analysis. 3D graphs permit visual exploration of the data in search of its distribution patterns.

Figure 1 shows two different views of the graph obtained using the 3D points, which were made from points for the 24 dimensions (in the mathematical sense) corresponding to those variables measured where significant differences were shown. The complete graph shows 3,749 data points in compressed form. Sammon's error was 0.12, suggesting that it should be interpreted with caution.

Figure 1

HDV graph of variables showing significant statistical differences.

Upon exploration of the graph, the following may be observed:

• 1.

  There is considerable overlap of the data for the two groups.

• 2.

  Data for the adult group is displaced to the right of data for the young group.

• 3.

  The two groups a similar data distribution (cone-shaped with apex pointing to the right).

• 4.

  Data for the young group shows higher dispersion than data for the adult group.

The data overlap (observation 1) is consistent with theories presenting FBM development as a process that is continuous and similar between young people and older adults. The displacement to the right of data for the adult group (observation 2) is consistent with theories of FBM development as changing gradually with age. The similarity of data distribution (observation 3) is also consistent with theories presenting FBM development as a process that is continuous and similar between young and adults.

The variability of the data is observed and used to describe and test hypotheses (see Appendix B ). The reduction in data variability with increasing age (observation 4) could be due to a more reliable memory process in older adults. This is consistent with a theory of FBM development in which young adults gain in reliability, accuracy, and stability of recall as they grow toward adulthood. The ratio of variability obtained for variables in which significant differences were shown was rv = .0969. Expressed as a percentage, this means that, in comparison to the young adults, older adult performances were 9.69% less variable. The older adults appeared to be more reliable. On the other hand, the improvement in reliability, as seen through variability reduction, is significant but not extreme.

The purpose of this study was to explore age-related differences in the phenomenal characteristics of long-term memories of the terrorist attacks that took place in Madrid (Spain) on March 11, 2004.

Differences between the two groups (adults and youths) were found in terms of memory quality, emotions associated with the event, and accessibility of the information remembered. The adults group obtained higher scores than the younger group on the following variables: definition, vividness, amount of detail remembered, fragmentation of the retrieval event, comprehension, complexity, viewpoint of recall, visual information, sound, odor, where and when the event took place, duration, year and hour when event happened, valence, intensity of feelings, relevance, feelings during the event and now, associated thoughts, effort to recall, remembering related events that took place before and after the attacks, multiple retrieval, and problems to talk about the event. On the other hand, the young group obtained a higher score only in one variable: doubts about their memories of the event. As the HDV graphs show, there was less quantitative variability within the adult group (differences).

Many empirical studies have demonstrated the existing relationship between the amount of retrieval of a FBM and the capacity to provide a detailed and confident report, after a long time period ( Bohn & Bernsten, 2007; Conway et al., 1994; Curci & Luminet, 2006, 2009; Finkenauer et al., 1998; Talarico & Rubin, 2003, 2007; Tinti et al., 2014 ). For this reason, the great importance that retrieval has in maintaining this type of memories becomes obvious.

In a review by Gordon, Baker-Word, and Ornstein (2001) regarding past experiences in children, results showed that simple exposure to an event is not enough to give rise to a FBM. As age increases, some cognitive s experiment changes that affects the acquisition and maintenance of the information in memory systems. That is to say that what we already know can determine what we can remember or not. According to these authors, it seems that there are some factors that influence the representation strength in the memory: a person's involved level of active participation, age or level of development and time of exposure to the event. An individual's previous knowledge and nature of the event itself would therefore influence details of the event codification and inclusion in long term memory system. This would explain results found in the present study.

On the other hand, some studies have found that the way older adults tend to process information with a particularly focusing on their affective response to the information ( Comblain, D’Argembeau, & Van der Linden, 2005 ), specifically favors, the preservation of memory with emotional relevance. The emotional-integrative model of FBMs also suggests that the emotional impact and personal importance of the event are factors crucial to the encoding and retention of FBMs ( Luminet & Curci, 2009 ). The individual's emotional reaction encourages post-event elaboration and rehearsal, both of which serve to enhance recollection, vividness, and confidence in memory accuracy ( Finkenauer et al., 1998 ). According to the importance-driven model, proposed by Er (2003) , personal consequences will determine the intensity of the emotional reaction. The emotional arousal will be different depending on the consequences the person thinks the event could have. The greater the degree of importance and the greater the emotional reaction, the more detailed and accurate the memory. The study conducted by Kensinger et al. (2006) found that, in some instances, emotional content may benefit the memories of older adults more than those of young adults. Emotional responses may make older adults more likely to rehearse the central information of the event (event-related details), as well as the personal details of emotional events. An event's emotional salience appears to enhance older adults’ memories. In another study conducted by Tekcan and Peynircio?lu (2002) analyzing possible age-related effects in FBMs formation, these authors found a positive correlation between retrieved information about an event that took place long time ago and older persons: the older they were at the moment the event took place, the more information they provided. This difference could be due, as we pointed out in a previous paragraph, to the existing difference in the ability to measure consequences of the events between both age groups and the greater ignorance of the real importance of the event by the younger participants ( Winograd & Killinger, 1983).

Given that the extent of consequences could be the key determinant of the intensity of an individual's emotional reaction (the emotional arousal), in our opinion, it is more difficult to younger people to identify the significance of an event (an event of this complexity, at least), for they probably do not attribute the same personal relevance to it as adults do. The fact that the emotional content of an event enhances the strength of the memory formed may be the most likely reason why the children's group obtained lower scores for relevance, comprehension, vividness, and intensity of feelings. Future studies should address this by comparing events that evoke emotions of similar intensity in both children and adults.

This paper is part of a research project on assessment of memories and psychological disorders associated to trauma in refugees and victims of war, developed by UCM Research Group on Eyewitness Testimony (ref. 971672), in the framework of the project financed by Santander-Universidad Complutense de Madrid (PR26/16-20330).

The authors of this article declare no conflict of interest.