Australasian Science: Australia's authority on science since 1938

Brilliant Memories for Dark Places

haunted house

By Oliver Baumann

We are usually not aware of it, but emotions exert a powerful influence over our memories by playing a key role in determining what we remember and what we forget.

Did you ever wonder why you can remember some places vividly while others seem fuzzy and fade with time? The answer could lie in your emotions.

Most of the episodes we experience in life are tinged to some degree by our emotional responses to them. We experience events that bring us joy, sadness, or make us afraid. Several lines of research now indicate that emotions are key players in controlling the accuracy, vividness and longevity of our memories.

In 1977, Brown and Kulik published their seminal paper on “flashbulb memory”, a term referring to the highly vivid and detailed memories of the moment and circumstances that can result from experiencing or hearing about an emotionally loaded event. For example, people typically have vivid memories of what they were doing when they heard about President Kennedy’s assassination or the 9/11 terrorist attacks on the World Trade Center.

In very extreme cases, traumatic emotional events can even cause a phenomenon that is called hypermnesia. Hypermnesic memories are exceptionally exact, vivid and strong, and are associated with certain mental disorders, such as post-traumatic stress disorder.

It is thought that extreme levels of emotional and physiological arousal during traumatic events lead to these abnormally complete and vivid memories. The more intense the emotion associated with an event, the more vivid the recollection.

Memories forged in the presence of strong emotions also seem to last longer. Research suggests that emotions slow the process of forgetting, so that emotional events might eventually be forgotten as well, but at a rate that is slower than for neutral episodes.

But the relationship between emotions and memory is very complex. Several lines of research suggest that universal memory enhancement is not the best description for how emotional arousal influences memory. Rather, it often leads to memory trade-offs, enhancing memory for select features of an event while impairing memory for other aspects.

One example is the “weapon focus effect”, where witnesses of a crime can give detailed descriptions of the weapon used but cannot remember any other details of the crime scene. These trade-offs have mostly been described for negative events but also appear for positive events, such as when a cheerful conversation while driving leads to partial amnesia for the scenery of the journey.

For the past few years I have been investigating the cognitive and neural mechanisms underlying people’s memory for places. Spatial memories are indispensable for finding our way in complex environments, planning routes to distant locations and even for simple tasks such as finding our way back home. While we know quite a bit about how the human spatial memory system works, we know very little about its interactions with emotions.

We therefore conducted an experiment to explore how emotions affect our memory for places. We not only wanted to investigate how emotions affect our ability to recall places, we also wanted to determine how emotions affect the brain systems that underlie our ability to remember places.

The basic design of our study was to test university students’ memory for places that had been previously paired with emotional events, while simultaneously recording the brain activity using functional magnetic resonance imaging (fMRI). We used a computer-simulated virtual environment for our study because we wanted to capture as much as possible of the true dynamism of real navigation while maintaining a high degree of control over the environment. The “virtual house” we created contained five distinct rooms that were all connected by a central, circular hall. The rooms included a kitchen, a living room, a bedroom, a bathroom and a study (Fig. 1).

Figure 1. A view experienced by participants when they entered a room in the vir
Figure 1. A view experienced by participants when they entered a room in the virtual house.

Our experiment consisted of an initial learning phase and a retrieval phase conducted a day later while we recorded the participants’ neural activity. In the learning phase we let the participants navigate through the different rooms of the virtual house and asked them to complete an object-location learning task. To do this we showed the participants simple drawings of everyday items, such as a trumpet or a coffee mug, and told them they had to discover and remember the room in which the object was hidden. Every room contained four objects.

Critically, there was no natural connection between the objects and the rooms. For example, the trumpet could be in the bathroom and the coffee mug in the bedroom. Further, the objects were hidden in the rooms so that they were not immediately visible upon entering. It sounds rather difficult, but in fact it took participants just half an hour or so to find and remember exactly which items in a set of 20 belonged in the five rooms.

To explore the effects of emotions on memory for places, we included another critical manipulation. Every time the participants entered a room we flashed a photograph that depicted either a pleasant, unpleasant or neutral event (Fig. 2). The pleasant images were pictures of couples, happy families and scenes of sport and adventure, whereas the unpleasant images included scenes of attack and threat, such as people brandishing guns or angry dogs. For the neutral events we used images of neutral-faced men and women as well as household items.

Figure 2. Example of an unpleasant image shown during the navigation task.
Figure 2. Example of an unpleasant image shown during the navigation task.

Three of the rooms were consistently coupled with pleasant, unpleasant or neutral events, whereas in the two other rooms a mix of neutral and either pleasant or unpleasant images were shown. We did this so that we could determine whether the likelihood with which an emotional event occurred would affect participants’ memories for the rooms.

On the next day, participants came back to have their brain activity measured using fMRI. While they were in the scanner we showed them static images from the virtual house they had seen the day before, and asked them to remember whether an object belonged in the room shown. We found that participants were much quicker to perform this task when the room they saw had previously been paired with either pleasant or unpleasant images rather than neutral images. This confirmed that emotional events during the learning phase had improved their memory.

Interestingly, this effect only occurred for the rooms that had been consistently paired with either positive or negative emotions. In fact, memory performance for rooms in which emotional events occurred only intermittently was no better than for the room in which only neutral images were seen. In other words, only rooms in which participants had been certain of encountering an emotional event were recalled faster.

Next we asked whether the emotional events also affected the brain systems underlying our memory for places. We predicted that if emotions affect memory for places, we should find differences in activity in brain areas that encode spatial scenes when participants viewed the different rooms from the virtual house. Indeed, we found that memory-related activity in the para­hippocampal gyrus – an important region for scene memory – was significantly enhanced when participants viewed images of rooms in which consistently unpleasant events had been encountered (Fig. 3).

Figure 3. Memory-related brain activity in the left and right parahippocampal gy
Figure 3. Memory-related brain activity in the left and right parahippocampal gyrus for rooms associated with unpleasant emotional events.

Interestingly, none of the participants realised that the rooms had been systematically paired with different emotional events. This suggests that memory enhancement was elicited automatically and did not require awareness on the part of the navigator.

Interestingly, we found no memory-enhancing effect for pleasant images observed during learning. This suggests there must be something special about unpleasant events when it comes to memory for places.

We believe the key to this question might lie in an almond-shaped brain region called the amygdala, which is located just adjacent to the parahippocampal gyrus and is crucial for the experience of fear. Back in 1888 it was reported that lesions of the amygdala in monkeys led to social and emotional deficits. But the most striking deficit was a complete loss of fear.

Later experiments in rats expanded on these findings by showing that the amygdala is also crucially involved in fear learning. For example, rats with lesions in their amygdala are unable to learn to avoid a location in their environment in which they receive electric foot shocks, even though they perceive the painful stimulus normally. It therefore appears that pain alone is not enough to mediate the formation of place avoidance memories, and that it is necessary for animals to experience fear via the amygdala for this to occur.

In our study in human volunteers, we failed to observe any amygdala activity. However, we only measured neural activity during the retrieval phase, after place learning was complete. It might therefore be the case that fear-related activity in the amygdala leads to a strengthening of place memories during learning, but that these memories are stored in the parahippocampal gyrus independent of any amygdala activity. In future experiments we aim to record brain activity during the learning phase as well to measure interactions between emotion-related and memory-related brain regions during the formation of place memories.

The famous philosopher William James once said that some impressions are so emotionally exciting as to leave a “scar upon the cerebral tissue”. And Sigmund Freud remarked that emotions affect our memories without us even being aware of it. Our recent work has shown that through the methods of cognitive neuroscience we can gather objective insights into how emotions affect our memories, without having to resort to introspection.

Our findings reveal that memories are stronger for places consistently associated with unpleasant events. This suggests that during navigation we are guided not only by visual landmarks but also by our emotions. In fact, threat-induced enhancement of spatial memory might constitute an important survival mechanism for avoiding future threats.

Better knowledge of the mechanisms underlying the interaction between emotions and memory might even lead to the development of treatments for patients with memory disorders by artificially stimulating the memory-enhancing neural processes that are associated with experiencing negative emotions.

Oliver Baumann is a Research Fellow in the Queensland Brain Institute at the University of Queensland.