- <div style="background-image:url(/live/image/gid/32/width/1600/height/300/crop/1/41839_V14Cover_Lynch_Artwork.2.rev.1520229233.png)"/>
The Influence of Cues and Beacons on Route Memory
Lake Forest College
Lake Forest, Illinois 60045
*This author wrote the paper as a senior thesis under the direction of Dr. Matthew Kelley
Part-set cueing facilitation refers to the retrieval advantage when part of the set of to-be-remembered information is provided as cues at test. The present experiments were designed to explore part-set cueing in the context of map and route memory. In two experiments, we set out to find how cues and beacons would affect the recall of routes, as well as to see if presentation during the memorization phase changed the outcome in the recall task. The use of both beacons and part-set cues facilitates the recall of routes, but to varying degrees. Implications of these findings will be discussed.
If you were asked to participate in a memory experiment and were given the option to attempt the test either with hints or without hints, what would you choose? Most people would likely choose to receive hints because they seem intuitively helpful. They would assume that the hints will help cue their memories. Sometimes, however, receiving hints could inhibit rather than aid a person’s ability to remember.
Hints could come in many forms. If you were trying to remember the name of a famous artist and were given the initials P.P., you would probably guess that it was Pablo Picasso. If you were trying to remember a word describing things relating to the law and were given the hint that it rhymed with “beagle,” you would probably guess that the word was “legal.” If you were asked to recall a specific word from a list and were given, as hints, the words that appeared near that word on the list on which it appeared, you might think these cues would similarly be helpful. This last example example is known as the part-set cueing procedure, where part of a particular set of presented information is given as a cue during the recall phase of a memory experiment. In part-set cueing experiments, participants are often shown a list of words to remember. Then, during recall, they are given half of the words as cues and are asked to recall the remaining words. While most people would likely naturally assume that these cues would enhance memory, research has found that, counterintuitively, these cues often hurt memory— people tend to recall fewer words in the presence of part-set cues as compared to an “uncued” situation in which no cues were given (e.g., Dong, 1972; Roediger, Stellon, and Tulving, 1977, Slamecka, 1968). This is known as part-set cueing inhibition.
Part-Set Cueing Inhibition & Facilitation
Slamecka (1968) was the first to examine the effects of part-set cueing on memory. In one experiment, participants were shown lists of 30 items that had no categorical relationship (e.g., random words; no repeated categories) and later were given either a cued or uncued test. To his surprise, Slamecka observed part-set cueing inhibition in the experiment—participants recalled more list items in the uncued (no hints) condition than in the cued (hints) condition. In his next three experiments, Slamecka varied the number of part-set cues given in the tests (between 5 and 29 cues) and observed the inhibition in each case. In his last two experiments, he tested to see whether cues were inhibiting recall because the experimental subjects had to read through the cues before being told to recall (which delayed recall), whereas the controls were asked to recall right away. In these experiments, the controls were made to wait the amount of time that the experimental subjects took to scan through the cued words before being told to recall. Despite the equal delay, the cues still were not helpful.
Roediger et al. (1977) studied the inhibitory effect of part-set cueing in another way. They had three main conditions: list items as cues (the standard part-set cueing procedure), “extralist” items as cues (meaning providing items that were not on the original list as cues), and a third condition with half part-set cues and half extralist cues. During some trials, participants were asked to recall all list items after studying the cues, whereas on other trials, they were asked to recall only the non-cued items. Roediger et al. found that both part-set cues and extralist cues had an inhibitory effect on recall, and that inhibition was greater for part-set cues than for the extralist cues. Also, participants performed better when they were asked to recall all word items than when they were asked to recall only the non-cued items.
Part-set cueing inhibition has proven to be a robust and easily replicable phenomenon over the years (e.g., Nickerson, 1984; Basden & Basden, 1995; Marsh, Dolan, Balota, & Roediger, 2004; Fritz & Morris, 2015). Clearly, inhibition is common when using lists of random words, then followed by a recall test.
Conversely, in some cases, part-set cueing has been shown to facilitate rather than inhibit recall. Research has shown that facilitation seems to depend on the type of material and the type of test (e.g., Bellezza and Hartwell, 1981; Pollio and Gerow, 1968; Tulving and Pearlstone, 1966). Serra and Nairne (2000) and Cole, Reysen, and Kelley (2013) observed part-set cueing facilitation in recent memory studies.
Serra and Nairne (2000) explored the effects of part-set cueing on order information across three experiments. All three experiments used reconstruction of order tests, which required participants to put scrambled lists of items back into its original order (positions) of presentation. In the first experiment, participants were given an eight-word list to memorize. On cued trials, four part-set cue words were put into their appropriate positions on the test sheet and participants had to reposition the remaining four items. On uncued trials, four positions were removed (replaced with Xs) and participants had to reposition the remaining four items. Serra and Nairne found that the cues helped participants remember the order of the uncued items; in other words, they found part-set cueing facilitation. They suggested that participants likely formed inter-item associations during presentation so that, at test, the cues were helpful because having one link in a memory chain was helpful in trying to recall other links in that chain (i.e., associative links). The next two experiments replicated part-set cueing facilitation in order memory. Experiment three also found that, if the cues are inconsistent with the original presentation order (e.g., item 1 shown in position three at test), the cues were detrimental to the subjects in the order test. In other words, part-set cueing inhibition occurred with inconsistent cues because they were not congruent with the “chaining” strategy.
Cole, Reysen, and Kelley (2013) also showed part-set cueing facilitation in their experiments on spatial memory. In their first experiment, they explored how part-set cues influenced their participants’ ability to reconstruct a snap circuit object. The snap circuit object contained 10 bottom pieces and 10 top pieces. All the participants viewed the completed circuit twice. On cued trials, participants received a board with five of the bottom pieces already placed in their proper locations and were asked to reconstruct the circuit. On uncued trials, they received a blank board and were asked to reconstruct all of the pieces. The results showed part-set cueing facilitated spatial location memory for both the top and bottom pieces of the circuit, even though only some of the bottom pieces were used as cues. In their second experiment, they added a third condition to the mix: a location cue group. Instead of having a full cue (i.e. the exact object in its exact position), the location-only cue had caps marking the locations of the bottom pieces, without revealing the identity of the object. It cued the location without giving the identity. Their results showed that there was greater part-set cueing facilitation for the fully cued group then there was for the location cue group, but that even the location cue facilitated performance compared to the uncued condition. The most effective cue for aiding recall provided a combination of both object identity and object location, but receiving location cues only facilitated recall as compared to the uncued condition. Overall, these results provided strong evidence that part-set cueing can facilitate spatial memory recall.
As with the Serra and Nairne (2000) order study, forming associations between the snap circuit items seemed to be a key factor in producing part-set cueing facilitation. In a follow-up study, Kelley, Parasuik, Salgado-Benz, & Crocco (2015) examined the effects of connection type (connected vs unconnected) and cue type (cued vs uncued) for spatial information. In their first experiment, participants were shown chess boards with pieces that were either connected to one another (the pieces were literally touching each other) or not (as in a normal chess game) and then gave them either a cued or uncued reconstruction test. The results showed that there was part-set cueing facilitation, but only when the chess pieces were explicitly connected to each other. In their second experiment, they replicated the results of the first experiment but with snap circuits. The important result was that, even in spatial memory, part-set cues had no influence on recall when the pieces were not explicitly connected to each other. Strong inter-item associations seem important to part-set cueing facilitation.
One of the most popular explanations of part-set cueing effects is known as the retrieval strategy disruption hypothesis (e.g., Basden & Basden, 1995). The hypothesis proposes that individuals develop their own personal strategy for encoding and recalling information such that, during the encoding process, they organize information in a particular way and, later during retrieval, they recall it in a manner that is consistent with how the information was initially organized during encoding. When information—such as part-set cues—is provided at test, the person’s retrieval strategy is disrupted because the cues are not consistent with that individual’s personal encoding and recall strategy. For instance, imagine that a person created an acronym of first letters when studying a certain list of words. On an uncued test, the hypothesis suggests that such a person would simply use the acronym he or she created during encoding to guide retrieval. When part-set cues are presented at test, the hypothesis suggests that they will likely disrupt the person’s original recall strategy—instead of trying to recall the acronym and then reconstruct the list from the acronym, participants might instead use the cues to try to remember neighboring words. The hypothesis suggests that this disruption leads to inhibition. Imagine, however, another example where a person’s original recall strategy was to associate word pairs during the encoding process. The hypothesis suggests that, in that case, having half of each word pair might enhance memory and lead to facilitation. This explanation of part-set cueing has shown itself to be quite successful (e.g., Todres & Watkins, 1981; Basden, Basden, & Stephens, 2002).
Memory for Maps & Routes
One drawback to the spatial part-set cueing studies explained above is that they used spatial tasks that were not common in the real world (i.e., reconstructing snap circuit and chess boards). However, people use spatial memory frequently in their daily lives, such as when they remember routes on maps. Routes, for example, can be studied and remembered via route descriptions or via route visualization (e.g., Buchner & Jansen-Osmann, 2008). Route description is the verbal description of the route from one person to another, such as when you recall a person explaining the route like this, “when you turn left at the post office, you will see the restaurant”. Route visualization is the process by which a map is studied with a route drawn on it. The route is developed through visual cues (e.g., showing landmarks, such as a post office) rather than verbal cues.
Historically, route memory has been studied with and without the presence of landmarks in a variety of ways. For example, Waller and Lippa (2007) categorized landmarks as either beacons or as associative cues. They defined beacons as “a landmark near enough to a goal that moving towards it leads to navigation closer to his/her goal,” whereas they defined associative cues as “stimuli that enable recall of directional responses that lead closer to the navigator’s goal” (Waller & Lippa, 2007, 911). They compared memory performance for each type of landmark across five experiments. Using a between-subjects design in Experiment 1, they found landmarks generally facilitated route learning, but that participants using beacons made fewer mistakes and learned faster than those using associative cues. In their second experiment, they replicated this result using a within-subjects design. In their remaining experiments, they showed that directional information is not retained in beacon route memory—the beacons themselves are the direction. Thus, beacons help us in guiding route memory.
In comparison, Westerbeek and Maes (2013) split landmarks into two general categories: route-external and route-internal. An example of a route-external cue would be “take a left at the pharmacy”, it describes the elements in the environment which are positioned along the path. In contrast, route-internal cues make reference to parts of the path being traveled itself, such as “take a left at the first intersection.” In their experiment, they showed participants maps with routes hand drawn on them for six seconds. Then, the participants were asked to speak into the microphone describing the route. The first key result was that, when the route was more visually cluttered, the route-external landmarks were preferred. The second result was that, when the route was longer and contained more choice points (turning left vs right vs going straight), the internal landmarks were preferred.
One could argue that route learning, with or without beacons, might be very similar to serial learning. After all, serial learning requires remembering information in order, just as route learning requires ordered memory. To explore this, Buchner and Jansen-Osmann (2008) presented participants with four conditions in a 2x2 within-subjects design to gain a better understanding of route learning. They manipulated whether set up was dynamic or static, with dynamic meaning that the observer approached the objects and passed them rather than a static, serial situation with the objects being presented one after the other with no context given. So essentially, they were seeing just objects, one after the other. They also varied the presence and absence of spatial context. If route learning was nothing more than serial recall, then the four groups would perform the same. That is, if route memory is just serial learning, the context of the four different conditions should not affect the learning—it would not matter whether there was spatial context or not or whether the condition was dynamic or static conditions. They showed that the recall of landmarks in a dynamic presentation with spatial context was better than the other three conditions. This showed that there is more to route learning than simple serial learning and memory.
Given that research on route memory is a well-established area, it is a prime candidate for studying the effects of part-set cueing. Before examining this link, however, this paper will explore the neural bases of memory in general and then, more specifically, the neural bases of spatial and route memory. After focusing on the neural level of analysis, the paper will return to the behavioral level of analysis and describe two novel experiments that examined the influence of part-set cues on route memory.
Neural Mechanisms of Memory
There are two dominant codes in long-term memory: episodic and semantic memory. Episodic memory has been described as “mental time travel” because it involves the experience of remembering not only the details of the event itself, but also the corresponding thoughts and feelings associated with that episode (Tulving, 1985). All memories begin as episodic memories, but over time, many memories lose their episodic details so that only the factual (semantic) information is retained. For example, imagine that a person learns about the reproduction of butterflies in a seminar series. Later that day, they can still remember episodic details such as where they sat, what the professor sounded like, how they felt during the lecture, etc. However, after a few weeks, all that remains from that seminar are the facts and the data of the butterfly reproduction, and one can no longer remember the details (Goldstein, 2015).
Present understanding suggests that the hippocampal region is where new memories are formed and the surrounding cortices are where these memories get stored (e.g., Leuner & Gould, 2010). As seen in Figure 1, the hippocampus is located under the cerebral cortex in the medial temporal lobe. The hippocampus is divided into three general portions: the dorsal, ventral, and intermediate regions. The ventral hippocampus (VH) is very close in proximity to the amygdala, which controls emotion in the brain. As a result, researchers believe that the VH is responsible in consolidating emotional memories (Fansleow & Dong, 2010). The dorsal portion of the hippocampus and the other structures of the medial temporal lobe (MTL) have been shown to be involved in spatial memory as well. Figure 2 displays these structures, which include: the hippocampal complex (CA fields, dentate gyrus, amygdala, and subicular complex) and the perirhinal, entorhinal, and parahippocampal cortices (Squire, Stark, & Clark, 2004).
Consolidation is an important process by which memory traces stabilize after they are initially acquired (Goldstein, 2014). There are two types of consolidation: synaptic consolidation and systems consolidation. Synaptic consolidation occurs when neurons are repeatedly stimulated, which causes structural changes to the involved neurons. These structural changes between the two neurons allow for more complex connections, increased release of neurotransmitters, and more rapid firing rates. Systems consolidation refers to enhancement of the connections between the hippocampus and the cortices. Reactivation—the key process in systems consolidation—occurs when “the hippocampus replays the neural activity associated with a memory” which “helps form direct connections between the various cortical areas” (Goldstein, 2014, p. 195). The processes of consolidation are aided by sleep. That is, going to sleep right after learning tends to enhance memory, presumably because consolidation
is more effective in the absence of new input.
Spatial Memory in the Brain
Multiple trace theory (MTT) is a leading explanation of spatial memory and it claims that the ability to re-experience the past in detail is a fundamental function of the hippocampus. ‘Multiple trace’ refers to the idea that memory ‘traces’ can be stored in multiple brain areas. According to the theory, cognitive maps of familiar environments are spatial analogs of semantic memory; that is, they are spatial representations enabling a person to navigate an environment but not necessarily re-experience it in rich detail (Moscovitch, Nadel, Winocur, Gilboa, & Rosenbaum, 2006). Neuronally, this means that the hippocampal complex (HC) contributes for a short time before the memories are solely supported by the cortices. This implies that damage to the cortices is potentially more detrimental than to the hippocampus itself when it comes to the storage and subsequent retrieval of memories.
Moser and Moser (1998) published research that confirmed the fact that spatial memory is not reliant just on the hippocampus. Using Long-Evans rats, they partially or fully lesioned parts of the hippocampus via surgery. They studied the ability to retrieve a learned spatial memory
task in a water maze post-operation, specifically with respect to the dorsal and ventral hippocampus areas. Moser and Moser found that, after the operation, the rats relearned the task in a single trial. These results suggest that lesions to the dorsal part of the hippocampus led to the rats having a harder time retrieving the tasks they had learned. Interestingly, they found that these lesions only partially disrupted retrieval, and that the important storage units for spatial memory were also located outside the hippocampus. These results support MTT’s assertion that the hippocampus is only a part of a more complex system for remembering spatial information.
Researchers have shown that the hippocampus is important in facilitating spatial memory in the form of navigation (Maguire, Woollet, & Spiers, 2006). For instance, animals that need to store food for the winter tend to have greater hippocampus volume. Similarily, Smulders, Sasson, and DeVoogd (1995) captured food-storing birds in the month of October, when food storing was at its peak, and compared them to birds captured during parts of the year when food storing was not as prevalent. They found that the hippocampus was significantly larger in size during food storing times, presumably because of the requirement of knowing exactly where the food was during the winter (Smulders et al., 1995). In other words, the volume of the hippocampus increased as the memory requirement increased.
Woolett and Maguire (2011) examined whether these morphological changes could happen in the human hippocampus as a function of spatial memory usage, specifically with human taxi drivers. They compared 31 qualified taxi drivers in London to 20 taxi trainees who failed to become taxi drivers. They were all given a structural MRI in order to get a comprehensive picture of the brain. They found larger hippocampi in the qualified taxi drivers. By comparing trainees who were qualified vs unqualified, they saw that the hippocampi indeed did get larger the more it was engaged. These findings support the MTT with the fact that the hippocampus is needed to acquire the spatial information.
In order to expand beyond the hippocampus, researchers examined the activation of the entire medial temporal lobe (MTL) during spatial memory tasks. Vann, Brown, Erichson, and Aggleton (2000) used 24 rats and ran them through a demanding radial arm maze (eight arms) and a less demanding arm maze (one arm). The next day, the rats were run through the maze again and then their brains were removed. The MTL was stained with Fos in order to see the activity of the brain using a microscope to count the number of stained nuclei. They found that the entorhinal cortex showed a significant increase in activation as the task became more demanding. While other areas of the MTL, such as the dentate gyrus, ventral hippocampus, and dorsal hippocampus were activated during retrieval, the entorhinal cortex (EC) was activated more intensely than other portions of the MTL (Vann et al., 2000). The fact that the EC was more strongly activated than the other parts of the MTL during retrieval suggests that it may be specific to the storage of spatial memories.
As mentioned earlier, sleep enhances memory in general because consolidation is more effective during sleep. Researchers showed that overnight sleep also specifically enhances spatial memory acquisition (Rauchs, Orban, Schmidt, Albouy, Balteau, Degueldre, Schnackers, Sterpenich, Tinguely, Luxen, Maquet, & Peigneux, 2008). Girardeau, Benchenane, Wiener, Buzsáki, and Zugaro (2009) went further by examining whether SPW-Rs (sharp wave ripple complexes), a type of slow sleep wave, were important for transferring the memories from the hippocampus to the cortices. They had rats learn a radial arm task during the day, and later that night, they stimulated the ventral hippocampal commissure, which interrupted SPW-Rs without affecting sleep. The following day, they put the rats through the task again and found that performance suffered. These results suggest that this type of slow-wave sleep is necessary for the memory consolidation between these two areas of the MTL. Although several questions are raised by this study, such as whether spatial memory is the only type of memory affected by this slow-wave sleep, it does further confirm the fact that the interaction between the hippocampus and the entorhinal cortex are necessary for long-term memory consolidation.
The Impact of Place and Grid Cells on Spatial Memory
The next step in understanding spatial memory is to determine how the hippocampus and the entorhinal cortex interact with each other specifically. The hippocampal-entorhinal spatial representation system contains place cells and grid cells. Place cells were first introduced by O’Keefe and Dostrovsky (1971) and were further studied by Henze, Borhegyi, Ciscsuari, Mamiya, Harris, and Buszaki (2000) among others. These researchers found that rats had place cells in the hippocampus that fired whenever they were in a certain place in a local environment and that the neighboring place cells fired when the rat was in nearby locations such that the entire environment was represented by the activity of the local place cells. They also found that these place cells activate in terms of the rat’s orientation, not the environment’s orientation. For instance, when a rat walks in circle, the firing pattern of place cells turns along with the rat. In contrast, when a rat stays in one place and the proximal environment changes, the corresponding place cells did not turn; they maintained their normal firing. Part A of Figure 3 displays place cells firing in one area that corresponds to an animals’ position in space.
Grid cells are found in the entorhinal cortex (EC) (e.g., O’Keefe & Burgess, 2005). These are called grid cells because the pattern that they fire in is ‘grid-like’ as seen in part B of Figure 3. Grid cells show a different firing pattern than place cells. A given cell could have multiple firing fields. Grid cells fire for spacing, orientation, and phase. Grid cells have periodic firing fields that form a regular triangular grid across the environment (see Part B of Figure 3). Grid fields are generated by path integration to serve as part of a neural map of self-location. The grid cells have a more detailed firing than the hippocampal place cells, because they fire for three things as opposed to one firing that each place cell does.
Researchers suggest that the difference in firing between place and grid cells could be caused by the grid cells (e.g., Rolls, Stringer, & Elliot, 2006). The hippocampal place fields may be formed by summing inputs from the grid cells. This means that the entorhinal cortex contains detailed information, which was stored with the help of the hippocampus and place cells. Subsequently, when it becomes time to retrieve the spatial memory, the place cells give a “summary” firing of what the grid cells have fired; the information is condensed into one firing rather than multiple ‘grid’ firings.
Receptors in Spatial Memory
In this next section, the receptors of the hippocampus will be discussed—beginning with a general process important to memory—long-term potentiation—and ending with specific processes related to spatial memory. Long-term potentiation is an essential feature of creating a long-term memory, and it represents the plasticity of memory (e.g., Bliss & Collingridge, 1993). The main idea of long-term potentiation (LTP) is that it can strengthen the synapse between two neurons that fire at the same time. This occurs through the release of glutamate into the synapse. Glutamate binds to AMPA and NMDA receptors on the postsynaptic site. When the glutamate binds the AMPA receptor, it releases sodium ions into the postsynaptic neuron, which causes a depolarization of the cell. This depolarization of the cell allows the magnesium block on NMDA receptors to be removed as glutamate binds it. As a result, calcium ions are released from the NMDA receptor. This release of calcium is critical to LTP because it makes the synapse more proficient for a long time. However, if either the NMDA or AMPA receptors are blocked, LTP will not occur. Hence, LTP is critical to learning and, therefore, subsequent remembering of information (e.g., Martinez & Derrick, 1996).
The next level of analysis involves exploring how spatial memory is influenced by different receptors and enzymes within the hippocampus. As mentioned above, AMPA receptors are necessary because they release the sodium that depolarizes the cell, which leads to NMDA receptor activation. Liang, Hon, Tyan, and Liao (1993) researched AMPA receptors in acquisition, consolidation, and retrieval processes of spatial memory. They placed cannula in the dorsal hippocampus of rats and they received 4 training trials on the Morris water maze for 6 consecutive days. Rats received infusion of an AMPA receptor antagonist, which blocks the receptor, into the hippocampus under one of the three schedules: Five minutes prior to each daily training session, immediately after each daily training session, or five minutes prior to the final testing trial. Pretest injections of the AMPA receptor antagonist severely impaired retrieval of the already-formed spatial memory. These results suggest that retrieving spatial memories requires AMPA receptor binding.
NMDA receptors play a critical role in long-term potentiation; their release of calcium is what makes the synapse more efficient for an extended period of time (e.g., Schiller, Schiller, and Clapham, 1998). Nakazawa, McHugh, Wilson, and Tonegawa (2004) examined NMDA receptors in the hippocampus as they related to retrieval of spatial information. In their literature review, these researchers found that previous research showed that a block on NMDA receptors in the hippocampus did not affect pre-trained spatial memory tasks. This makes sense since it is now understood that the spatial memory is not stored in the hippocampus, but in the entorhinal cortex. In their experiment, they used NMDA-receptor-knockout mice, which means that these mice could not release calcium, along with a water maze task where the platform moved around. After the initial trial of learning the cues of where the platform would be, the knockout mice took significantly more time to find the platforms then the normal mice, which indicates that the encoding of spatial memory was impaired. These results show that NMDA receptors are essential for consolidation of the spatial memory.
Conrad, Lupien, and McEwen (1999) examined adrenal steroid receptors of the hippocampus. These receptors are of interest because they bind corticosterone, which is a stress hormone, and researchers know that memory generally, memory declines as a stress increases (e.g., McEwen, 2000). There are two types of receptors: Type I and Type II receptors. The researchers injected these rats with both Type I and Type II agonists and antagonists to see what levels of binding of each type of receptor affected spatial memory. They measured corticosteroid serum levels for both types of receptors, as well as spatial memory capacity. They found an inverted U relationship between Type II binding and spatial memory capacity. When binding was absent or high, memory capacity was poor. When the levels of binding were low-moderate, spatial memory capacity was good. Memory capacity was unaffected by Type I receptors. The results suggest that since Type II receptors are the ‘stress detectors’, very low or very high levels of stress will impair memory.
Mitogen-activated protein kinase (MAPK) is a specialized enzyme that, when activated in the human body, leads to the eventual death of the cell in which it was activated (e.g., Sang, Stiehl, Bohensky, Leshchinsky, Srinivas, & Caro, 2003). Blum, Moore, Adams, and Dash (1999) researched the role of MAPK in long-term spatial memory. They trained rats with a spatial memory task in a water maze and then they injected them with JNK, which is an antibody that binds to MAPK. This is essential to detect the precise levels of MAPK in the hippocampus when the rat brains are sectioned. They found that MAPK levels in the dorsal hippocampus were significantly higher than in the ventral hippocampus. More precisely, they found that the subfields of CA1/CA2 of the hippocampus, which project to the neocortices (e.g., entorhinal cortex) had more significant MAPK activation than CA3 and the dentate gyrus. These results showed that the dorsal portion of the hippocampus is crucial for long-term spatial memory storage and that MAPK plays a role in the permanent storage of the memories.
Summary of Neural Mechanisms of Memory
In conclusion, our brain encodes spatial memory in the hippocampus and then distributes that information to the cortices where it is (semi)permanently stored. The hippocampus is the key location where memory is initially formed and subsequently retrieves the memory. The hippocampus increases in volume with experience and usage in both animals and humans. If the hippocampus is lesioned or damaged, spatial memory formation and acquisition is disrupted. The entorhinal cortex and the surrounding cortices are critical for long-term storage of the spatial memories. Consolidation is a key process by which memories get permanently stored in the cortices, and if slow-wave sleep is interrupted, the memories do not get stored. There is evidence of place cells in the hippocampus and gird cells in the entorhinal cortex. The grid cells store a detailed version of the spatial memory and when they fire, the place cells take the sum of these firings and provide the animal or human with a direction in which they should move. As stress levels in the brain increase, spatial memory capacity decreases. MAPK activation in the dorsal hippocampus is critical for long term spatial memory. AMPA receptor binding is necessary for the retrieval of spatial memories. Finally, NMDA receptor activation is necessary for the consolidation of spatial memories.
Overview of Present Experiments
In the present study, we built upon ideas from the previous spatial part-set cueing experiments in order to design our own set of novel experiments. We know how part-set cuing effects spatial memory, we know how beacons effect route memory but we do not know how part-set cues effect route memory and we do not know how the part-set cues and beacons interact with each other. In both experiments, participants viewed maps with routes on them and then were asked to reconstruct the routes on a test map. When beacons were present, small recognizable images of places or things were present on the map, such as a pharmacy or train station; when beacons were not present, the maps did not have these images available. When cues were present, there were parts of the route already printed on the map test sheet; when the cues were not present, the maps were blank. In Experiment 1, the participants were given whole map with route to study for two minutes; in Experiment 2, participants were shown each step in a route sequentially. The first situation is analogous to being given a map to study, whereas the second is analogous to seeing someone use a finger to trace route on a map or looking at a map while hearing step-by-step directions. We chose to consider both types of presentation because, whereas most part-set cueing experiments use step by step presentation (Experiment 2), most map study experiments provide the maps and entire route simultaneously (Experiment 1). In order to cover our bases, we decided to use both conditions.
Based on previous part-set cueing research on spatial memory and serial memory, participants should show part-set cueing facilitation in these studies (e.g., Serra & Nairne, 2000; Cole et al., 2013; Kelley et al., 2015). That is, participants should perform better on cued trials than on uncued trials. Similarly, based on previous research with landmarks and beacons, participants in the beacon group should perform better than those with no beacons (e.g., Buchner & Jansen-Osmann, 2008; Westerbeek & Maes, 2013). It is unclear whether the presence or absence of cues and beacons will interact with one another. Having part of the route given back as cues might be more important than having beacons because participants may memorize using the arrow cues instead of beacons, so if both are present the beacons could be a distractor to the cues. On the other hand, having beacons might be important because participants might integrate the maps and the beacons, possibly at the expense of the arrow cues. Or, perhaps, both are of equal importance. In this case, one would not distract from the other and the participants would choose the method of memorization that is easiest for them. It is also possible that the combination of the two would allow for the participant to do better than if one or the other was present. Unfortunately, current theories of part-set cueing are relatively silent on what to predict in this situation. Experiment 1
Participants & Setting
Forty-eight introductory psychology students from Lake Forest College received extra credit for participating in this experiment. Participants performed the experiment individually while sitting in a cubicle with a computer and the stimulus materials.
Four maps were created using Map Stack (“Map Stack,” 2015). All maps were taken from areas in the western suburbs of Chicago. Each street was drawn as a black line and no street names were given on the maps; a 20-step route was drawn on each map with orange arrows using Microsoft Paint (see Appendix A). The study used a 2 (beacon type: beacon, no-beacon) x 2 (cue type: cue, uncued) repeated-measures design, where beacon type was manipulated during the presentation phase and cue type was manipulated during the test phase. Accordingly, each of the four basic maps was adapted to fit each of the four conditions—No Beacon-Uncued, No-Beacon-Cued, Beacon-Uncued, and Beacon-Cue.
The beacons were images of emoticons from an Apple iPhone keyboard. The beacons chosen were distinguishable objects/places that one might see on the road (e.g. barber shops, hospitals, and street lights) and in each map they varied so that the participant did not see the same beacon across maps (see Appendix A, all emojis that were used are displayed on the maps). Four beacons were presented along the route and four beacons that were presented outside of the route. On a cued test, five arrows from the original route were presented in their appropriate locations on the map whereas, on an uncued test, none of the original route arrows were shown. Using a Latin-square (an incomplete counterbalance technique), all map-beacon-cue combinations appeared equally often across each of the four trials. So on each trial, a participant experienced a different map, so they saw all four maps by the end of the experiment.
Presentation of the instructions and materials was controlled via Microsoft PowerPoint. Participants completed four trials. On each trial, participants were shown a map with a route for two minutes and then completed a digit-tracking task for 10 seconds. Digit-tracking task is when a participant saw a single number flash on screen every 0.8 seconds and had to say each aloud. After the distracting task, the participants moved to one of four cubicles where they were given 90 seconds to reconstruct the original route by drawing it on a blank map that matched the experimental condition. This process was repeated until the participants finished all 4 trials. Two of these trials offered cues on the blank maps (5 steps given; 15 to reconstruct), whereas the remaining trials had no cues (0 steps given; 20 to reconstruct). Following the final trial, the students were debriefed and thanked for their efforts.
A strict scoring criterion was used such that, to be marked correct, the given step had to be in its exact proper location on the map. On cued trials, the proportion correct was taken out of 15 whereas, on uncued trials, the proportion was calculated out of 20.
Figure 4 displays the mean proportion correct placement as a function of beacon type and cue type. A two factor repeated-measures analysis of variance (ANOVA) revealed a significant main effect of cue type, F(1, 31) = 33.33; p < .001, ŋ2 = .518. Overall, performance was significantly higher in the cued (.867) compared to uncued (.695) condition. The main effect of degree of beacon type was significant, F(1, 31) = 11.23; p = .002, ŋ2 = .266. Performance was significantly higher in beacon (.827) compared to no-beacon (.734) trials.
The interaction between beacon type and cue type was significant, F(1, 31) = 6.284; p = .018, ŋ2 = .169. On the no-beacon trials, part-set cueing facilitation was at 24% which means that the cues were particularly helpful in the absence of beacons. Although part-set cueing facilitation was significant when the beacons were present, it was reduced to an 11% advantage, so the cues were only moderately helpful when the beacons were present.
Serra and Nairne (2000) showed that, as long as the items were associated with each other, part-set cueing should help facilitate recall. Kelley et al. (2015) took their results a step further showing that part-set cueing facilitated spatial memory recall when the pieces touched in space. The results of the current experiment are consistent with these studies involving spatial memory and order memory—specifically, part-set cues enhanced memory for a route of associated information. These results were consistent with the thought that the cues and beacons separately would aid in memory facilitation. In the end, the data suggest that part-set cueing facilitates spatial memory recall, as does the presence of beacons. When the two appeared together, participants remembered the routes particularly well, but not significantly better than just with part-set cues alone. Of course, overall performance was quite high in this task. Maybe if performance wasn’t so near to the ceiling, we might have seen a difference between these two conditions.
Participants & Setting
Thirty-two introductory psychology students from Lake Forest College received extra credit for participating in this experiment. Participants performed the experiment individually while sitting in a cubicle with a computer and the stimulus materials.
Materials were the same exact ones used in Experiment 1.
Presentation of the instructions and materials was controlled via Microsoft PowerPoint. Participants completed four trials. On each trial, participants were shown a map and the route developed step by step (six seconds for each step, two minutes total) and then completed a digit-tracking task for 10 seconds. After the distracting task, the participants moved to one of four cubicles where they were given 90 seconds to reconstruct the original route by drawing it on a blank map that matched the experimental condition. This process was repeated until the participants finished all 4 trials. Two of these trials offered cues on the blank maps (5 steps given; 15 to reconstruct), whereas the remaining trials had no cues (0 steps given; 20 to reconstruct). Following the final trial, the students were debriefed and thanked for their efforts.
A strict scoring criterion was used such that, to be marked correct, the given step had to be in its exact proper location on the map. On cued trials, the proportion correct was taken out of 15 whereas, on uncued trials, the proportion was calculated out of 20.
Figure 5 displays the mean proportion correct placement as a function of beacon type and cue type. A two factor repeated-measures analysis of variance (ANOVA) revealed a significant main effect of cue type, F(1, 31) = 18.57; p < .001, ŋ2 = .375. Overall, performance was significantly higher in the cued (.772) compared to uncued (.613) condition. The main effect of degree of beacon type was significant, F(1, 31) = 17.82; p < .001, ŋ2 = .365. Performance was significantly higher in beacon (.772) compared to no-beacon (.614) trials.
The interaction between beacon type and cue type was significant, F(1, 31) = 21.175; p < .001, ŋ2 = .406. On the no-beacon trials, part-set cueing facilitation was at 29% which means that the cues were particularly helpful in the absence of beacons. However, when beacons were present, the part-set cueing facilitation was not significant (p = 0.516), and reduced to a 3% advantage, so the cues were not helpful when the beacons were present.
These results showed that part set-cueing and presence of beacons, individually, facilitated the recall of spatial memory. The main difference in the results of Experiments 1 and 2 was that, when beacons were present with the cues in Experiment 2, the part-set cueing advantage was lower (3% vs. 11%). So the cues were only minimally helpful in the presence of beacons when participants received step-by-step presentation of the route. Currently, the reason for this reduction in part-set cueing facilitation is unclear. Aside from this, the results of Experiment 1 and 2 were largely consistent with each other, which suggests that both studying of the map and finger tracing techniques of memorization are likely to use cues and beacons equally to memorize a route.
In our first experiment, we gave participants two minutes to memorize a route as it appeared on a map. Across four different trials, the participants were faced with four conditions (no cue-no beacon, no cue-beacon, cue-no beacon, cue-beacon) during recall. We predicted that the conditions using beacons or cues, and the one condition with both of them present would lead to a significant advantage in spatial memory recall compared to condition without cues and beacons. Our results supported these predictions and were consistent with previous research by Cole et al. (2013). We also found that participants did not perform significantly better with both cues and beacons present than when they were just given cues alone. Generally, part-set cueing facilitation was reduced in the presence of beacons. Performance was worst when participants received neither cues nor beacons.
In Experiment 2, we changed the presentation; instead of seeing the entire route on the map for two minutes, the participants were given the map and then each step of the route appeared sequentially. Every six seconds the next part was added until all 20 steps were shown. We replicated the main effects of cue type and beacon type that were reported in Experiment 1. The major difference in the two studies was that, in Experiment 2, there was reduced part-set cuing in the presence of beacons, from 11% (Experiment 1) down to 3%. This could be due to the presentation. When participants were studying the maps, beacons likely were helpful to remember because the route was being presented one by one, so they were probably paying attention to how the route formed around the beacons. One could argue that they were able to focus on one small chunk at a time and everything in that chunk could be processed. In contrast, participants might not have used this strategy when seeing the entire map all at once in Experiment 1. Further research would be needed to test this explanation.
The present results have several potential implications. First, when studying a route for later recall, a person is likely to make use of beacons both during encoding and during retrieval. For instance, a person is likely to encode and retain that on their way to work they go straight past a Shell gas station and make a right near a bank. These beacons help guide the person to their destination and facilitate their recall.
Next, part-set cues seem to facilitate route memory when those cues are part of the to-be-remembered route. What this means is that if a person is to receive a cue while they are traveling a route they are more likely to remember the rest of it. For example, if you went to visit your cousin in Kansas once and you can kind of remember how you got there the next time you go and your spouse next to you says, “Make a left at the third intersection,” that could cue you to remember the rest of the route, even though you only traveled it once.
Moreover, the current results are consistent with the explanation provided by Kelley et al. (2015) that part-set cueing facilitation occurs when there are strong associations between items (see also Bäuml and Aslan, 2006). They showed, with chess pieces and snap circuits, that cues only facilitated memory if they were directly connect to one another during encoding. Our results confirm these findings as our routes were connected; no step had a gap or jump in between. This further suggests that participants might have used a chaining technique (one step in the route connected to the next step in the route).
The present study had a variety of strengths. For instance, none of the maps were similar to each other, so the participants were not likely to have any carryover effects (similarity advantage; interference deficit). The use of the Latin square design to counterbalance the conditions assured that no map/route or variable combination would have the advantage over the other. The environment in which the participants took the study was strictly controlled, and there was not any excess noise or distractors around them. The experiments themselves were well designed.
One limitation of the present studies was that the participants had a very high performance whenever a beacon or cue was present (or both). Ceiling effects could be an issue. This high performance might have been due to the fact that the routes and maps themselves did not have enough options for turns or alternative routes present in a typical route trying to reach a destination. Another potential limit was that we only tested participants in a laboratory setting. That is, participants were not required to actually go out and follow routes in the real world. By only testing their route navigation on paper, we are restricted in how far we can generalize these results.
Some of these limitations motivated an additional pair of pilot experiments. In an attempt to bring performance down from the ceiling, we chose maps with more of a grid-like structure, which allowed for more possible routes, and we increased the number of steps in the route from 20 to 25. We thought these changes might make the task more challenging for the participants. To our surprise, the first pilot experiment with 10 participants showed that they scored even higher on these tests than on those reported in Experiments 1 and 2 above. Although we tried to make the test harder, we actually made it easier for them. This prompted a second pilot study with 30 participants in which we used the grid-like maps again but reduced the amount of encoding time from two minutes to 90 seconds given to the participants and increased the length of the distracting activity from 10 seconds to 20 seconds. Although performance dropped a bit, performance still remained higher than in both Experiments 1 and 2. When we asked participants why they found the task so easy, some commented that there were many dead ends in the routes, which ultimately limited the number of possible routes. These participants suggested that they really only needed to know whether to go right or left at many of the choice points along each route. An obvious, and important, next step would be to attempt to replicate Experiments 1 and 2 with different maps that are more challenging. Clearly, we tried to do this but we ended up failing in the execution of the new maps and we simply ran out of time in the semester try any other modifications.
As reviewed earlier, given the specific parts of the brain that are activated during memory tasks and, more specifically, during spatial memory consolidation, we expect that portions of the brain like the hippocampus were active during the present experiments. The hippocampal complex is likely activated during these trials because the memory consolidation has not occurred 30 seconds after attempting to memorize the route. We expect that the medial temporal lobe, which is critical for spatial memory tasks, to be active during the memorization and recall portions of the study because this was a spatial memory study.
One next step to enhance our understanding of the neural mechanisms involved in route memory and part-set cueing would be to examine brain activity during these tasks. For example, a future study could use the present design while scanning the participants in an fMRI and see what parts of the brain are most active. We would expect the hippocampus, medial temporal lobe, and more specifically the entorhinal cortex to be active. As explained in the Introduction, we have an understanding of how the brain is active during memory processes and there is some understanding of the spatial memory process, but this type of study would be more relevant to the current set of questions.
Another compelling line of future research would be to design a study where a participant uses virtual-reality to simulate driving a car, which increases the real-world nature of the task. In the presentation phase, the participant could be shown a video where they are viewing the virtual environment in first-person as they progress along a certain route (i.e., from behind the virtual wheel). Then in the testing phase, they would have to choose whether to go left, right, or straight at choice points. Beacons could be presented as seen in real-life. To manipulate cueing in this situation, when a choice point is reached, the car would automatically go to the next point. The implications of this research would show how our brains recall in a detail-rich environment that closely resembles reality. If we learned that participants paid attention to other things in the virtual environment, then we could open up more research avenues to see what things in our environment guide our spatial navigation.
A final interesting direction for future research would be to see whether street names or beacons are preferred in route memorization and recall. For instance, we could set up a task where participants are given a virtual reality again, and they are lead along the route. We can have both beacons and street names equally visible to the participants. After they are lead along the route, the participants must speak into a microphone and tell someone how to get along this route. We would predict that the participant would likely remember more beacons then they would street names. This research would help to show what is more important for people when they are memorizing routes in real world context.
Note: Eukaryon is published by students at Lake Forest College, who are solely responsible for its content. The views expressed in Eukaryon do not necessarily reflect those of the College.
Basden, D. R., & Basden, B. H. (1995). Some tests of the strategy disruption interpretation of part-list cueing inhibition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21(6), 1656.
Basden, B. H., Basden, D. R., & Stephens, J. P. (2002). Part-set cueing of order information in recall tests. Journal of Memory and Language, 47(4), 517-529.
Bäuml, K. H., & Aslan, A. (2006). Part-list cueing can be transient and lasting: The role of encoding. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32(1), 33.
Bellezza, F. S., & Hartwell, T. C. (1981). Cueing subjective units. The Journal of Psychology, 107(2), 209-218.
Bliss, T. V., & Collingridge, G. L. (1993). A synaptic model of memory: Long-term potentiation in the hippocampus. Nature, 361(6407), 31-39.
Blum, S., Moore, A. N., Adams, F., & Dash, P. K. (1999). A mitogen-activated protein kinase cascade in the CA1/CA2 subfield of the dorsal hippocampus is essential for long-term spatial memory. The Journal of Neuroscience, 19(9), 3535-3544.
Buchner, A., & Jansen-Osmann, P. (2008). Is route learning more than serial learning? Spatial Cognition & Computation, 8(4), 289-305.
Cole, S. M., Reysen, M. B., & Kelley, M. R. (2013). Part-set cueing facilitation for spatial information. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39(5), 1615.
Conrad, C. D., Lupien, S. J., & McEwen, B. S. (1999). Support for a bimodal role for type II adrenal steroid receptors in spatial memory. Neurobiology of Learning and Memory, 72(1), 39-46.
Dong, T. (1972). Cued partial recall of categorized words. Journal of Experimental Psychology, 93(1), 123.
Fanselow, M. S., & Dong, H. W. (2010). Are the dorsal and ventral hippocampus functionally distinct structures? Neuron, 65(1), 7-19.
Fritz, C. O., & Morris, P. E. (2015). Part-set cueing of texts, scenes, and matrices. British Journal of Psychology, 106(1), 1-21.
Girardeau, G., Benchenane, K., Wiener, S. I., Buzsáki, G., & Zugaro, M. B. (2009). Selective suppression of hippocampal ripples impairs spatial memory. Nature Neuroscience, 12(10), 1222-1223.
Goldstein, E. B. (2015). Cognitive psychology: Connecting mind, research, & everyday experience (4 ed.). Stamford, CT: Cengage Learning.
Henze, D. A., Borhegyi, Z., Csicsvari, J., Mamiya, A., Harris, K. D., & Buzsáki, G. (2000). Intracellular features predicted by extracellular recordings in the hippocampus in vivo. Journal of Neurophysiology, 84(1), 390-400.
Kelley, M. R., Parasiuk, Y., Salgado-Benz, J., & Crocco, M. (2015). Spatial part-set cuing facilitation. Memory, 39, 1615-1620.
Leuner, B., & Gould, E. (2010). Structural plasticity and hippocampal function. Annual Review of Psychology, 61, 111.
Liang, K. C., Hon, W., Tyan, Y. M., & Liao, W. L. (1993). Involvement of hippocampal NMDA and AMPA receptors in acquisition, formation and retrieval of spatial memory in the Morris water maze. The Chinese Journal of Physiology, 37(4), 201-212.
Maguire, E. A., Gadian, D. G., Johnsrude, I. S., Good, C. D., Ashburner, J., Frackowiak, R. S., & Frith, C. D. (2000). Navigation-related structural change in the hippocampi of taxi drivers. Proceedings of the National Academy of Sciences, 97(8), 4398-4403.
Maguire, E. A., Woollett, K., & Spiers, H. J. (2006). London taxi drivers
Eukaryon is published by students at Lake Forest College, who are solely responsible for its content. The views expressed in Eukaryon do not necessarily reflect those of the College.
Articles published within Eukaryon should not be cited in bibliographies. Material contained herein should be treated as personal communication and should be cited as such only with the consent of the author.