1. Radial Arm Maze Task (RAM)
1.1. Free-Choice and Forced-Choice Version
The Radial Arm Maze (RAM) task, developed by Olton and Samuelson (1976) to assess the spatial abilities in rodents, is also used in several studies on children
[1][2][3][4][5][6] and adults
[7][8][9].
RAM consists of a central holding area from which several identical arms, commonly eight, radiate, and the task’s difficulty depends on the number of them. There is a hidden reward at the end of each arm, generally a coin or a little toy for children.
Different RAM paradigms take into account the environmental cues and the kind of spatial memory process to be investigated. Generally, it is possible to distinguish two main classic paradigms: free-choice and forced-choice RAM versions (
Figure 1).
Figure 1.
Free-choice (
a
) and forced-choice (
b
) paradigms are represented.
In the free-choice version, the subjects have to take all the rewards and know that the arms are rewarded only once (declarative rule). To solve the task without errors, the subject has to make use of mnesic and mapping abilities, as well as proficient explorative strategies
[3][10][11]. Carrying out the RAM in several trials, these competencies can be learned during the various phases of the task.
For many years, the free-choice RAM version has been considered appropriate for evaluating the correct functioning of working-term memory abilities by detecting the number of errors (e.g., returning to arms already visited). However, some authors observed that the longest sequence of correctly visited arms, corresponding to spatial span parameter, can also depend on the type of strategy put into action to explore the maze, suggesting the potential that RAM offers to evaluate procedural memory processes
[12][13]. In addition, employing different parameters, the free-choice RAM version allows to efficaciously study the explorative strategies used by the subject. For example, it is possible to analyze if he/she visits a specific sequence of arms or always beginning a run from the same arm (“praxic” strategy), or if he/she solves the task by referring to specific environmental stimuli (“taxic” strategy), or finally, if he/she exploits mapping abilities to build a cognitive spatial map (“place” strategy)
[14][15][16]. To refer to these different strategies, numerous terms, sich as “motor”, “cue”, or “relational” strategies, respectively, have been used in other studies
[3][15].
Therefore, to distinguish explorative from working mnesic components, it is possible to use the forced-choice RAM version. In this protocol, each trial consists of two phases. In the first phase, the subjects have to collect only four rewards, while the remaining ones are inaccessible. In the second phase, he/she has to collect the rewards of the four arms not visited in the first phase. Success depends on remembering the arms visited in the first phase (rather than putting into action particular search patterns), thus emphasizing working memory requirements. Putting into action a specific exploratory strategy is avoided using different angles to separate the opened arms (i.e., arms 2,4,5,8). Therefore, although both paradigms investigate spatial memory processes, they analyze different aspects of these processes.
Both versions of RAM can be cued or uncued
[3][17]. In the cued RAM version, each arm is made physically distinct by visual stimulus at its end. In the uncued RAM version, visited arms can be remembered by the subjects in relation to their spatial relationship to distal extra maze cues. It is easy to understand that when using the cued RAM version, the subject is forced to apply a taxic strategy to solve the task, while an allocentric strategy in the uncued RAM version is more appropriate. Again, the type of paradigm allows to investigate different aspects of spatial exploration.
Therefore, the choice of a specific RAM paradigm depends on the type of study objective to be achieved, and the normotypical and clinical population to be studied. For example, for children around the age of four years, who have not yet developed short-term memory processes, the use of free choice may be more appropriate.
In
Table 1, the main parameters analyzed in free-choice and forced-choice RAM versions are reported to emphasize the different faces of memory components that can be studied throughout real and virtual RAM tasks.
Table 1.
Illustration of the main parameters used to analyze the performances in RAM task.
| Free-Choice RAM Version |
|
| Forced-Choice RAM Version |
| (Referred to the Second Phase of the Task) |
|
| Total time to complete the entire task |
| Time to reach each reward |
|
| Total time to complete the second phase of the task |
|
| Latency to select the first arm |
|
| Latency to select the first arm |
|
| Total entries (arms correct and incorrect visited) |
|
| Total entries (arms correct and incorrect visited) |
|
| Distance travelled |
|
| Distance travelled |
|
| Movement speed |
|
| Errors |
|
| Frequency of successes/Percentage of correct visits/Search efficiency |
|
| Across-phase errors |
|
| Errors/Error-free trials |
|
| Within-phase errors |
|
| The longest sequence of correctly visited arms |
|
| The longest sequence of correctly visited arms |
|
| Percentage of angles turned (45°, 90°, 135°, 180° or 360°)/Angle change/Strategy fixation |
|
|
| Perseverations (consecutive entries into the same arm or the re-entries into a fixed sequence of arms) |
|
|
| Declarative mastery |
|
|
It is appropriate to note that in free-choice and forced-choice RAM versions, the subject walks around the maze, and this promotes the integration of the mechanisms that link perception to action. This feature suggests that RAM is a complete task as it allows to also analyze perceptive and motor processes. Furthermore, the exploration of an environment through moving in it accelerates the spatial learning processes, allowing the formation of a spatial cognitive map
[18], thus indicating RAM as a tool for devising virtual personalized neurorehabilitation training, as is already being done with other experimental protocols
[19].
1.2. Table RAM and Visuospatial Peripersonal Abilities
The RAM is cataloged among large-scale behavioral tasks since it is a walking task. The subjects are inside the maze and see it from the inside, thus promoting an allocentric and egocentric encoding. The participant is compelled to build a spatial cognitive map of RAM to orient and move himself/herself in it. In this way, the declarative competence of the environment is probably built through procedural competence
[13]. Recently, Foti and collaborators have developed a RAM table version that allows studying the visuospatial peripersonal abilities through body–objects interaction
[20]. In fact, in this table RAM version, the participant is forced to explore the portion of space accessible with the limbs in order to resolve the task, which was presented to children as the “Ladybug game”. The child had to move the older sister ladybug, placed on the central platform, to find its sisters hidden inside the caps at the end of each arm
[20]. The child is seated in front of the RAM and has visual access to the maze in all its completeness. Seeing it from above, it is likely that the construction of the spatial cognitive map may be facilitated because declarative knowledge is promptly formed. In addition, recent scientific literature reported that tactile and visual stimuli inside the peripersonal space elicit stronger processing and induce a powerful multilevel activation
[21], inducing an integration of perceptive, motor, and cognitive processes. When we see an object and recognize its function, we also know how to grasp it, preparing ourselves for the action to be enacted upon it. These characteristics related to the process that links the perception to the action suggest the table RAM is an advantageous tool to improve peripersonal spatial abilities. Furthermore, in this RAM table version, the two RAM paradigms, free choice and forced choice, were administered. However, this time it is necessary to point out that the free choice paradigm served as habituation to the setting. In contrast, the forced choice paradigm constituted the experimental part of the study. The reason for this is easy to understand, as on a small scale, free exploration is elementary, even for children. In the future, it may be helpful to administer free choice to populations with marked cognitive deficits, such as in neglect syndrome.
In this line of thinking, it is interesting to note that another group of researchers has used a small-scale RAM model to investigate the age at which children begin to integrate the increasing flexibility in the conjoint use of egocentric and allocentric frames of reference
[22], obtaining data comparable to those of classic neuropsychological spatial tests, such as Corsi Block task or block construction
[1][14], indicating also the reliability of this ecological task. In the past, O’Connor and Glassman used a radial maze analog drawn on paper to study short-term memory
[7], first suggesting the RAM as a tabletop tool.
2. Applications of RAM Task in Real Environment
As described above, RAM is a behavioral ecological task on a large and small scale that allows the analyses of different facets of spatial memory. In humans, several clinical and psychological studies have extensively used the walking RAM version for analyzing the navigational abilities in individuals with typical development (TD) and the spatial deficit in specific clinical populations.
In the late ‘80s, walking RAM was used in children to study spatial memory and understand from what age it could be administered
[2][4]. These studies have shown that even preschool infants can walk in RAM. However, the variable dimensions regarding length and number of arms and the experimental setting have confused the results. About ten years after these pioneering studies, Overmann and colleagues developed a RAM built to human scale in which children were tested without explicit verbal instructions and with a longitudinal procedure for up to 16 consecutive weekdays, using free-choice and forced-choice versions
[3]. In a sense, Overmann’s study confirms the precedents, even though it aimed to observe the development of mapping abilities rather than evaluate the age of administration of RAM. Successively, other behavioral studies on TD children were carried out employing both versions of the walking RAM to investigate the ontogenesis of spatial competencies and eventually gender differences
[1][23][24][6] as well, so as to better characterize the spatial deficit in adolescents with Williams and Prader-Willi syndromes
[25][14] and to evaluate the spatial orientation of intrauterine growth retarded children
[5]. All these studies have shown how RAM can analyze the development of a process (spatial abilities) and highlight the presence and severity of a spatial deficit.
Recently, the walking RAM task has also been used to compare learning by observation to learning by doing in TD children
[26]. In this study, the authors have made clear that the observation of the correct explorative strategy showed by the experimenter promotes the development of spatial declarative and procedural knowledge, thus suggesting the RAM task as a useful tool for improving and facilitating spatial memory. In particular, the authors highlighted that the observation of a correct exploration strategy, such as the entry into the adjacent arms, induces an early development of the spatial cognitive map in the observing child. This study suggests that RAM can also be an educational tool to facilitate and accelerate learning processes.
Even in adults, the first studies that used the walking RAM date back to the 1980s. In some of these, participants’ performances were compared to those of the rats in analog mazes
[7]. Successively, the RAM task was mainly used to study human navigation behavior in health and clinical populations
[7][8][9][6][17][27], highlighting once again how RAM can be used for diagnostic purposes. Recently, a RAM version has been also used to evidence physical activity effects on spatial abilities
[28]. In fact, by comparing the performance of athletes with those of a sedentary group, it was possible to highlight how physical exercise improves spatial memory.
However, in these studies, the behavioral procedure is not always comparable. For example, in some of them, it is preferred to use the free choice version with only part of the arms baited
[29], or to insert specific cue intramaze, or change the starting arm
[17]. As already pointed out, the choice of one or the other version of the RAM task depends on the age of the participants and on the type of memory process to be studied.
Although these differences make the results confusing and not homogeneous, they demonstrate once again the extent to which RAM task is a flexible tool that can be easily adapted to the type of spatial process to be investigated and the type of deficit to be rehabilitated.
3. Potentiality and Applications of RAM Task in Virtual Environment
The RAM task is a highly ecological test because it is administered outside hospital environments and experimental settings of research laboratories. Aside for a few examples, RAM is a large-scale task that is presented as a game, especially in children. When considering the different RAM paradigms and versions, overall, on the one hand, they have favored objectives and reliable results, also correlating to aseptic paper and pencil tests. However, on the other hand, their design has hampered RAM use, as it is very expensive to assemble them in real environments. Furthermore, as RAM tasks are very often performed outdoors and generally last a few days, they are also affected by weather conditions. All these difficulties may explain why the RAM task is only partially used in humans compared to its extensive application in animal research and the numerous evidences in the implementation in virtual modality made possible by VR technology progress. As has been already pointed out, VR offers several advantages, such as the possibility to evaluate people in complete safety
[30][31]. Another possible advantage consists of manipulating the environment, for example, making it increasingly complex or easier to explore, thus allowing for more personalization as well as a more interactive subject-environment. In addition, the changes that can be made in virtual modality allow to specifically investigate the type of strategy used by the participant to solve the task. While in real RAM version, for example, it is not certain whether the participant has oriented himself/herself according to the external cues, which, although kept under control cannot be stable (for example, a strong wind, variable brightness, etc.). In virtual RAM version, it is possible to modify the surrounding landmarks and keep other conditions constant, analyzing the procedural competences in more detail. Despite this, most of the studies that have used RAM in virtual modality have adopted the forced-choice version of the task, which allows analyzing working and short-term memory processes rather than the type of strategy used by the subject. A possible explanation could be that the free choice version is apparently easier than the forced-choice one, and since the participants were mainly young adults, the researchers believed it more useful to administer a RAM version emphasizing working memory requirements.
Other potential advantages relate to the fact that the digitized versions of the RAM task can be easily shared by several groups of researchers, and that the data obtained can be entered into scientific databases. With the aim of eventually implementing rehabilitative intervention, it could be possible to imagine a sort of “videogame training” that the patient can perform inside his/her home when motivated to do so.
To the researchers' knowledge, the first evidence of a virtual RAM task goes back to the Iaria et al. study in 2003. The authors created an eight-arm radial maze with a central starting location. The maze was surrounded by a landscape (mountains and sunset), two trees, and a short wall located between the landscape and the tree. At the end of each arm, there was a staircase leading to the location where an object could be picked up in some of the arms. The participants were young, healthy adults who used a keypad to move in any direction
[32]. Successively, joysticks were also used to navigate through the virtual RAM, but the subjects were always seated in front of the computer
[33][34][35][36][29][37][10][38][11]. In 2012, a study of 599, including TD children and younger to older healthy adults, demonstrated the virtual RAM task to be a useful tool with which to investigate the changes in exploratory strategies over life span
[39], confirming the results of the studies conducted with real RAM, but adding valuable information as to environmental factors that can modulate the development of navigational strategies. In the last twenty years, studies with virtual RAM tasks have greatly increased, and more and more evidence also relates to clinical populations and children, as well as the suggestion of new paradigms of task-based RAM
[40][41][42][43][44][45][46][47][48][36][49][29]. For example, Marsh and colleagues (2015) had administered a virtual eight RAM version during fMRI scanning of adults with obsessive-compulsive disorder (OCD) in order to study the functioning of mesolimbic and striatal areas involved in reward-based spatial learning
[44]. Furthermore, other authors investigated navigational strategies in Attention Deficit Hyperactivity Disorder (ADHD) children
[43]. The use of virtual RAM in these clinical populations suggests how it is suitable for individuals exhibiting behavioral alterations.
Recently, a digitalized version of the RAM task was also used to investigate the impact of the COVID-19 pandemic on the spatial exploration in Italian University students, allowing to evidence an increase in pseudoneglect through analysis of the lateralization of the first explored arm
[50]. However, many studies use RAM in non-immersive modality to evaluate spatial abilities
[51][52][53][54][55][56][57][58][59][60][61][62][63].
To date, only two studies reported the virtual RAM task in full immersive modality
[64][65]. In particular, Kim et al. have developed a virtual RAM task with a head-mounted display to produce information about travel distance and head movement, demonstrating that this virtual task was just as competent as the walking task one in measuring spatial learning and memory
[64]. More recently, Ben-Zeev and colleagues have produced a virtual RAM task in which the subjects wore specific virtual reality goggles as a display that enabled them to see the room in a first-person perspective, as well as a rotating tool of the view, due to its capability to translate head movements in real-time as shifts of the viewpoint
[65].
From the analysis of the studies carried out, it is clear that most of them use the forced choice paradigm, and this observation deserves careful consideration. Once again, the two paradigms allowed to evaluate different facets of spatial memory. Still, the forced choice method is more sensitive to the short memory components, and is also more challenging to perform. However, in virtual modality, it is easier to modify scenarios by reducing (or increasing) the complexity of the task. Perhaps this could be why forced choice in virtual modality is more frequent than free choice.
Funding
This research was supported by funding from the Department of Humanities, University of Naples Federico II (Fondi ricerca dipartimentale 30% and 70% 2020/2021) to L.M.