A Preliminary Development of the Morris Maze Procedure in Virtual
us Moreno
, Juan M. Jurado
, Jos
e E. Callejas-Aguilera
and J. Roberto Jim
Computer Graphics and Geomatics Group, University of Ja
en, Ja
en, Spain
Psychology, University of Ja
en, Ja
en, Spain
Virtual Reality, Human-centered Computing, Psychology and Graphic Input Devices.
The Morris Water Maze (MWM) has become one of the most widely used laboratory tools in behavioural
neuroscience. It has been used in some of the most sophisticated experiments in the study of spatial learning
and memory with animals. However, human-based studies have been very limited due to the use of unrealistic
scenarios, usually presented on a computer screen where participants’ attention is poorly controlled. Recent
advances in virtual reality (VR) enable the generation of 3D environments with a high level of realism and
user’s immersion. The user’s attention plays a key role in spatial learning. Current VR systems integrate eye-
tracking devices to measure the user’s attention over virtual entities. In this paper, we present an easy-to-use
game-based simulator of the MWM, using eye-tracking VR technology to extract information about the user’s
attention. This research still in progress has achieved important hints according to the design of the virtual
scenario, user interaction and experimentation. The study conducted in this paper validates the technology as
a novel way to perform MWM focused on spatial learning and memory with human participants.
The Morris Water Maze (MWM) was defined 40
years ago as a procedure to investigate spatial learning
and memory in laboratory rats. It has become one of
the most widely used laboratory tools in behavioural
neuroscience. This procedure consists of setting a
large circular pool filled with opaque water in which
a small platform is hidden. In a standard experiment,
during the training phase animals learn that differ-
ent fixed landmarks lead to the position of the hidden
platform. Despite being a relatively basic procedure,
it has been used in some of the most sophisticated
experiments related to neurobiology and neurophar-
macology (D’Hooge and De Deyn, 2001). Regard-
ing spatial learning, monitoring attention is crucial.
Focusing on monitoring the human’s attention, there
are multiple devices based on eye-tracking, which
are able to record the user’s gaze and detect those
objects or areas where the human pays more atten-
tion (Falck-Ytter et al., 2013) (Marcos and Gonz
Caro, 2010). However, previous studies are usually
carried out in non-realistic scenarios using computer
screens as context for user interaction. The arrival of
virtual reality (VR) allows us to design and generate
realistic virtual environments, in which humans live
immersive experiences. Moreover, the subsequent in-
corporation of eye-tracking into VR systems enables
monitoring the user’s attention with a high spatial res-
olution. Accordingly, this technology opens new op-
portunities to assess the human’s attention based on
already validated procedures within the MWM. Thus,
we aim to expand our knowledge about spatial cog-
nition in humans, considering the role of attention in
spatial learning.
Figure 1 summarizes some tasks proposed for the
study of learning and spatial cognition. In this paper,
first steps are presented for the simulation of MWM
on VR focused on spatial learning and memory with
humans. We propose an easy-to-use framework that
enables the creation of multiple spatial learning and
memory tasks focused on the user behaviour assess-
ment considering eye-tracking data in VR.
The rest of the manuscript is organized as follows:
Section 2 describes the related work, Section 3 is the
core of the study, where the experiment design is ex-
plained and all the material used is presented, Section
4 shows and discusses the results. Finally, Section 5
concludes the paper and introduces future works.
Moreno, J., Jurado, J., Callejas-Aguilera, J. and Jiménez-Pérez, J.
A Preliminary Development of the Morris Maze Procedure in Virtual Reality.
DOI: 10.5220/0010897500003124
In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) - Volume 1: GRAPP, pages
ISBN: 978-989-758-555-5; ISSN: 2184-4321
2022 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
a b c d
Figure 1: The progress of spatial cognition and learning studies over time. (a) MWM with rats (SCAC-IRIB, 2020) (b)
screen-based human experiments (c) eye-tracking experiments (University, 2014) and (d) our approach.
The Morris water maze (Morris, 1984) is often con-
sidered the “gold standard” test to assess animal
learning and spatial memory in a much more acces-
sible and controlled way (Thornberry et al., 2021).
The MWM generally consists of a large circu-
lar pool of water maintained at room temperature (or
slightly above) with a fixed platform hidden just be-
low the surface of the water. Within this pool, trials
are carried out on rats. These trials consist of the exe-
cution of a given task by a rat during a fixed time. In
each trial, rats are introduced into the pool from a dif-
ferent position (North, East, South, West) and tested
individually. The time elapsed and/or the distance tra-
versed to reach the hidden platform is recorded. Dis-
tinctive landmarks (geometric images or objects such
as circles, squares, triangles, etc.) are often placed
on the surroundings. Rats can use these visual cues
to guide their navigation. Trial by trial, rats progres-
sively become more efficient at locating the platform.
Thus escaping the water by learning the location of
the platform relative to the distal visual cues. The
MWM provides a highly controlled environment for
landmark manipulation, behavioural observation, and
lesion studies.
The theoretical development in spatial cognition
based on the results obtained with animals in MWM
has motivated the interest in establishing similar tasks
with which to replicate these results in humans. In
addition, these results encourage the evaluation of
the adequacy of the theoretical developments initi-
ated with animals, and to advance in the knowledge
of spatial cognition. Several procedures have been
devised for the study of spatial cognition in humans.
One of the most relevant has been the use of virtual
reality technology for the implementation of spatial
learning tasks (Commins et al., 2020) (Machado et al.,
2019). The immediate antecedent of these procedures
is found in the use of tasks in which participants had
to learn about spatial information presented through
a computer screen (Astur et al., 2004) (Livingstone-
Lee et al., 2011) (Newhouse et al., 2007) (Newman
et al., 2007) (Piper et al., 2010) (Spiers and Maguire,
2008). These tasks favour the rapid completion of
spatial cognition studies with a high degree of exper-
imental control. Nevertheless, their results have low
external validity since the spatial experience recreated
in these tasks differs significantly from actual experi-
ence in the natural environment.
Recent advances in virtual reality technology have
favoured running tasks such as the evaluation of nav-
igation systems (Roth et al., 2020) (Sitzmann et al.,
2018) and the study of spatial memory (Antonova
et al., 2011) (Astur et al., 1998). Moreover, other
studies have been proposed in order to assess how
auditory stimuli can affect the user’s visual perfor-
mance (Malpica et al., 2020). These results vali-
date virtual reality technology concerning traditional
laboratory experimentation. According to the MWM
procedure, there is not any VR framework with eye-
tracking integration as a common standard to replicate
and customize the target tasks over different laborato-
ries. Consequently, the comparison between different
laboratory results is challenging.
The subsequent incorporation of eye-tracking into
VR systems enables monitoring the user’s attention
with a high spatial resolution. Even though some re-
searchers have used eye-tracking for spatial condi-
tioning and learning, these studies do not integrate
this technology in VR (Kiefer et al., 2017). In some
approaches, eye movements and physiological mea-
sures of memory (pupillometry) are analysed in or-
der to compare visual exploration of spatial orienta-
tion. However, this experimentation is limited due to
non-immersive tasks just using the computer screen
(Mueller et al., 2008).
Thus, the integration of eye-tracking and VR may
leverage the knowledge in spatial cognition experi-
ments. In this work, we present the first steps for the
development of a game-based simulator in order to
study spatial cognition according to the MWM pro-
A Preliminary Development of the Morris Maze Procedure in Virtual Reality
cedure. The main contribution of our proposal is the
fusion of both virtual reality and eye-tracking tech-
nologies to monitor user behaviour during the experi-
ment. Moreover, the proposed framework is designed
to be replicated under specific constraints in order to
achieve comparable results between different scenar-
ios. Our solution integrates multiple records and treat-
ments such as time elapsed and/or the distance tra-
versed in the scenario to reach the experiment goal.
In this work, a game-based simulator of the MWM
procedure is presented. This novel approach aims
at providing a new framework to set up experiments
in the case of human beings by monitoring both the
user’s position and gaze in virtual reality. As a con-
tribution of this research, we enable the acquisition
of eye-tracking data, which are crucial to analyze the
user’s attention and the impact of surrounding virtual
elements that help to achieve the goal.
To ensure a correct adaptation of the MWM pro-
cedure to virtual reality, for the study of spatial cog-
nition and memory, a virtual simulator has been pro-
posed. This simulator includes a realistic scenario and
software modules to define the cues and goals of the
experiments as well as to capture the time elapsed, the
distance traversed or the attention path among others.
The proposed methodology is depicted in Fig-
ure 2. Firstly, the MWM logic has to be implemented
taking advantage of virtual reality. The adaptation of
the original procedure is based on searching for a hid-
den treasure located in a diaphanous area that simu-
lates the well-known pool of the MWM procedure. To
this end, the user can move around and excavate over
the area by intuitive gestures using a pick. This area is
formed by 12 x 12 digging cells where the treasure is
buried (see Figure 3). Secondly, the virtual scenario is
modelled. A dense forest is created in which the men-
tioned area is located. Moreover, target surrounding
cues are placed in order to help the user toward the
treasure position. This environment ensures immer-
sive experiences enabling users to behave naturally.
Thirdly, the MWM procedure must be adapted to be
agile enough so multiple designs can be implemented.
This allows the realization of different experiments
in spatial cognition, memory, learning and condition-
ing. An important step for the validation of our proce-
dure consists of the replication of basic phenomena in
spatial learning such as acquisition and interference.
Taking this into account, we have designed an exper-
iment consisting of two learning phases. First, an ac-
quisition phase and then, an interference phase. For
both phases, participants aim to look for the treasure
whose position is always correlated to two surround-
ing virtual objects, notated as target cues. In the first
phase (acquisition), these entities are located close
to the treasure. In the second phase (interference),
the treasure position is modified concerning the tar-
get cues, being now far away from them, as shown in
Figure 4. Thus, a spatial learning function guided by
target cues is expected during the acquisition phase,
and this one should be disrupted during the interfer-
ence phase. A spatial learning is expected during the
acquisition phase as the time the user spends close to
the target cues increases over trials. This learning is
also inferred by the number of excavations that take
place in the same area. During the interference phase,
in which the treasure is located farther away from the
target cues, a decrease in behaviour (permanence and
digging) in the previous region should be observed.
In contrast, an increase of this behaviour should be
noticed in the new area in which the treasure is now
located. Fourthly, the experiment is conducted with
a group of participants and all the relevant informa-
tion for the study is collected. We recorded the user
position, movements, excavations, eye-tracking fixa-
tion and transition between entities (related to the at-
tention of each environment object). Heat maps are
generated in order to analyze the user’s attention and
its impact on the user’s behaviour. According to the
number of excavations, the user path and gaze data,
a statistical analysis is carried out. This stage is still
in progress. Currently, we are testing the experiment
with 32 participants. For the time being, only prelim-
inary results have been obtained corresponding to six
users. Finally, the last stage consists of validating the
whole process once the data is analyzed.
3.1 Participants
Although the experimentation phase is still in
progress, in this study we have collected results from
six students aged between 18 and 23 years old of the
University of Ja
en. The experimental series were ap-
proved by the ethics committee of the University of
en, protocol number CEIH 250914-2.
3.2 Apparatus and Stimuli
To carry out this study, HTC VIVE Pro Eye hardware
(HTC, 2021) and Unity game engine software (Unity,
2021) were used. The study was conducted using a
NVIDIA Geforce 1060 3Gb, a recommended GPU by
HTC for VR. Stable 90 fps were achieved to avoid
sickness from the participants (Hagita et al., 2019).
GRAPP 2022 - 17th International Conference on Computer Graphics Theory and Applications
Figure 2: Methodology followed for the adaptation of the MWM procedure to VR and the replication of a spatial cognition
Figure 3: 12x12 grid located in the diaphanous area where
the user is able to dig. Treasure is buried in one grid.
Figure 4: The torch and the barrel represent random target
cues. Both, left and right images represent the design of the
acquisition and extinction phase respectively.
The final environment consisted of a realistic for-
est that served as a starting point for our study. It was
based on a Unity asset called ”Forest Environment -
Dynamic Nature” (NatureManufacture, 2021).
Once a realistic environment was selected, the
user had the possibility to move around, observe its
entities and interact. It ensures a feeling of total im-
mersion and helps users to behave in a natural way
which facilitates getting reliable data. The proposed
movement within the VR environment is based on the
use of teleports. The SteamVR SDK (Valve, 2021)
provides teleportation based on zones, distance con-
straints and user feedback.
For the adaptation of the MWM procedure to VR,
some changes are proposed: instead of using a pool as
a search space, a diaphanous area delimited by ropes
is modelled. A treasure that is buried somewhere in
the search area has been proposed as a substitute for
the platform that had to be searched in the pool. Ob-
jects (landmarks) that help to identify the platform in
MWM were replaced by environmental objects (cues)
in our adaptation.
As cues, eight items were used. Three of them
were irrelevant fixed cues (well, menhir, cabin) that
do not indicate the position of the treasure and al-
ways remain in the same position. The other five cues
were randomly located. Two of them were random
target cues (torch, barrel), which indicates the posi-
tion of the treasure by always maintaining the same
relation to it. The other three are random irrelevant
cues (skull, little stone, buckets) that do not indicate
the position of the treasure.
As a search action, in the Morris maze task, the rat
had to swim to find a platform and be in a safe place.
In this case, the action consists of digging and moving
around the search area until the user finds the treasure
for which he/she will be rewarded.
A Preliminary Development of the Morris Maze Procedure in Virtual Reality
3.3 Procedure
First, participants were asked to read and sign an in-
formed consent before starting the task. Afterwards,
once the headset was fitted, the task began.
At the beginning of the experiment, a cover story
was told to participants to better acclimate them into
the environment. They were told that a friend had
taken them to a forest near his home. An ancient civi-
lization lived in that forest and it was attacked by van-
dals. Before fleeing, they buried their most valuable
belongings somewhere in the forest. Their friend had
a defined area where he believed they were located.
The users’ task consisted of finding the treasure and
receiving a reward accordingly. In addition, the re-
sources are limited as each dig consumes some coins.
If the participants run out of coins, they will not be
able to dig until their friend lends them a few.
Just after the cover story, a tutorial began and users
were introduced into the environment. The objective
was to get the user used to the movement and the en-
vironment. After a while, their friend found a trea-
sure and they are moved to its position. At this point,
the importance of surrounding elements to achieve the
task goal is highlighted.
As mentioned above, the goal of this experiment is
the replication of basic phenomena in spatial learning
such as acquisition and interference. For this purpose,
two phases were created: an acquisition phase and an
interference phase. In the acquisition phase, the trea-
sure was hidden close and with the same distance re-
lated to the random target cues. In the interference
phase treasure was far away from the random target
cues (see Figure 4). Each phase consists of 8 training
trials. In these trials both treasure and cues were avail-
able. Participants were able to dig in the search area
and move through the environment. They were intro-
duced into the search area, randomizing their starting
position among the North, South, East and West co-
ordinates defined on each side of the search area. The
selection of the starting coordinate was done by sam-
pling without replacement to ensure that the user does
not repeat a position until he/she has passed through
all of them. The search area had a treasure hidden in
a different position from trial to trial to avoid that the
user could identify the treasure position by using ele-
ments of the environment instead of being guided by
random target cues. The maximum duration of each
training trial was 60 seconds. When the user reached
the treasure, he/she stayed at the position for 10 sec-
onds. This delay is done to facilitate the inspection
of the entities concerning the position of the treasure.
Similarly, if participants did not find the treasure after
60 sec, they were gently pushed to the treasure, where
they also remained for 10 additional seconds. Then, a
new trial began.
For data analysis, a probe trial is introduced after
4 training trials. Probe trials are identical to train-
ing trials with the difference that there is no treasure.
Users remain for 60 seconds and then they are ad-
vised that this trial has no treasure. This is done to
prevent the probe trials from becoming interference
trials, i.e., to prevent the participants from thinking
that the target cues were no longer good predictors of
treasure location. The probe trials are the most rel-
evant since they allow us to observe how the partic-
ipants distribute their behaviour among the different
quadrants during a fixed time of 60 sec.
3.4 Data Analysis and Dependent
Figure 5: Quadrants defined for analysis. They persist be-
tween phases. In this trial from the acquisition phase, the
green quadrant is defined as the target quadrant as it corre-
sponds to the quadrant where the target keys are located.
For further analysis of the user’s performance, 4 equal
quadrants are defined (see Figure 5). These quad-
rants are afterwards delimited zones (not perceptible
by the user during the experiment) within the search
area. It allows the user’s behaviour and learning to
be analyzed. In addition, it enables exporting relevant
information on the study variables (both, spent time
and number of excavations by quadrant). For sub-
sequent analysis, an objective quadrant is defined on
which the statistical study of the values will be carried
out. This target quadrant corresponds to the quadrant
where the target keys are located (green quadrant in
Figure 5). Target quadrant position changes from trial
GRAPP 2022 - 17th International Conference on Computer Graphics Theory and Applications
Figure 6: Heat map extracted from the execution of a training trial in the acquisition phase. Grids that appear on the ground
represent the different areas where the user could dig. It can be observed how the random target cues (barrel and torch)
received the most attention, as well as the digging grids closest to them.
to trial based on the target keys position. The logic
of the data analysis is simple: if the task leads to par-
ticipants learning that the target cues point to the lo-
cation of the treasure, then as training progresses par-
ticipants will stay longer, and perform a greater num-
ber of digs, in the quadrant where the treasure is hid-
den. For data analysis, the probe trials should confirm
the acquisition and interference that appear during the
training trials.
In addition, the eye-tracking system integrated
into the VR technology uses a series of algorithms
to determine the position at which the user is looking
in the virtual world by detecting pupils’ position and
gaze direction. Using the Unity SDK, objects placed
in gaze direction are collided by a ray. It can be ob-
tained all the desired information about the collided
objects such as its name, its type (a cue, a digging
area, a tree), the position from which it has been ob-
served, time at which the user started to observe the
object, number of looks and duration of the gaze fix-
ation, etc. All this information is crucial to evaluate
the role of attention in spatial learning.
For the information obtained from eye-tracking,
we propose the generation of 360º heat maps to rep-
resent the users’ attention to different entities in the
environment. These heat maps help to validate user’s
learning through all the trials. In this way, after an
execution, a map can be generated where each ob-
ject takes a colour depending on the total time that
the user has been observing it. The continuous range
of colours used goes from blue to red (following the
colour spectrum), representing lack of attention and
maximum time looking at a given entity, respectively.
Heat maps are generated for each trial and each user.
As an example, Figure 6 shows the heat map from the
execution of a training trial by a participant. As can be
seen, the random target cues received the most atten-
tion from the user as well as the digging area around
them. This may confirm that the user has learned
which are the cues that indicate the treasure position
and the relation between them.
Time spent in the target quadrant. Figure 7 presents
the mean time spent in the target quadrant during the
acquisition phase (left side) and the interference phase
(right side). In addition to the probe trials, the first
training trial during the acquisition phase and the first
two training trials of the interference phase have been
represented. It is not foreseeable that during these
trials the participants would find the treasure, thus
maintaining the same characteristics as the probe tri-
als (only one of the six participants found the treasure
in these three training trials). As can be expected,
time spent in the target quadrant was increasing as
the acquisition trials progressed, suggesting that par-
ticipants learned that the target cues point to the po-
sition of the hidden treasure. Moreover, in the sec-
ond interference trial, a greater permanence is ob-
served, something known in the literature as interfer-
ence burst (a transient increase in response rate over
those observed in baseline during the period imme-
diately following discontinuation of reinforcement of
a response (Lattal et al., 2020) (Skinner, 2019)). Fi-
nally, a decrease in the time participants spent in the
target quadrant can be observed as interference trials
followed one after the other. This suggests that the
extinguishing treatment was effective. These initial
impressions were confirmed by the planned statistical
analyses. Comparison of the time spent in the target
quadrant between the first training trial in the acquisi-
tion phase (when participants do not yet know which
A Preliminary Development of the Morris Maze Procedure in Virtual Reality
Figure 7: Mean time spent in the target quadrant in the first
training trial and the two test trials during the acquisition
phase (left side) and the two first training trials and the two
test trials during interference phase (right side). The hori-
zontal line at second 15 indicates the mean time expected if
the user moves randomly.
cues may be relevant for treasure location) and the
second training trial in the interference phase (when
the behavioural expression is maximal as a function
of prior learning during the acquisition phase) was
statistically significant, t(5) = 3.67, p = 0.01. The
comparison between the second training trial and the
last probe trial (when interference should be higher)
during the interference phase was statistically signif-
icant too, t(5) = 3.35, p = 0.02. These results show
that participants learned to relate the target cues to
the treasure location during the acquisition phase and
then, that this relationship was disrupted during the
second training phase.
Diggings in the target quadrant. Figure 8 presents
the mean diggings in the target quadrant in the same
training and probe trials as in the previous dependent
variable. As can be seen, the number of excavations
increased until it reached its maximum value in the
second training trial during the interference phase.
Thereafter, the number of excavations decreased un-
til reaching its lowest level at the end of the interfer-
ence phase. These impressions were confirmed by the
planned statistical analyses. Comparison of the dig-
gings in the target quadrant between the first training
trial in the acquisition phase and the second training
trial in the interference phase was statistically signifi-
cant, t(5) = 3.20, p = 0.02. Comparison between the
second training trial and the last probe trial (when in-
terference should be higher) during the interference
phase was statistically significant too, t(5) = 2.88,
p = 0.03.
The present study attempted to validate a novel way to
perform MWM focused on spatial learning and mem-
Figure 8: Mean diggings in the target quadrant in the first
training trial and the two test trials during the acquisition
phase (left side) and the two first training trials and the two
test trials during interference phase (right side).
ory with human participants using an eye-tracking VR
system. For this purpose, a simple learning acquisi-
tion and subsequent interference experiment was de-
signed. The results obtained validate the task and
the procedure from a behavioural perspective, show-
ing the tool potential for the joint research of be-
havioural and attentional processes. However, the va-
lidity of the task relies on the confirmation of the re-
sults found from statistical analyses guided by the-
ory and a posteriori planned contrast. Although this
methodology may be adequate in the replication of a
study, it is not sufficient in studies that seek to fur-
ther advance knowledge of the problem they are deal-
ing with. The lack of results in this study with the a
priori analysis strategy may be due to different fac-
tors, although three are worth highlighting: First, the
number of participants, only 6, which negatively af-
fects statistical power. So, increasing the number of
participants could help address the lack of statisti-
cal power. On the other hand, the number of irrele-
vant cues may have hindered the learning of the target
cues. Therefore, increasing the number of training tri-
als or decreasing the number of irrelevant cues should
enhance the learning function during the acquisition
phase. Finally, increasing the duration of training tri-
als should facilitate learning acquisition. Future work
in our laboratory will aim at finding parameters that
favour task sensitivity for the study of spatial learning
and memory. These will also attempt to relate be-
havioural responses (i.e., displacements and digging)
to measures of attention obtained from the record-
ing of participants’ eye-tracking during their perfor-
mance. This may represent an important advance in
the understanding of the basic processes governing
spatial learning and memory. In addition, we want to
exploit the eye-tracking technology, not only through
heat maps but also by obtaining the eye transitions of
the user on the different entities in the environment.
Eye transitions will allow us to know the order in
which users were paying attention to different objects
and how this could influence their actions.
GRAPP 2022 - 17th International Conference on Computer Graphics Theory and Applications
Research was supported by Grant PGC2018-097769-
B-C22 from the Spanish Ministry of Science and In-
novation and by the Ministerio de Econom
ıa y Com-
petitividad and the European Union (via ERDF funds)
through the research project (TIN2017-84968-R).
Antonova, E., Parslow, D., Brammer, M., Simmons, A.,
Williams, S., Dawson, G. R., and Morris, R. (2011).
Scopolamine disrupts hippocampal activity during al-
locentric spatial memory in humans: an fMRI study
using a virtual reality analogue of the morris water
maze. Journal of Psychopharmacology, 25(9):1256–
Astur, R. S., Ortiz, M. L., and Sutherland, R. J. (1998). A
characterization of performance by men and women in
a virtual morris water task:: A large and reliable sex
difference. Behavioural Brain Research, 93(1):185–
Astur, R. S., Tropp, J., Sava, S., Constable, R. T., and
Markus, E. J. (2004). Sex differences and correla-
tions in a virtual morris water task, a virtual radial
arm maze, and mental rotation. Behavioural Brain
Research, 151(1):103–115.
Commins, S., Duffin, J., Chaves, K., Leahy, D., Corcoran,
K., Caffrey, M., Keenan, L., Finan, D., and Thorn-
berry, C. (2020). NavWell: A simplified virtual-
reality platform for spatial navigation and memory ex-
periments. Behavior Research Methods, 52(3):1189–
D’Hooge, R. and De Deyn, P. P. (2001). Applications of the
morris water maze in the study of learning and mem-
ory. Brain Research Reviews, 36(1):60–90.
Falck-Ytter, T., B
olte, S., and Gredeb
ack, G. (2013). Eye
tracking in early autism research. Journal of Neurode-
velopmental Disorders, 5(1):28.
Hagita, K., Matsumoto, S., and Ota, K. (2019). Study of
commodity VR for computational material sciences.
HTC (2021). Htc vive pro eye overview. Accessed on 2021-
Kiefer, P., Giannopoulos, I., Raubal, M., and Duchowski,
A. (2017). Eye tracking for spatial research: Cogni-
tion, computation, challenges. Spatial Cognition &
Computation, 17(1):1–19.
Lattal, K. A., Kuroda, T., and Cook, J. E. (2020). Early ex-
tinction effects following intermittent reinforcement:
Little evidence of extinction bursts. Journal of the Ex-
perimental Analysis of Behavior, 114(1).
Livingstone-Lee, S. A., Murchison, S., Zeman, P. M.,
Gandhi, M., van Gerven, D., Stewart, L., Livingston,
N. J., and Skelton, R. W. (2011). Simple gaze analysis
and special design of a virtual morris water maze pro-
vides a new method for differentiating egocentric and
allocentric navigational strategy choice. Behavioural
Brain Research, 225(1):117–125.
Machado, M. L., Lef
evre, N., Philoxene, B., Le Gall, A.,
Madeleine, S., Fleury, P., Smith, P. F., and Besnard,
S. (2019). New software dedicated to virtual mazes
for human cognitive investigations. Journal of Neuro-
science Methods, 327:108388.
Malpica, S., Serrano, A., Gutierrez, D., and Masia, B.
(2020). Auditory stimuli degrade visual performance
in virtual reality. Scientific Reports, 10(1):12363.
Marcos, M.-C. and Gonz
alez-Caro, C. (2010). Compor-
tamiento de los usuarios en la p
agina de resultados de
los buscadores. un estudio basado en eye tracking.
Morris, R. (1984). Developments of a water-maze proce-
dure for studying spatial learning in the rat. Journal
of Neuroscience Methods, 11(1):47–60.
Mueller, S. C., Jackson, C. P. T., and Skelton, R. W. (2008).
Sex differences in a virtual water maze: An eye track-
ing and pupillometry study. Behavioural Brain Re-
search, 193(2):209–215.
NatureManufacture (2021). Forest environment - dynamic
nature. Accessed on 2021-10-15.
Newhouse, P., Newhouse, C., and Astur, R. S. (2007).
Sex differences in visual-spatial learning using a vir-
tual water maze in pre-pubertal children. Behavioural
Brain Research, 183(1):1–7.
Newman, E. L., Caplan, J. B., Kirschen, M. P., Korolev,
I. O., Sekuler, R., and Kahana, M. J. (2007). Learning
your way around town: How virtual taxicab drivers
learn to use both layout and landmark information.
Cognition, 104(2):231–253.
Piper, B. J., Acevedo, S. F., Craytor, M. J., Murray,
P. W., and Raber, J. (2010). The use and validation
of the spatial navigation memory island test in pri-
mary school children. Behavioural Brain Research,
Roth, D., Purps, C. F., and Neumann, W.-J. (2020). A
virtual morris water maze to study neurodegenarative
disorders. In 2020 IEEE International Symposium
on Mixed and Augmented Reality Adjunct (ISMAR-
Adjunct), pages 141–146. IEEE.
SCAC-IRIB (2020). Mwm with rats image reference. Ac-
cessed on 2021-11-02.
Sitzmann, V., Serrano, A., Pavel, A., Agrawala, M., Gutier-
rez, D., Masia, B., and Wetzstein, G. (2018). Saliency
in VR: How do people explore virtual environments?
IEEE Transactions on Visualization and Computer
Graphics, 24(4):1633–1642.
Skinner, B. (2019). The behavior of organisms: An experi-
mental analysis.
Spiers, H. J. and Maguire, E. A. (2008). The dynamic nature
of cognition during wayfinding. Journal of Environ-
mental Psychology, 28(3):232–249.
Thornberry, C., Cimadevilla, J. M., and Commins, S.
(2021). Virtual morris water maze: opportunities and
challenges. Reviews in the Neurosciences.
Unity (2021). Unity game engine. Accessed on 2021-10-15.
University, L. (2014). Eye tracking image reference. Ac-
cessed on 2021-11-02.
Valve (2021). Steam vr sdk for unity. Accessed on 2021-
A Preliminary Development of the Morris Maze Procedure in Virtual Reality