RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation

Platform on Patients with Multiple Sclerosis

Margarida Henriques

, Maria Irene Mendes

, Ana Martins

, Carla Quint

1,2 a

and Cl

audia Quaresma

1,2 b

Physics Department, NOVA School of Science and Technology – FCT NOVA, Universidade Nova de Lisboa,

2829-516 Caparica, Portugal

LIBPhys-UNL, Physics Department, NOVA School of Science and Technology – FCT NOVA,

Universidade Nova de Lisboa, 2829-516 Caparica, Portugal

Neurology Department, Hospital Garcia de Orta, 2805-267 Almada, Portugal

Keywords:

Multiple Sclerosis, Visuomotor Skills, Eye Tracker.

Abstract:

Multiple Sclerosis (MS), the most prevalent immune-mediated inﬂammatory demyelinating disease affecting

the Central Nervous System (CNS), has an estimated global incidence of 2,8 million individuals. Although

its symptomatology is highly varied and unpredictable, depending on the lesions’ location in the CNS, visual

impairments are among the most common manifestations. However, conventional methods for assessing and

rehabilitating visuomotor competences are not sufﬁcient to deliver objective assessments or personalized ther-

apies. The current study addresses this gap by adapting and testing the RehabVisual platform’s usability in

MS patients. RehabVisual, developed in previous studies, aims to objectively assess visuomotor skills through

an integrated low-cost eye tracking system, offering speciﬁc clinical intervention. Before clinical application,

a normative base was established using 50 healthy individuals for later comparison. The experimental group

comprised 25 MS patients with and without conﬁrmed visuomotor alterations. The protocol involved viewing

three visual stimuli for later calculation of the mean Euclidean distance between the gaze and stimulus posi-

tions using the eye tracker, for further assessment of the patients’ performance in tracking the stimulus. The

ﬁndings conﬁrmed diagnosed visual impairments, along with their quantiﬁcation and storage for monitoring

and rehabilitation purposes, highlighting the platform’s potential as an auxiliary tool for healthcare profession-

als.

1 INTRODUCTION

Multiple Sclerosis (MS) is the most prevalent

immune-mediated inﬂammatory demyelinating dis-

ease affecting the Central Nervous System (CNS),

with an estimated incidence of 2,8 million individu-

als worldwide (Walton et al., 2020).

Patients suffering from this chronic pathology

may present a variety of symptoms depending on the

location of the CNS lesions, making it difﬁcult to

predict the disease’s course. Visual impairments are

among the most common symptoms and are often the

ﬁrst manifestation of the disease, signiﬁcantly alter-

ing the patients’ quality of life. However, the evalua-

tion of these competencies is typically based on the

https://orcid.org/0000-0003-1015-4655

https://orcid.org/0000-0001-9978-261X

subjective observation of the physicians, since it is

usually performed by the naked eye, resulting in over-

looked impairments (Sheehy et al., 2018).

In this sense, it is clear that the clinical practice

would beneﬁt from the inclusion of objective and ac-

curate methods to adequately monitor the oculomo-

tor function throughout the evolution of the pathol-

ogy to avoid neglecting possible pathological alter-

ations. Previous studies have included eye tracking

systems to achieve this outcome, namely a research

carried out in 2020 (Sheehy et al., 2020), which uti-

lized a retinal eye tracking system to objectively mea-

sure ﬁxational microsaccades (small, rapid, and subtle

eye movements that occur during ﬁxation on a station-

ary target) in MS patients. The results indicated that

these could serve as an effective measure of disability

in MS, with a higher frequency of ﬁxational microsac-

164

Henriques, M., Mendes, M., Martins, A., Quintão, C. and Quaresma, C.

RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation Platform on Patients with Multiple Sclerosis.

DOI: 10.5220/0012463700003657

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2024) - Volume 1, pages 164-171

ISBN: 978-989-758-688-0; ISSN: 2184-4305

cades associated with greater neurological disability.

Additionally, these objective methods should also

aid in the rehabilitation area, allowing the creation of

rehabilitation plans tailored to the patients’ needs, en-

suring a higher quality of life and independence in

daily tasks.

RehabVisual is a visuomotor skills stimulation

web-based platform designed both to evaluate the

oculomotor behaviour and to develop personalized in-

tervention plans. It was originally developed in a part-

nership between students and professors of biomed-

ical engineering of NOVA School of Science and

Technology, and occupational therapists and physi-

cians from the Service of Physical Medicine and Re-

habilitation of the Hospital Dona Estef

ania (Machado

et al., 2018) to improve the methodology used in

infants with developmental abnormalities. Subse-

quently, an eye tracking system was developed to in-

tegrate the platform and objectively quantify visual

impairments (Dias et al., 2020). Currently, the plat-

form has been adapted and tested in post-stroke pa-

tients (Ferreira et al., 2020) and the eye tracker has

been improved and validated (Fonseca, 2022).

The present paper aims to describe the adaptations

made to the RehabVisual platform for its application

on individuals diagnosed with MS, along with a com-

parison between the patients and a control group of

individuals without associated pathology. The entire

process was undertaken in collaboration with the Hos-

pital Garcia de Orta (HGO), both the platform’s alter-

ations and the clinical application.

2 MATERIALS AND METHODS

The current chapter addresses the instruments used

and expanded in this work, the RehabVisual platform

and its integrated eye tracking system, in Section 2.1,

along with the data acquisition methodology in Sec-

tion 2.2.

2.1 Instruments

2.1.1 RehabVisual Platform

RehabVisual was designed using different program-

ming languages, such as HTML, PHP, JS, CSS to

create the web application, and SQL to create the

database (Machado et al., 2018). Also, it allows four

different proﬁles with distinct permissions within the

platform: administrator, physician/technician, occu-

pational therapist and caregiver.

The platform has two sections: assessment and

intervention. The ﬁrst one consists of a database to

record all the relevant clinical information of the pa-

tient, namely their clinical record and ophthalmologi-

cal, behavioral, neuropsychological and functional as-

sessments, facilitating a long-term monitoring. The

intervention program presents a variety of protocols

with different stimuli according to the visuomotor

skills status, allowing the selection of a more adequate

set of stimuli for a speciﬁc patient.

The current study focused on the conﬁrmation and

monitoring of diagnosed visual alterations, namely in

the functional assessment of MS patients, so the inter-

vention program was not altered nor approached.

Regarding the speciﬁcities of the studied pathol-

ogy, it was added a new database entry for the neu-

ropsychology assessment, typically carried out in MS

patients to evaluate their executive system. Also, the

functional assessment was altered to include the ex-

perimental protocol employed in the current study, ex-

plained in Section 2.2.2.

2.1.2 Eye Tracking System

The eye tracking system was created using Matlab

software and operates ofﬂine, only requiring prior

recording of the participant’s face during stimulus

observation. Figure 1 summarizes the eye tracker’s

semi-automated workﬂow.

Figure 1: Eye tracking system operation summary.

The system needs a manual input of the video to

be analysed, the tolerance value for each eye, and the

position of the eyes and a reference point. These tol-

erance values are the input argument for the image

segmentation technique used in the image binariza-

tion process (performed with the Matlab command

grayconnected()). After the abovementioned steps

that require user interaction, the system automatically

analyses the whole video, processing every frame.

RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation Platform on Patients with Multiple Sclerosis

165

The eye detection is obtained by the

imfindcircles() function, which ﬁnds black

circles with a radius ranging from 80% to 120% of

the ﬁrst found circle’s radius. The image coordinates

of both irises and the reference point are saved in

matrices and converted into screen positions in pixels

following a calibration.

With the pixel coordinates of the subject’s gaze

and of the stimulus, it is possible to correlate these to

metrics and calculate the distance between them. The

program automatically presents the mean Euclidean

distance between the gaze and the stimulus positions

for the chosen video and respective stimulus, which

is a metric that showed a promising result in the val-

idation of this system (Fonseca, 2022). Additionally,

it also generates graphs depicting the overlay of the

stimulus positions with the gaze’s direction of each

eye (both in pixels) as a function of the video frame

number. These graphs are separated between vertical

and horizontal directions and can be used to assess a

subject’s performance in following the visual stimu-

lus presented.

2.2 Data Acquisition

Data acquisition was carried out in two separate sam-

ples, forming a control group and an experimental

group. The Ethics Committees of the HGO and the

NOVA School of Science and Technology reviewed

the present protocol and allowed its execution in their

facilities, for the construction of the experimental and

control groups, respectively.

Section 2.2.1 presents a characterization of both

groups. The experimental protocol is detailed in Sec-

tion 2.2.2.

2.2.1 Samples Characterization

Control Group. Two inclusion criteria were de-

ﬁned for the selection of the control group: minimum

age of 18 years old and absence of known pathol-

ogy that could affect ocular movements in any way.

Additionally, all subjects were asked to remove their

glasses to prevent interference with the detection of

the eyes by the eye tracker. Nonetheless, it was en-

sured that the stimulus recognition was not affected

in order to allow an accurate stimulus tracking by the

participant.

Data collection was carried out in a sample of 50

volunteers, among whom 38 (76%) were female and

12 (24%) were male. Ages ranged from 19 to 63 years

old (mean 30,3 ± 13,3 years), while the female pop-

ulation presented a mean value of 32,7 years of age

with a standard deviation of 14,4 years and the male

cohort presented a mean value of 22,5 years of age

with a standard deviation of 1,6.

Additionally, all participants willingly agreed to

collaborate in the study, providing their free consent

before initializing any experiment.

Multiple Sclerosis Group. Regarding the experi-

mental group, the inclusion criteria deﬁned were hav-

ing a diagnostic of MS and not having a relapse

in more than six months. Similarly to the control

group, the participants were asked to remove their

glasses due to the same reasons explained earlier. The

study population involved 25 participants, 18 females

(72%) and 7 males (28%). Ages ranged from 19 to

63 years old (mean 41,8 ± 11,7 years), with a mean

value of 41,2 years of age for the female cohort and

a standard deviation of 11,5 years, whereas the male

population had a mean value of 43,3 years of age and

a standard deviation of 12,9 years.

It was also taken into consideration the MS sub-

type of each patient and whether they had a diagnosis

of internuclear ophthalmoplegia, optic neuropathy, or

executive alterations, which are common symptoms

related to MS and may affect the patient’s perfor-

mance in following a visual stimulus. The most com-

mon subtype presented was Relapsing-remitting Mul-

tiple Sclerosis, accounting for 22 (88%) of the partic-

ipants, while the other 3 (12%) were diagnosed with

Secondary Progressive Multiple Sclerosis. Regarding

the neurological symptoms, 9 patients (36%) had al-

ready been diagnosed with internuclear ophthalmo-

plegia, 9 patients (36%) with optic neuropathy, and

12 (48%) with executive alterations.

Before initializing the experimental protocol, all

participants were fully informed about the aim of the

study and its procedures, and all provided free in-

formed consent.

2.2.2 Experimental Protocol

All participants from both groups were asked to visu-

alize three videos containing different visual stimuli

with increasing complexity, while resting their head

on a support, and an external camera recorded their

face. The subjects were instructed to follow the stim-

ulus solely with their eyes, keeping their head immo-

bile throughout the entire videos. The recording of the

participants’ face must include a clear image of their

eyes, uncovered and aligned with the screen. Fur-

thermore, the lighting conditions should be favorable,

minimizing reﬂections or shadows on the subjects’

ocular surfaces, and this video should be recorded at

approximately 30 frames per second, since recording

at a lower frame rate could result in the loss of rele-

vant movement information. Lastly, it is essential that

BIODEVICES 2024 - 17th International Conference on Biomedical Electronics and Devices

166

their head is rest and immobile, ensuring that the only

present movements are the eyes’.

Experimental Setup. The experimental setup is

shown in Figure 2 (on the left) and is constituted by

a laptop, a head-immobilizer, an external webcam,

and an external screen. The laptop was used to con-

trol the RehabVisual platform, while the extra moni-

tor was employed to reduce the visual clutter for the

subject, displaying only the visual stimuli intended to

be shown. Accordingly, an extra camera was neces-

sary to record the participant’s face while they visual-

ized the videos. The support, Figure 2 (on the right),

was used to immobilize the head, as the subject was

instructed to rest their chin and forehead during the

acquisition.

Figure 2: Experimental setup (on the left) and head rest (on

the right).

The monitor and the camera were positioned at

approximately 20 cm in height to align with the eye

level, and at a distance of approximately 60 cm from

the subject, allowing a comfortable viewing of the

stimuli in the participant’s ﬁeld of view.

The camera used was a Logitech C920 HD PRO

Webcam, which offers a 78º ﬁeld of view and a

recording resolution of 1920x1080 pixels (full HD)

at 30 frames per second, while the display (22”) pre-

sented a resolution of 1680x1050 pixels at 60 Hz. The

laptop utilized to control the stimuli, as well as record

and process the webcam data was an Acer Aspire E15.

Stimuli. Three different stimuli were elaborated in

collaboration with the Neurology Service of the HGO

and were shown during the experimental protocol.

The stimulus itself was the same, comprised by a

black circle with a red center (Figure 3, on the left),

differing only in the trajectory followed. Addition-

ally, all three videos started with a calibration se-

quence of 15 seconds, equivalent to 450 frames, cre-

ating a correspondence between the maximum and

minimum amplitude of the eyes and the screen edges.

Considering the scheme represented in Figure 3 (on

the right), the calibration procedure followed the se-

quence B-D-H-F, after an initial ﬁxed position for 3 s

at the ﬁrst location and with a ﬁxation on each of the

marked locations of 1 s.

Figure 3: Stimulus (on the left) and possible locations of the

stimuli’s paths (on the right).

The ﬁrst video has a duration of 28 s, in which

the stimulus moves to location E, where it remains

stationary for 10 s until the end, following the cali-

bration abovementioned. This stimulus was chosen to

ascertain the presence of nystagmus, by investigating

the capability of maintaining a steady gaze at a ﬁxed

point.

The second video aimed to assess if the subject

could achieve a smooth pursuit of the stimulus, which

could translate into the presence (or absence) of sac-

cadic intrusions. Accordingly, in this 40 s video, the

stimulus describes the path E-B-H-E-D-F after the

calibration, comprising vertical and horizontal move-

ments with 1 s ﬁxations at each marked location.

Lastly, the third video comprehends a trajectory

along the screen corners, as well as an intermittent

movement at the end, resulting in a duration of 1

minute and 40 seconds, in order to assess the subject’s

visual ﬁled and visual perception. The initial contin-

uous movement corresponds to the path E-B-H-I-A-

G-C with a ﬁxation of 1 s in each location following

the calibration. Subsequently, the stimulus fades and

reappears in another area, where it remains stationary

for 3 s, describing the unpredictable sequence E-A-I-

D-F-B-G-C-H.

3 RESULTS AND DISCUSSION

3.1 Control Group

After applying the experimental protocol to the con-

trol group, the mean Euclidean distance between the

stimulus and the gaze positions was calculated for the

ﬁrst two videos and for each subject in order to es-

tablish a reference value for further comparison with

the experimental group. The third video was analysed

separately. Table 1 summarizes the results obtained.

RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation Platform on Patients with Multiple Sclerosis

167

Table 1: Descriptive statistics of the mean Euclidean dis-

tance between the stimulus and the gaze positions (in pix-

els) for the ﬁrst two videos and for both eyes of the control

group.

Mean Euclidean

distances (pixels)

Video 2

Video

Right Left Right Left

Maximum 125 119 137 149

Minimun 49 44 53 61

Mean 84 89 109 116

As it was expected, in general the obtained val-

ues are greater for the second video. This result may

be related to the participants’ performance, as well

as to the eye detecting system. A longer video de-

mands a longer attention span and can lead to visual

fatigue, therefore resulting in a more imprecise track-

ing. On the other hand, a longer time interval implies

a higher chance of situations where the eye is not cor-

rectly detected, namely due to blinking or momentary

changes in brightness, and a higher chance of the sub-

ject moving their head. Moreover, an initial impre-

cise calibration results in more inaccurate values in a

longer video, thus leading to higher mean Euclidean

distances.

Following the same methodology of the study that

validated the eye tracking system (Fonseca, 2022), it

was established a threshold value for each stimulus

according to the approximate maximum value. There-

fore, mean Euclidean distances above 130 pixels and

150 pixels, for the ﬁrst and second stimulus respec-

tively, were indicators of difﬁculties in tracking the

stimulus. The three ﬁgures below depict the graphs

generated by the eye tracking system for the three

stimuli of one healthy participant’s left eye.

Figure 4 shows the graphs representing the move-

ments of the ﬁrst stimulus (red) and the left eye

(green) tracking it. The graphs are separated in the

two directions, horizontal (top) and vertical (bottom),

and both measure screen positions in pixels as a func-

tion of the video’s frame number. As the stimulus

remains static at the end of the video, there is a hori-

zontal line in both graphs representing the unchanged

coordinates.

Accordingly, Figure 5 depicts the graphs repre-

senting the movements of the second stimulus (red)

and the left eye (green) following it, for the horizon-

tal (top) and vertical (bottom) directions. During the

vertical movements, the graph corresponding to the

movement in the x axis shows a horizontal line, as

the stimulus does not change its x coordinate. On the

other hand, during the horizontal movements, there is

no change in the y coordinate of the stimulus, hence

the horizontal line in the graph that corresponds to the

Figure 4: Coordinates in pixels of the ﬁrst stimulus (red)

and the gaze (green) as a function of the frame number for

the left eye.

movements in the vertical direction.

Lastly, the graphs in Figure 6 detail the movement

of the third stimulus (red) and of the left eye (green)

tracking it. In this case, the stimulus fades and ap-

pears in another location to produce an intermittent

movement, which is evident by the abrupt changes

in both graphs representing each direction, horizon-

tal (top) and vertical (bottom).

Analysing the generated graphs, and taking into

consideration that the graphs for the right eye are

identical, the curves representing the participant’s

gaze are noisier than the curves representing the stim-

ulus. Yet, all the transitions and ﬁxations are correctly

identiﬁed. Another outcome is that the y coordinate

estimation of the eye movement seems less accurate

than the horizontal coordinate, for the three stimuli.

Said discrepancy might be caused by differences in

eyelid opening during vertical movements, which do

not occur during horizontal ones.

3.2 Multiple Sclerosis Group

In accordance with the previous subsection, Table 2

represents the results obtained for the experimental

group of MS patients.

Upon initial examination, we can conﬁrm that, on

average, the values are higher for this group of indi-

BIODEVICES 2024 - 17th International Conference on Biomedical Electronics and Devices

168

Figure 5: Coordinates in pixels of the second stimulus (red)

and the gaze (green) as a function of the frame number for

the left eye.

Figure 6: Coordinates in pixels of the third stimulus (red)

and the gaze (green) as a function of the frame number for

the left eye.

viduals than for the group constituted by healthy indi-

viduals, which was undoubtedly expected since this

Table 2: Descriptive statistics of the mean Euclidean dis-

tance between the stimulus and the gaze positions (in pixels)

for the ﬁrst two videos and for both eyes of the experimental

group.

Mean Euclidean

distances (pixels)

Video 2

Video

Right Left Right Left

Maximum 336 346 367 457

Minimun 58 48 81 72

Mean 113 128 159 171

sample includes patients with diagnosed visual im-

pairments. Furthermore, during the acquisition, some

patients were unable to keep their head still and track

the stimulus only with eye movements, resulting in a

relatively mobile reference point and, consequently,

in disparate results. This limitation may be linked to

the presence of executive impairments, which can af-

fect, for example, the ability to maintain focus and

follow instructions.

As mentioned above, the ﬁrst stimulus was used

to assess the presence of nystagmus. However, it was

not possible to draw any conclusions regarding this

visual alteration. This could be due to a low accuracy

of the eye tracker or a low incidence or intensity of

nystagmus in the patients studied. On the other hand,

it was possible to conﬁrm the presence of abnormal

eye movements in some cases, conﬁrming the diag-

nostic of visuomotor alterations, which can be seen in

the following ﬁgures.

Figure 7 represents the right eye movements of pa-

tient 15 while visualizing the ﬁrst stimulus. This pa-

tient is a female diagnosed with Relapsing-remitting

Multiple Sclerosis and executive alterations. The

mean Euclidean distances between the stimulus and

the gaze positions were 159 ± 96 pixels for the right

eye and 134 ± 113 pixels for the left eye.

As this patient has no diagnosed visual alterations,

it was expected that the graphs would show a smooth

pursuit of the stimulus. Yet, by the presence of several

peaks, it is clear that the patient was not able to fol-

low the stimulus’s continuous movement. Also, in the

highlighted area, it is possible to see a higher slope in

the green line, suggesting that the patient anticipated

the movement and had to adjust their gaze.

Regarding the second stimulus, its aim was to

evaluate the participants’ ability to track continuous

motion. Although the ﬁrst video had already provided

some information about these movements, it was pos-

sible to observe visuomotor alterations with this video

as well, as expected. Patient 5 had already been di-

agnosed with internuclear ophthalmoplegia and ex-

ecutive alterations, so the presence of abnormal eye

movements were predicted. The graphs generated by

RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation Platform on Patients with Multiple Sclerosis

169

Figure 7: Coordinates in pixels of the ﬁrst stimulus (red)

and the right eye’s gaze (green) of patient 15 as a function

of the frame number, with the x positions on the top and the

y positions on the bottom.

the eye tracking system corresponding to the visual-

ization of the second video by patient 5 are presented

in Figure 8. In this case, the mean Euclidean distances

between the stimulus and the left eye’s gaze positions

were 153 ± 161 pixels for the right eye and 157 ± 144

pixels for the left eye.

As can be observed, there is a peak at approxi-

mately frame 380, leading to a steeper slope of the

graph line representing the gaze position (green) com-

pared to the slope of the line representing the stim-

ulus position (red). Similarly to the ﬁrst case, this

indicates that the patient anticipated the movement,

resulting in a discontinuous motion and having to

later adjust the gaze to properly follow the stimu-

lus’s movement. The same phenomenon occurs at

approximately frame 980, during a horizontal move-

ment from the middle of the screen to the left.

The third stimulus aimed to assess the patients’

ﬁeld of vision, as well as their visual attention. In MS,

the visual ﬁeld may be affected due to the inﬂamma-

tion of the optic nerve (optic neuritis). However, at

the time of the acquisition, none of the patients was in

this situation, and as a result none of them exhibited a

Figure 8: Coordinates in pixels of the second stimulus (red)

and the left eye’s gaze (green) of patient 5 as a function of

the frame number, with the x positions on the top and the y

positions on the bottom.

loss of the visual ﬁeld. Therefore, no new results were

anticipated when compared to the outcomes obtained

with the two previous stimuli. This hypothesis was

corroborated, as the patients did not encounter difﬁ-

culties that had not already been identiﬁed with the

initial videos, nor did they experience difﬁculties in

locating the stimulus during intermittent movement.

However, it is important to integrate a stimulus to as-

sess these competences in future studies as well, as

they are usually affected in individuals with MS.

4 CONCLUSIONS AND FUTURE

WORK

The main objective of the present study was to expand

and adapt the RehabVisual platform to MS and test its

usability on patients diagnosed with this pathology,

following its adaptation and expansion for this popu-

lation as a continuation of previous work.

To achieve this goal, the experimental proto-

col was performed in 50 healthy volunteers (control

group) to establish normative values for further com-

parison and 25 MS patients from the HGO (exper-

BIODEVICES 2024 - 17th International Conference on Biomedical Electronics and Devices

170

imental group). Subsequently, the mean Euclidean

distances between the gaze and the stimulus positions

were calculated and it was evident that the values, in

general, were higher for the experimental group, as it

was expected.

The results obtained show that it is possible to

record and save a quantiﬁcation of the visual impair-

ments, as the eye tracking system was able to con-

ﬁrm diagnosed visual alterations. This platform en-

ables an easier monitoring of the disease’s progress,

along with a possible auxiliary tool for the rehabili-

tation planning. In this sense, it is evident that MS

patients may beneﬁt from the use of the RehabVisual

platform.

Although the results obtained are promising, there

are some limitations that were encountered and

should be addressed in future studies. Regarding the

eye tracking system itself, there were some errors

demonstrated in the eye detection, leading to inac-

curate values, thus the system’s general performance

should continue to be improved. Additionally, the

computational time needed to process the data was

greater than desired, taking approximately 5 minutes

to process a video of 29 seconds (with around 600

frames). On another note, the metric used for assess-

ment (mean Euclidean distance between the stimulus

and the gaze positions) was not always a good indica-

tor of difﬁculties in tracking the stimulus, since it is a

mean value. In this regard, analysing Euclidean dis-

tances in speciﬁc positions should be interesting and

could bring satisfactory results.

Also regarding future work, it is important to con-

duct a usability test in order to evaluate the platform’s

performance in the hands of healthcare professionals,

as well as to assess areas for improvement. Addition-

ally, as the present work did not focus on the reha-

bilitation area, it would be interesting to use the plat-

form to choose individualized rehabilitation programs

in future studies and test its results and relevance.

The present study offers a summary of the

methodologies and results obtained, intending to

highlight the potential of the RehabVisual platform,

as mentioned earlier. Furthermore, RehabVisual al-

lows for systematic and standardized monitoring of

patients with visuomotor impairments over time. Ad-

ditionally, it facilitates the identiﬁcation the interven-

tion methodologies’ impact, allowing necessary adap-

tations to address the speciﬁc needs of each patient

when required.

ACKNOWLEDGEMENTS

Research was supported by Fundac¸

ao para a

encia e a Tecnologia through research Grants

UIDB/FIS/04559/2020 and UIDP/FIS/04559/2020

(LIBPhys).

REFERENCES

Dias, P., Ferreira, A., Vig

ario, R., Quaresma, C., and

Quint

ao, C. (2020). Rehabvisual: Implementation of

a low cost eye tracker without pre-calibration. In Pro-

ceedings of the 13th International Joint Conference

on Biomedical Engineering Systems and Technolo-

gies (BIOSTEC 2020) - BIODEVICES, pages 235–

241. SciTePress.

Ferreira, A., Santos, P., Dias, P., Alves, A., Carmo, B., Vil-

hena, F., Costa, S., Quaresma, C., and Quint

ao, C.

(2020). Rehabvisual: Application on subjects with

stroke. IFIP Advances in Information and Communi-

cation Technology, 577:355–365.

Fonseca, P. (2022). Validac¸

ao da plataforma rehabvisual:

Ferramenta para estimulac¸

ao das compet

encias visuo-

motoras - aplicac¸

ao a doentes com avc. Master’s the-

sis, NOVA School of Science and Technology.

Machado, R., Ferreira, A., Quint

ao, C., and Quaresma, C.

(2018). Rehabvisual: Development of an applica-

tion to stimulate visuomotor skills. In Proceedings of

the 11th International Joint Conference on Biomedi-

cal Engineering Systems and Technologies (BIOSTEC

2018) - BIODEVICES, pages 173–178. SciTePress.

Sheehy, C. K., Beaudry-Richard, A., Bensinger, E., Theis,

J., and Green, A. J. (2018). Methods to assess ocular

motor dysfunction in multiple sclerosis. Journal of

Neuro-Ophthalmology, 38:488–493.

Sheehy, C. K., Bensinger, E. S., Romeo, A., Rani, L.,

Stepien-Bernabe, N., Shi, B., Helft, Z., Putnam, N.,

Cordano, C., Gelfand, J. M., Bove, R., Stevenson,

S. B., and Green, A. J. (2020). Fixational microsac-

cades: A quantitative and objective measure of dis-

ability in multiple sclerosis. Multiple Sclerosis Jour-

nal, 26(3):343–353.

Walton, C., King, R., Rechtman, L., Kaye, W., Leray, E.,

Marrie, R. A., Robertson, N., Rocca, N. L., Uitdehaag,

B., van der Mei, I., Wallin, M., Helme, A., Napier,

C. A., Rijke, N., and Baneke, P. (2020). Rising preva-

lence of multiple sclerosis worldwide: Insights from

the atlas of ms. Multiple Sclerosis Journal, 26:1816–

1821.

RehabVisual: Adapting and Testing the Visuomotor Skills Stimulation Platform on Patients with Multiple Sclerosis

171