A New Rehabilitation Device for Balance Impaired Individuals

Wiktor Sieklicki

, Robert Barański

, Szymon Grocholski

, Patrycja Matejek

, Mateusz Dyrda

and

Konrad Klepacki

Faculty of Mechanical Engineering, Gdańsk University of Technology, Gdańsk, Poland

Faculty of Mechanical Engineering and Robotics, AGH University of Science and Technology, Kraków, Poland

Faculty of Management, University of Gdańsk, Poland

Keywords: Robotic Rehabilitation, Balance Training, Biomedical Electronics, Ataxic Cerebral Palsy.

Abstract: In the paper authors present a device designed to improve the rehabilitation process of people with balance

impairment. The discussed device (JStep) utilizes a commercially available static standing frame (stander)

modified in order to fit force sensing units under the feet and in the pillows around the hips of a patient. While

executing rehabilitation tasks, the patient may compensate his balance deficiency by leaning on the pillows

around his hips. Information about weight distribution between left and right leg together with the information

about the force applied to the pillows supporting the patient's body is further presented on a display in front

of the patient. Such a setup allows physicians to work with the patient while having direct information about

compensation necessary for completing a task or gives the patient a visual biofeedback about how well he is

doing the exercise. The system is based on an ATmega controller, load cells and analogue amplifiers. In this

framework a case study is presented of a 16 y.o. patient with Cerebral Palsy affecting his cerebellum, labelled

as ataxic Cerebral Palsy. Two exercise scenarios utilizing the proposed device are discussed and results of a

6-week exercise are further presented. They show a decrease in necessary compensation in order to maintain

a standing posture as well as a better accuracy in achieving the desired force distribution between right and

left leg while standing upright.

1 INTRODUCTION

Keeping an upright standing position requires a

cooperation of nervous, muscular, skeletal and

fasciae systems and is based on the integrity of

reactions, reflexes, tonus, sensory system information

as well as the intellectual, emotional and social

capacity (Horak, 2006)(Błaszczyk, 2004)(Matyja,

2012). Free body movements, walking and finally

locomotion are the results of this complex

task (Błaszczyk, 2004), and the capacity for free

movement gives the feeling of independence and

personal safety. It’s futile to try to pick one of the

systems, as the most important as they are all directly

responsible for completing the task and their

functioning is mutually dependent (Horak,

2006)(Matyja and Domagalska, 1998).

For infants, the spinal cord and primitive

reflexes are the first to develop with prevalence of the

latter, after a short period of time after birth. Next, the

righting reflexes emerge, which lead to the

development of stability and equilibrium reactions at

the age of 6 months (Matyja, 2012)(Gilfoyle et al.,

1990). Integrity of those is essential (Batra et al.,

2011)(Zafeiriou, 2004) and leads to a correct postural

tonus in coronal, sagittal and transverse planes.

Resulting postural reactions last for the whole

life (Zafeiriou, 2004). The integration of postural

reactions takes place in cerebral cortex whereas

reticular formation is responsible for the control of

the process together with basal ganglia and

cerebellum

(Hurło and Kowalski, 2003). Proper

postural reaction gives the ability to coordinate the

position of body segments in respect to each other, in

order to maintain the desired position or to reclaim

equilibrium in the presence of a gravitational

force (Matyja, 2012)(Gilfoyle et al., 1990).

Ataxic Cerebral Palsy is one of the examples in

which the development of a person's reflexes is

stunted at some point and keeping a standing posture

is often impossible. The lack of the skill of

maintaining stability affects the possibility of social

interactions. The lack of axial limb loading can lead

to circulatory, respiratory, urinary and skeletal system

Sieklicki, W., Bara

nski, R., Grocholski, S., Matejek, P., Dyrda, M. and Klepacki, K.

A New Rehabilitation Device for Balance Impaired Individuals.

DOI: 10.5220/0007374300260035

In Proceedings of the 12th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2019), pages 26-35

ISBN: 978-989-758-353-7

dysfunctions (Drużbicki et al., 2013). Due to the

complex nature of Cerebral Palsy, resulting mobility

disorders have various forms.

It is important to differentiate between stability

and equilibrium. Equilibrium is a state in which the

adjustment of the position of body segments is a result

of bringing the resultant of acting forces on the body

down to a minimum. It is achieved by adjusting the

proper tonus of the muscular system in a stationary

environment (Błaszczyk, 2004). Whereas stability is

the ability to maintain equilibrium in a dynamic

environment where disturbances may occur (Horak,

2006)(Kuczyński et al., 2012).

In this paper we focused on the condition affecting

a particular patient, so the proposed device is

designed to meet the requirements of Ataxic Cerebral

Palsy patient rehabilitation.

Various devices are used for maintaining an erect

posture like: walking (walkers) and standing

(standers, parapodia) assistance devices. They are

usually modular with a wide base and support a

patient in an upright position by various elements of

the device. Walkers enable their users to move

around. The parapodia can be of two kinds: static

parapodium, which stabilizes the body of the patient

in an upright position, and may provide support in the

chest, hip, lumbar and knee areas, or a dynamic

parapodium, which stabilizes a patient’s body while

allowing one to move around in limited space.

Movement of the device is achieved here via repeated

body movements from side to side.

Active forms of work with a patient held upright

with a parapodium device usually focuses on

activating their manual and cognitive skills. It is

possible, because a parapodium allows the patient to

free the arms from supporting the body, while

simultaneously blocking their legs in most cases.

The available standers are very limited in their

role as assisting devices for ameliorating stability or

training correct gait patterns for patients with a

GMFCS (Gross Motor Function Classification

System (Palisano et al., 2007)) level of IV-V. Even

the dynamic parapodium doesn’t allow the patient to

train their gait, because the movement of the device

is achieved through a side to side rocking movement,

which does not help with developing correct walking

patterns. There are no strong arguments to back the

thesis, that exercising with the use of walkers allows

to minimise coordination dysfunction, which in turn

would allow the patient to be able to retain their

balance without the help of assisting devices

(Livingstone and Paleg, 2016)(Paleg and

Livingstone, 2015). Even though, a parapodium may

be a good starting point for exercise which have a goal

of getting to know the correct muscle tonus while in

an upright position.

It is a common practice, that gait rehabilitation is

performed with a help of balancing platforms and

force platforms (Woollacott et al., 2005)(Shanahan et

al., 2018). If the patient leans on the stabilizing device

though, posturography with its CoP (Centre of

Pressure) analysis does not provide correct results.

There are just a few tools, which would help training

and measuring stability based on the CoP for people,

who need standing assisting devices to maintain an

upright position (e.g. most of patients with a GMFCS

level of IV-V).

An example of a device where designers reached

beyond the goal of just keeping the patient upright is

the static-dynamic parapodium BalanceReTreiner

BalanceReTreiner (Matjacić et al., 2003)(Gałęcki et

al., 2013)(Michalska et al., 2011). It allows to keep

the patient in an upright position, while allowing for

an inclination in coronal and sagittal planes for up to

10 degrees, where the inclination is assisted with a

resistance from a spring. The feet of the patient

remain attached to the floor. The device uses a visual

feedback that shows the patient the current inclination

of their upper body. Measurements of the inclination

are registered via accelerometers. The patient is

requested to control the inclination of the body, based

on the direction and amount of inclination shown on

a screen in front of him. The assessment of the

patient’s abilities is based upon the concordance of

the directions and amounts of inclination requested by

the program and executed by the patient. The

patient’s actual CoP is affected by a possible leaning

on the device. There is no information gathered about

the pressure applied by the patient to various parts of

the device. Therefore the device does not allow to

monitor the patient’s CoP. Spring mechanism is the

resistor for patient movements. Therefore, the force

required to perform an inclination rises, as the

inclination gets deeper. This behaviour is very

different to what the patient experiences when he has

no support from the device.

There is a lack of adequate rehabilitation devices,

which would enable monitoring of forces necessary

to compensate the disturbances of stability for

individuals, which are not able to maintain an upright

position without the aid of a physiotherapist and

orthopaedic aids. This led authors to develop the idea

of the JStep device described further in this paper.

This paper is organized as follows: the device

construction and sensing system are described in

Section 2; specific case study done with presented

device and its results are described in Section 3; the

discussion is presented in Section 4.

A New Rehabilitation Device for Balance Impaired Individuals

2 REHABILITATION DEVICE

The device was designed and built so that it provides

a real-time visualisation of the body weight

distribution between the left and the right leg. It also

provides information about leaning of the patient on

left, right and rear side pillow of a static parapodium.

The device can be used as a rehabilitation exercise

platform. The patient’s goal is to follow a therapist’s

orders with visual information on the body weight

distribution and the force applied to the supporting

pillows of the parapodium. The design assures that

the patient is safely held upright.

If the pressure applied to the supporting pillows is

treated as a necessary help in maintaining stability,

then the information about the usage of the supporting

areas and the measured amount of pressure applied to

them can be an additional tool for the therapist during

stability exercises. The information about the amount

of pressure applied to the base of the left and right

foot allows identifying asymmetry of leg load.

This device is designed using basic and affordable

electronic and mechanical elements in order to make

it more accessible for potential patients.

2.1 Mechanical Platform

A static standing frame PJMS 180 (Figure.1, left) has

been chosen as a platform to be modified in order to

fit the sensors. It is a commercially available stander,

already designed to be safe for the patient. It’s crucial

that the position/length of each component of the

stander is easily adjusted, thus adapting the device to

our needs with a reasonable amount of work.

The position of the two independently mounted

platforms on which the patient stands can be shifted

in coronal plane. Pillows on the sides of the stander,

as well as the one in the front and in the rear hold the

patient in the upright standing position. Those can be

raised or lowered depending on the height of the

patient. The distance between them can also be

adjusted. Two cups straightening the knees that were

originally installed in the device have been removed

so that the patient is able to lift the leg and bend the

knee. In case the patient has little or no control over

the extensor muscles of the leg, the cups may be

replaced with a rubber band so that the legs return to

the straightened position.

The measurement of the applied force is done by

load cells mounted in the frame. Each of the

highlighted surfaces on the left in Figure 1 is capable

of sensing the force acting perpendicularly to the

surface. Load cells used in the pillows (Figure 1a and

Figure 1b) are beam type load cells NA27-005. Their

measurement range is wide enough so that a person

weighing 90 kg and with a height between 155 and

185 cm may lean on any pillow and the reading will

be adequate to the leaning force. In order to measure

the body weight distribution between right and left

leg, load cells YZC-161 are used (Figure 2, left).

Their advantage is that they have very low profile

(12mm) while being able to measure and withstand

significant loads (up to 40 kg each).

Both left and right standing platform is equipped

with a vertical barrier on the inside edge of the

platform (Figure 2, right). This is due to the fact that

the patient with whom we were working had

problems with putting his feet back on the platform

after raising them. A 10 cm barrier between his feet is

enough to manage this problem.

The device is designed targeting both private

customers and medical centres that focus on the

therapy of CP patients. With all that in mind the

design allows the device to be lightweight and easy

to move in a house or at a gym, so that it can be placed

in a convenient place. The weight and size of the

device allows carrying it around with minimal effort.

Figure 1: Static stander in a basic version (left), sensing

surfaces after mounting sensors (right); a) green - two side

pillows surfaces; b) blue - rear pillow surface; c) red -

sensors under the feet.

Figure 2: Four YZC-161 load cells bundled to measure

weight of the patient (left); standing platforms mounted on

the top of the load cells (right).

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

Shifting body weight from one leg to another

implies shifting the position of the hips to the sides as

well. For this reason the distance between each of the

side pillows and the patient’s hips remains

approximately 3 cm. This distance was chosen after

trial tests. We noticed it was still enough for the

patient to feel comfortable and safe, knowing that he

can lean on the side pillow, while it was enough for

him to shift the body weight as well. Lack of constant

pressure between the patient's body and the side

pillows of the stander requires more physical strength

from the patient, therefore it has to be adjusted

specifically for the patient.

The complete device, prepared for tests is

presented in Figure 3 with LED (Light Emitting

Diode) display mounted in front of the patient. Since

no one is in the device only two LEDs indicating null

weight applied to the right and left standing platforms

are illuminated. In this setup, the two holding bars in

front of the patient are present but during later tests

they were removed.

Figure 3: A photograph of the JStep platform with LED

display mounted in front of the modified stander.

2.2 Sensory System and Data

Acquisition Module

Each load cell mounted in the pillows is a separate

measuring cell. Load cells mounted under the

standing platforms are placed in bundles of four and

create two groups of Wheatstone half-bridges

configuration on each of the two platforms. This gives

possibility to measure body weight distribution bet

ween left and right leg.

The signal flow diagram is presented in Figure 4.

The entire device is powered only through USB 2.0

5V, therefore the signal from load cells has to be

amplified with low voltage analogue amplifiers

TL084. A central unit ATmega328 controller is

equipped with a built-in analogue-digital converter

with a resolution of 10-bit, and six channels. In order

to light up adequate number of LED on a display in

front of the patient, shift registers are used. Fast

switch ing between illuminated LEDs gives the

possibility to control nine independent LEDs with one

output from the ATmega unit. Collected data includes

amount of time spent in order to execute an exercise

as well as amount of force applied to the device by

the patient. All together is transmitted through UART

via USB to a personal computer with a 10 Hz

frequency.

Figure 4: A diagram of information flow of the force

applied to the device.

2.3 LED Display and Control Panel

Significant difficulties in reading and understanding

complex information by patients with vestibular

system damage within Cerebral Palsy in cerebellum

(Ojoga and Marinescu, 2013) implies that the

information about the forces applied to the device has

to be shown in a simple and informative way. A

numerical display with values changing in respect to

the force applied to the device was tested, but the

patient could not understand the readouts during the

A New Rehabilitation Device for Balance Impaired Individuals

exercises, thus there was no biofeedback. Any results

from these tests are inconclusive.

Authors decided, that a solution to this problem

would be a LED display (shown in Figure 5)

composed of five bars, each representing one sensing

surface (presented earlier in the Figure 1 on the right).

Applying force to a sensing surface results in lighting

up a correlated LED bar. The more force is applied,

the more LEDs in the bar are illuminated. Force

applied to the side and rear pillows is shown in green

bars on the sides and at the bottom of the display.

Body weight distribution between left and right leg is

shown as two yellow-red tapering bars. When the

force is equally distributed all yellow LEDs are

illuminated. As soon as the weight is transferred to

one of the sides, more LEDs on this side is

illuminated, whereas the bar showing readings from

the opposite side is diminishing.

Such a display setting allows to understand the body

weight distribution at a glimpse of an eye, which

allows to see the reactions between the patient and the

device. Moreover it is very easy to set a goal for an

exercise and the patient clearly sees if he is succeed-

ing or not. This lets the patient put more attention to

execute the task rather than focusing on reading the

information from the display.

The simplicity of the display is achieved at the

expense of readings resolution. As it is mentioned by

a number of authors (Matyja, 2012)(Matyja and

Domagalska, 1998)(Batra et al., 2011) the asymmetry

of movements and reflexes is crucial in the

examination of motor skills. For that reason a system

that visualizes movement asymmetry rather than

precise forces applied to the device by the patient

gives better feedback information and is more

informative. Independent to what is shown to the

patient on the display, raw data is sent directly to the

PC (Personal Computer) for further analysis.

Figure 5: LED display mounted in front of the patient with

LED bars and button "P" indicated.

2.4 Sensor Calibration

The spasticity of some muscle groups in CP can cause

the inability to equalise the pressure of left and right

leg on the base. It does not, however conclude for the

inability of maintaining stability. The device allows

calibrating the pressure range for sensing surfaces, so

that it is tailored to the patient and their abilities.

The microcontroller used in the control system

enables adjustment of the device performance to the

patients of various weight and independently of how

much force they apply to the device when leaning on

the side or rear pillows. This feature has been

achieved by implementing a sensor calibration.

The calibration is based on measuring maximum

force applicable by the patient to the specific sensing

surface and scaling the range of the readings

accordingly. The result is, that the first LED and all

LEDs up to the last one of a selected bar on the

display are illuminated when minimum and

maximum force is applied respectively.

Each time the device is switched on, an average

reading during first three seconds is considered a null

force reading for each sensor. During that time the

patient cannot be in the device. The system enters the

operational mode after three seconds and the patient

may enter the device.

At this point the patient is asked to apply as much

force to this sensing surface as he is capable. The

system picks maximum value read during that time

and sets it as the maximum value readable for this

sensor. This information is mapped to eight levels, so

that the LED bar can be lit properly. Afterwards the

system goes back to operational mode and in order to

calibrate another sensor, this procedure has to be

repeated. The presented procedure does not influence

readouts sent to PC via USB. The calibration refers

only to the LED display resolution and range of

signals.

3 CASE REPORT

The device has been designed and built in close

cooperation with a patient and his family. The patient

was given the device to test it for the duration of six

weeks. During this time he was asked to work with

the assignments and complete each of them every

day. The two assignments are discussed in the

3.2 Subsection. He was also asked to carry out the

calibration procedure for null weight and maximum

weight for all the sensors each day before he did the

assignments.

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

3.1 Subject

The patient is a 16 y.o. boy with Cerebral Palsy in the

cerebellum diagnosed when he was 20 months old.

Based on MRI tests his palsy affected multiple areas

of his brain: extrapyramidal nerves in his brain stem,

cranial nerves, the diencephalon and the pyramidal

nerve system of cortex. More precise examination is

presented in a case study (Koczyk, 2015) which

discusses EEG, MNRI and BAEP tests results. The

patient is classified with GMFCS in between level IV

and level V. His motor abilities allow him to sit and

keep the torso upright while sitting. He is also capable

of keeping his head upright and execute reaching

tasks although dysmetria, tremor and dyssynergia are

evident. He has strong astigmatism and difficulties in

controlling the eyes movements when tired.

The patient moves around his home on all fours.

This is the only way he can translocate from one place

to another in a safe environment on his own. Outside

his home he is able to move in his wheelchair but only

with an assistance. He is unable to stand still without

any support but he can hold on to furniture and rise as

well as stand holding on to it. Making steps is

challenging for the patient even when he is using

external support. An assistant is necessary to walk

him around the house but then the patient leans on the

assistant giving him full control over the equilibrium.

Due to the spasticity of his legs' rear muscles e.g.

plantar flexion is present permanently. Hence, the

patient struggles to maintain an upright standing

position because he has to flex the knees, keep his

weight on the toes and rotate his hips backwards

achieving anterior pelvic tilt. The result of such a

posture is that the patient does not bring his hips in

contact with the front pillow of a stander.

3.2 Testing Procedure and Exercises

Scenarios

The device was located in a room, where the patient

could easily enter the stander. In case the patient

requires assistance, the assistant helps him walking

into the device, locking the rear clasp and turning on

the device. The patient performs twice two exercise

scenarios each day, what results in approximately

30 minutes of standing in the device daily. In case the

patient is sick or unable to perform his duty that day

he is asked to complete his task the other day

additionally to the scheduled exercises for that day.

Exercise scenarios:

1. Illuminate specific number of LEDs related to the

body weight distribution – goal is to shift the weight

in order to reach a given number of LEDs randomly

chosen from two tapering LED bars as discussed in

Subsection 2.3. In order to complete the exercise the

given LED has to be kept illuminated for five

consecutive seconds. The patient has 120 seconds to

succeed. Otherwise the system automatically ends

this round of an exercise. The exercise consists of two

rounds.

2.Illuminate specific number of LEDs related to the

body weight distribution while refraining from

illuminating the LEDs related to the side or rear

pillows – the patient has to execute the same

algorithm as in the first exercise but this time leaning

on the device results in lighting up the green bars on

the LED display while the LED bars related to the

standing platform are switched off. The patient is

obliged then to diminish the pressure on the side

pillows and find the correct weight distribution again

During the exercises the patient was asked not to

grasp the stander with his hands. To facilitate that, he

was given two cylindrical objects to hold.

For the sake of completing the second exercise a

feasible threshold of force applied to the pillows has

to be set, otherwise a person with stability deficits

will not be able to finalize this task. After some trial

tests the threshold value was set as 40% of the

maximum force applicable to the pillows for this

particular patient. This is in line with a very important

aspect of the gamification, which is the win-lose ratio.

In order to keep the patient actively involved in the

game, hence in the rehabilitation process, a game

designer has to maintain the game challenging for the

patient but at the same time plausible to succeed.

All the tests and exercises were performed in a

room temperature varying between 21C and 30C.

To achieve results independent from the temperature

variations, load cells readouts were verified in respect

to the temperature changes. This was done thanks to

the sensors' null calibration feature, which is

essentially the readout of a non-loaded sensor. This

data was collected from each day the device was used

and stacked with room temperature changes in order

to avoid any temperature influence onto the results.

3.3 Results and Discussion

During six week time period the patient performed

the first exercise scenario 154 times and the second

exercise scenario 113 times. It took him 23 days to

complete all these exercises which was the result of

patient's frequent respiratory infections. It was

notable that after each break (the longest was five

days in a row) he struggled with achieving a correct

body weight distribution.

All the results gathered include: date and time;

current number of exercise (1 or 2); current round

A New Rehabilitation Device for Balance Impaired Individuals

number (1 or 2); current LED number illuminated as

a target for this exercise, which together with the

round number refers to the requested body weight

distribution (1 through 8); amount of time spent on

completing the exercise (in milliseconds); amount of

time spent with a correct body weight distribution (in

milliseconds); and for the second exercise, the

amount of time the patient applied force to the pillows

higher than the threshold (in milliseconds).

Data acquired during tests is divided based on the

side to which the patient had to transfer the body

weight as well as the amount of body weight applied

to the particular side. There are four weight levels:

63% or less, 63 − 75%, 75 − 88% and 88 − 100% of

body weight transferred to a particular leg. Such a

division is done in order to identify if the patient has

a problem with acquiring a specific body weight

distribution.

The first quality criteria for both of the exercises

is T

which is the amount of time necessary for the

patient to achieve and maintain the desired body

weight distribution for five consecutive seconds. The

exemplary results showing T

gathered for exercise 1,

when the task was to keep the body weight evenly

distributed between left and right leg with maximum

63% of the weight on one leg in a timespan of 42 days

is shown in Figure 7, top. A trend function:

y=-1.15+68 with coefficient of determination

=0.15 is embedded into the plot.

For both legs and all four body weight distribution

levels, the trend function during the six weeks

training time resulted in a negative slope constant

which is presented in Table 1 as a linear trend

function of subsequent plots. Additionally, the

percentage of decrease of time necessary to complete

the exercise ΔT

for both legs combined based on a

trend function is also included. A significant decrease

of time necessary to succeed in this task as the

training proceeded can clearly be noticed. The

activity which took approximately 70 seconds, after 6

weeks of training was more often achieved in only 20

seconds. Also, at first glance, an improvement in case

of 63 − 75% level is least impressive, but in this case

was the lowest out of all four load cases which is

further shown in Table 2. This shows, that for tasks

which are generally easier for the patient, the

improvement is less notable.

The second quality criteria for both of the

exercises is T

which is the amount of time during

which the desired body weight distribution was

achieved but not necessarily maintained for

consecutive five seconds. The T

in its nominal value

does not include that the overall time of completing

the task may change reducing possibility to achieve

higher T

values. Hence, T

is further considered a

percentage value of a T

time, where:

𝑇

𝑠%

𝑇

𝑠

𝑇

𝑎

∙ 100%

(1)

Figure 6: On the top - time necessary to complete the task

in exercise 1 - T

; on the bottom - percentage of time

necessary to complete the exercise during which patient

achieved desired body weight distribution but not

necessarily maintained it for consecutive 5 seconds - T

Data presented for case where the patient's weight had to be

distributed evenly with up to 63% of weight on one leg.

Table 1: Linear trend function y=ax+b of T

and T

function of days (t) the exercises were performed for

subsequent body weight distribution levels. LL stands for

the left leg, RL stands for the right leg, and TO stands for

both legs combined. Additionally, ΔT

is a percentage

decrease of time necessary to complete the exercise for both

legs combined based on a trend function T

(TO).

100-88% of

weight on

one leg

88-75% of

weight on

one leg

75-63% of

weight on

one leg

up to 63%

of weight

on one leg

(LL)

-0.37t+31

-2.09t+81

-0.44t+44

-0.99t+51

(RL)

-1.89t+129

-2.0t+89

-0.1t+35

-1.2t+79

(TO)

-2.15t+113

-2.0t+87

-0.27t+39

-1.15t+68

(LL)

-1.91t+105

0.91t+25

0.06t+48

0.74t+33

(RL)

0.66t+8.95

1.3t+24

0.17t+49

0.68t+27

(TO)

0.52t+23

1.14t+25

0.1t+49

0.71t+30

ΔT

[%]

An exemplary result for T

is presented in Figure 7,

bottom, where data was gathered for the first exercise

with body weight evenly distributed between left and

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

right leg with maximum 63% of the body weight on

one leg in a timespan of 42 days. Here the trend

function y=0.71+30 with coefficient of

determination R

=0.26 is embedded into the plot.

Calculated trend functions of T

for each leg

separately as well as for both legs combined and for

all of the four body weight distribution levels are

presented in Table 1.

In most cases T

rises as the patient progresses

with the training with an exception for 88 − 100% of

body weight transferred to the left leg. This could be

explained by a very low T

values for this load case

(fastest was 11 seconds and slowest was 25 seconds).

Having in mind, that for 5 seconds patient has to

maintain the correct balance anyway, achieving better

results becomes very challenging. Another aspect is

that T

is inversely proportional to the T

values.

This implies that decreasing T

readouts with

consecutive training days (what evidently occur) will

always result in T

getting higher. Therefore, even

though T

shows improvement over time, changes of

nominal time spent on standing with proper body

weight distribution in most cases were found not to

have an apparent trend nor change significantly.

The combined results of T

from six-weeks

training period are calculated in order to achieve an

overall view of patient's abilities. Mean time value T

is presented in Figure 8 and Table 2 for each body

weight distribution level separately what provides

information about posture dissymmetry.

It is notable, that tasks involving shifting body

weight to the right leg are much more difficult for the

patient. Also, the more weight is to be transferred to

the right leg, the more difficult the task becomes.

Standard deviations of up to SD=22 for the body

weight shifted mainly onto the left leg and up to

SD=43 for the body weight shifted mainly onto the

right leg are the result of frequent failures in body

weight distribution task in 120 seconds time limit. For

the task of transferring body weight onto the left leg

patient did not succeeded only twice (for 87 times the

task was assigned) whereas in the case of transferring

body weight onto the right leg it happened 11 times

throughout all 67 times the task was assigned.

Observation of the patient while he was in training

though revealed that most of the times he did not

succeeded happened when he was distracted by a

nearby discussion or was anxious to do something

else in the time he was exercising. It is disputable

therefore if a drop-out ratio should be considered as

an important parameter for this study.

Restriction inflicted in the second exercise by the

presence of side and rear pillows significantly

extended the time necessary for the patient to achieve

and maintain the desired body weight distribution for

Figure 7: Mean time value 𝑇

𝑎

necessary to complete the

task in exercise 1 - data divided for four levels of body

weight distribution and presented for each leg separately as

well as for both legs combined (total).

Table 2: Data referring time value 𝑇

𝑎

necessary to complete

the task in exercise 1 presented for both legs in total as well

as for the left and right leg separately in all four body weight

distribution cases. LL stands for the left leg and RL stands

for the right leg.

Percentage

of weight

on one leg:

100-88%

88-75%

75-63%

63-50%

𝑇

𝑎

total [s]

𝑇

𝑎

(LL) [s]

𝑇

𝑎

(RL) [s]

five consecutive seconds (T

). Mean time value 𝑇

𝑎

presented in Figure 9 as well as in the Table 3. Here

the differences in shifting body weight over left or

right leg are not so obvious and improvement in 𝑇

𝑎

only notable in respect to amount of body weight

shifted to the side. This suggests, that it was more

difficult for the patient to stand without leaning on the

pillows than it was to shift the body weight to a

correct position.

The task of transferring the body weight properly

to the left and right leg in the second exercise was

assigned 57 and 56 times respectively. In case of the

left leg the patient did not manage to achieve success

in the 120 seconds time limit six times and for the

right leg it was eight times. This resulted in standard

deviations of 𝑇

𝑎

up to SD=38 for the left leg and up to

SD=54 for the right leg results. Again, the drop-out

Case were mostly inflicted by external disturbances.

A New Rehabilitation Device for Balance Impaired Individuals

Figure 8: Mean time value 𝑇

𝑎

necessary to complete the

task in exercise 2 - data divided for four levels of body

weight distribution and presented for each leg separately as

well as for both legs combined (total).

Table 3: Data referring time value 𝑇

𝑎

necessary to complete

the task in exercise 2 presented for both legs in total as well

as for the left and right leg separately in all four body weight

distribution cases. LL stands for the left leg and RL stands

for the right leg.

Percentage

of weight

on one leg:

100-88%

88-75%

75-63%

63-50%

𝑇

𝑎

total [s]

𝑇

𝑎

(LL) [s]

𝑇

𝑎

(RL) [s]

The results in case of the second exercise are

much more consistent and show very similar 𝑇

𝑎

values for all four cases where the weight had to be

shifted onto the left leg. We noticed also very little

improvement of T

and T

parameters over the six-

weeks training time in case of second exercise. This

is probably due to a very high difficulty the patient

had with the task of shifting the body weight while

simultaneously caring not to lean on the device.

4 CONCLUSIONS

We have presented a new rehabilitation device for

people with stability deficits. Discussed assisting

device ameliorates keeping the equilibrium by

training the body weight shifting in coronal plane

while visualizing patient's current body weight

distribution and forces inflicted onto the pillows

keeping him in an upright standing position. The

device is designed to be used by the people who may

not necessarily be able to keep the upright standing

position on their own but have muscle strength

sufficient to maintain an upright position with use of

a stander.

A set of exercises is proposed in order to train the

stability with use of the device. They involve shifting

patient's body weight so that it matches requested

body weight distribution while avoiding leaning onto

the side and rear pillows of the stander.

Forces inflicted onto the device are shown in form

of LEDs being illuminated as the patient applies more

force to the device or shifts the body weight. Such

functionality gives physiotherapists real time

information about body weight distribution while the

patient is performing exercises in an erect position.

Moreover it provides a wider knowledge about the

compensation the patient may need in order to keep

the upright standing position. It is also considered that

the device may play a significant role in an evaluation

process of patients' stability. Performing proposed

exercises allows to estimate the capabilities of the

patient.

For the patient, on the other hand, important

aspect of using the device is that, with the instant

biofeedback about the body weight distribution, the

patient may adjust his position by himself. This can

be done even if the person's reflexes are stunted or the

nervous system doesn’t work correctly.

Thanks to the presented calibration process the

measurement system is versatile making the device

suitable for work with patients of any weight or

height.

The device was tested with an Ataxic Cerebral

Palsy patient during a six-week time period. During

this time the patient performed 267 times the

exercises described. It means the patient spent

approximately 30 hours in the device safely standing

and performing exercises, what is already a great

success. Detailed results given by the system were

discussed with physiotherapists and several

conclusions were drawn:

1. The patient tends to grasp the stander. This was

notable mainly when the patient tried to produce

dynamic movements with relatively high

acceleration. Holding to a steady stander gave the

patient more confidence in movements and requested

tasks were performed faster. If the aim is to induce a

correct muscle tonus while in erect position, then

body dynamics should be drawn from body mobility

and control rather than external support.

2. The randomness of body weight distribution

requests should be introduced after repetitive

movements requests. This is due to the fact, that the

BIODEVICES 2019 - 12th International Conference on Biomedical Electronics and Devices

integrity of reflexes of ataxic patients is not

developed properly and they struggle to control their

body. Giving them a task, to shift the body weight and

ask to pay attention to leaning on the device is far too

much for them at the beginning of the training.

3. Physiotherapists struggled with too much

information from the system. It is considered for

future works therefore to limit the possible body

weight levels to: left, centre, and right, rather than

eight separate levels.

In future works we would like to develop more

sophisticated environment, where the patient is

motivated to pursue best results based on gratification

system embedded in a computer game.

ACKNOWLEDGEMENTS

The authors would like to express their gratitude and

thanks to Jan Koczyk and his family for continuous

support of this project and involvement in testing.

Valuable assistance during the project realization was

provided by physiotherapist Katarzyna Sęk. Further,

we thank Barbara Karlik for careful reading of the

manuscript and language suggestions.

REFERENCES

Batra, M., Sharma, D.V.P., Batra, D.V., Malik, D.G.K.,

Pandey, D.R.M., 2011. Postural Reactions: An

Elementary Unit for Development of Motor Control.

Disabil. CBR Incl. Dev. 22, 134–137.

Błaszczyk, J.W., 2004. Biomechanika Kliniczna:

podręcznik dla studentów medycyny i fizjoterapii.

Wydawnictwo Lekarskie PZWL, Warszawa.

Drużbicki, M., Przysada, G., Rykała, J., Podgórska, J.,

Guzik, A., Kołodziej, K., 2013. Evaluation of the

effectiveness of selected scales and methods used in the

assessment of gait and balance after a cerebral stroke.

Przegląd Med. Uniw. Rzesz. i Nar. Inst. Leków w

Warszawie.

Gałęcki, S., Walasik, M., Rokicki, R., Sikorska, K.,

Dudkiewicz, Z., 2013. Effectiveness of rehabilitation

and exercises on static-dynamic parapodium with

biofeedback in relation to body balance in patients after

ischaemic stroke. Kwart. Ortopedyczny 2013, 314–325.

Gilfoyle, E.M., Grady, A.P., Moore, J.C., 1990. Children

adapt : a theory of sensorimotor-sensory development.

Slack Inc.

Horak, F.B., 2006. Postural orientation and equilibrium:

What do we need to know about neural control of

balance to prevent falls? In: Age and Ageing. pp. ii7-

ii11.

Hurło, L., Kowalski, M.I., 2003. Zaburzenia postawy ciała

w wieku rozwojowym. Uniwersytet Warmińsko-

Mazurski w Olsztynie, Olsztyn.

Koczyk, M., 2015. Case study of Janek: MNRI for a Child

with Severe Cerbellum Damage. Reflexes Portal to

Neurodev. Learning., Svetlana Masgutova Educational

Institute, A Collect. Work. 271–280.

Kuczyński, K., Podbielska, M., Bieć, D., Paluszak, A.,

Kręcisz, K., 2012. Podstawy oceny równowagi ciała:

czyli co , w jaki sposób i dlaczego powinniśmy

mierzyć? Acta Bio-Optica Inform. Medica. Inżynieria

Biomed. 18, 243–249.

Livingstone, R., Paleg, G., 2016. Measuring Outcomes for

Children with Cerebral Palsy Who Use Gait Trainers.

Technologies 4, 22.

Matjacić, Z., Hesse, S., Sinkjaer, T., 2003.

BalanceReTrainer: a new standing-balance training

apparatus and methods applied to a chronic hemiparetic

subject with a neglect syndrome. NeuroRehabilitation

18, 251–259.

Matyja, M., 2012. Neurorozwojowa analiza wad postawy

ciała u dzieci i młodzieży. Wydawnictwo Akademii

Wychowania Fizycznego w Katowicach, Katowice.

Matyja, M., Domagalska, M., 1998. Podstawy

usprawniania neurorozwojowego według Bert i Karela

Bobathów. ląska Akademia Medyczna, Katowice.

Michalska, A., Dudek, J., Bieniek, M., Tarasow-Zych, A.,

Zawadzka, K., 2011. The application of the balance

trainer parapodium in the therapy of children with

cerebral pals. Fizjoterapia Pol. 11, 273–285.

Ojoga, F., Marinescu, S., 2013. Physical Therapy and

Rehabilitation for Ataxic Patients. Balneo Res. J. 4, 81–

84.

Paleg, G., Livingstone, R., 2015. Outcomes of gait trainer

use in home and school settings for children with motor

impairments: A systematic review. Clin. Rehabil. 29,

1077–1091.

Palisano, R., Rosenbaum, P., Bartlett, D., Livingston, M.,

Walter, S., Russell, D., Wood, E., Galuppi, B., 2007.

Gross Motor Function Classification System Expanded

and Revised. Ref. Dev Med Child Neurol 39, 214–223.

Shanahan, C.J., Boonstra, F.M.C., Cofré Lizama, L.E.,

Strik, M., Moffat, B.A., Khan, F., Kilpatrick, T.J., van

der Walt, A., Galea, M.P., Kolbe, S.C., 2018.

Technologies for Advanced Gait and Balance

Assessments in People with Multiple Sclerosis. Front.

Neurol. 8, 708.

Woollacott, M., Shumway-Cook, A., Hutchinson, S., Ciol,

M., Price, R., Kartin, D., 2005. Effect of balance

training on muscle activity used in recovery of stability

in children with cerebral palsy: A pilot study. Dev.

Med. Child Neurol. 47, 455–461.

Zafeiriou, D.I., 2004. Primitive reflexes and postural

reactions in the neurodevelopmental examination.

Pediatr. Neurol.

A New Rehabilitation Device for Balance Impaired Individuals