Space Perception by Acoustic Cues Inﬂuences

Auditory-induced Body Balance Control

Shinichi Yamagiwa

and Naka Gotoda

and Yuji Yamamoto

Faculty of Engineering, Information and System, University of Tsukuba/ JST PRESTO, Tsukuba , Japan

Department of Sports Science, Japan Institute of Sports Sciences, Tokyo, Japan

Research Center of Health, Physical Fitness and Sports, Department of Psychology and Human Developmental Sciences,

Nagoya University, Nagoya, Japan

Keywords:

Acoustic, Visibility, Blind, Vection, Body balance.

Abstract:

The auditory-induced illusion called vection has been investigated for decades. However, it is not conﬁrmed

how the illusion affects to body balance. Especially, during a dynamic activity such as walking, it has not

conﬁrmed whether any clear effect to the body balance control appears or not. This paper focuses on inves-

tigating the effect of vection during walking. Especially this paper will discuss the space perception induced

by acoustic stimuli that indicate the directions. The authors of this paper measured the response time from the

acoustic cue to body balance control and the rotation amount using a small sensor system with accelerometer,

magnetic and gyro sensors. According to the experiments with sight/blind participants from young to old ages,

requesting them to walk to the directions of the acoustic cues with/without sight, the authors conﬁrmed a close

relation between the vision and the auditory of human during a dynamic activity.

1 INTRODUCTION

We are surrounded by sounds in our life such as the

winds in a green ﬁeld, whisper in a silent room, noise

in a hard trafﬁc between high buildings, chat of peo-

ples in a restaurant and so on. A sudden large sound

lets your head turn up and see its direction. In a fog,

you would try to know the circumstance around you

hearing sounds and then you try to know if any danger

is coming to your side or not. These human natural

behaviors are coming from potential ability of space

perception (Blake and Sekuler, 2005) with auditory

and visual information acquired from the environ-

ment. However, the certainness of the space percep-

tion from the auditory information includes many un-

known cognitive responses, especially the inﬂuence

of the auditory related to the body balance control

during active movements like walking.

Several advanced research results have been re-

ported to show that how cognitive behavior would

happen when the body balance is inﬂuenced by au-

ditory cues given to human. There exists a cognitive

illusion called vection (Valjamae, 2009) that causes

a misunderstanding of changing posture by the au-

ditory cues which is experienced in an environment

with multiple speakers surrounded in a human (Lack-

ner, 1978). However, it is not known how auditory

cues inﬂuence the space perception during dynamic

body balance control. For example, how correct does

a human follow auditory cues with walking? How

fast is the response from the cues to the body balance

controls? And how the vision inﬂuences to the body

balance when the different types of auditory cues are

given?

This paper focuses on these unknown human na-

ture responses regarding the dynamic body balance

control using auditory cues that changes the sound

pan directions among left, center and right. To in-

vestigate the inﬂuences by the auditory cues given to

human, we categorized two types of cognitive behav-

iors. The ﬁrst one is the interference to the body bal-

ance control potentially affected by the cues. Another

is the interference to the human body balance control

guided by the cues when he/she chases the cues. We

focus on the latter one by performing experiences at

walking situation giving acoustic cues that the sound

pan is changing suddenly or smoothly among right,

center and left randomly. In the experiments, we ap-

plied the auditory cue patterns to congenital and ac-

quired blindness and healthy sighted (with/without

eye mask) participants to compare the cognitive be-

haviors from the acoustic cues to the body balance

Yamagiwa S., Gotoda N. and Yamamoto Y..

Space Perception by Acoustic Cues Inﬂuences Auditory-induced Body Balance Control.

DOI: 10.5220/0004601200300040

In Proceedings of the International Congress on Sports Science Research and Technology Support (icSPORTS-2013), pages 30-40

ISBN: 978-989-8565-79-2

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

control during walking situation.

According to the experiments, three main ﬁndings

are resulted in this paper: 1) space perception between

blindness and sighted people differs. Especially, con-

genital blindness people have a keen space perception

from the auditory cues used as the guide for the walk-

ing direction. 2) Patterns of sound pan as the acoustic

cues induce different body balance control behaviors

in the aspects of response speed and rotation amount

of body. The sudden change of the sound pan in-

duced fast and large balance control. Finally 3) vision

of sighted or acquired blindness people is combined

to control their body balance. They had performed

faster to response against the acoustic cues without

eye mask than invalidating the vision with eye mask.

In the next section, we introduce the advanced re-

search results related to our experiments. Section 3

will explain the methods of our experiments and pro-

pose hypotheses expected from the experiments. Sec-

tion 4 will explain the results of the experiments and

will show proofs for our hypothesizes. Finally we will

discuss the experimental results, the ﬁndings and will

conclude this paper.

2 RELATED WORKS

The relationship between acoustic cue and body bal-

ance control has been investigated since 1970’s.

The early investigations were related to the

auditory-induced illusion (Lackner, 1977)(Lackner,

1978)(Lackner, 1983). It is well known experiments

using rotating sound source with 10-20Hz sound pan

change of each speaker. Sitting in the environment,

the participant can percept that either ”I am moving”

or ”Surround is moving”. Extending the illusory

phenomenon, researchers investigate the posture

perception with multiple rotating sound sources.

Following the rotating direction of the sound, the

participant percepts illusory posture movement

(actually posture does not change).

On the other hand, the clinical approach reports

an experiment that gives electrical stimuli to vestibu-

lar organ of the human participants. It results that the

stimuli absolutely affects to his/her body balance con-

trol (Aoki et al., 2000). It is now very hard to perform

the experiment with the vestibular stimuli due to the

ethical problems. Therefore, clinically the participant

is a patient who is in hospital to aid auditory-related

diseases.

When we consider applications of the vection or

the vestibular stimuli, the biofeedback application is

one of the interesting directions for attractive utiliza-

tions. The biofeedback is a trend word used in the

application area of the auditory-induced body bal-

ance control. For example, (Dozza et al., 2006), (Gi-

ansanti et al., 2009) and (Brunelli et al., 2006) have

reported the body balance control using a feedback

system to keep a vertical standing posture. We think

these research projects are going to the frontline of

the research regarding auditory cues with sound pan

changes that affect the right and the left bias of pos-

ture. The research is focused on the static posture

control with the healthy sighted, the blindness, and

some Parkinson’s disease participants. However, we

are going to focus on the body balance control against

auditory cues of the blindness and the sighted people

under a walking situation. Under such an active en-

vironment, the mechanisms of the auditory-induced

body balance control has not been clearly explained in

the previous researches because the experiments had

been focused on illusions occurred at a static posture.

Regarding the Parkinson’s patients, a biofeedback

system called Walkmate has been developed by the

researches (Miyake, 2009) (Hove et al., 2012). The

Walkmate generates rhythmic pacing sequences to the

Parkinson patients using nonlinear limit cycle oscil-

lators. The rhythmic pacing timings are generated

by feedforwarded the timings taken from a pressure

sensor put in the patient’s foot. The system induces

self-walking from the potential characteristics of the

disease. Therefore, this type of application using au-

ditory cues during active body balance control can be

treated as one of special cases. Therefore, we want to

try to ﬁnd how the auditory and the body balance con-

trol of physically healthy persons are involved during

active movements like walking.

As we discussed above, it is not well-known that

the cognitive response from vection to the body bal-

ance regardingphysically healthy people. The vection

is recognized as an internal psychological and cogni-

tive perception. Therefore, the body balance might be

inﬂuenced by the vection in a dynamic body balance

control such as during walking, running or hard sports

activity. However, the research regarding the vection

is still under progress in the static body balance con-

trol or self-cognitive issue due to the complex rela-

tion to the sight. Thus, although the space perception

of human in a dynamic body balance control would

be related to the vection and the vision, it is not ex-

plored that the dynamic body balance control relates

to the auditory stimuli. Finally the new ﬁndings will

be utilized in sports trainings, new society design to

eliminate dangerousness induced by acoustic illusion,

and engineering applications such as a novel rehabili-

tation system using a sound guide.

In this paper, we focus on exploring inﬂuences

to the space perception regarding the dynamic body

SpacePerceptionbyAcousticCuesInfluencesAuditory-inducedBodyBalanceControl

−

)(

Sigmoid function

changing in a second

Blue wave represents

the right sound pan.

Red wave represents

the left sound pan.

Sigmoid

Center Heviside Heviside

Sigmoid

(a) (a)

Figure 1: Sound pan change waves when the Heviside and

the Sigmoid function are applied. The blue part is a sound

pan from right side. The red part is the one from left side.

(a) shows the pan change with the Sigmoid function. (b)

is the one with the Heviside. The center pan is shown in

(c). The sound volume was adjusted to be felt in the same

among (a), (b) and (c). Therefore, the range of sampling

data during the center pan is half of the right or left sound

pan.

balance control induced by the auditory cues. We

have questions for the perception: 1) If human can

follow auditory cues and walk the expected direction

with/without vision or not. 2) If the sight is not related

to the space perception when we follow the acoustic

cues or not. And 3) if we have any difference in the

response time from the acoustic cues to the dynamic

body balance with/without the sight or the sound type

of the cues or not. Thus, we report the experiments

to explore answers of the questions above and discuss

the semantics of the results.

3 METHODS AND HYPOTHESIS

3.1 Methods and Experimental Setup

To investigate the space perception regarding the dy-

namic body balance control induced by the auditory

cues, we use a quite simple method following the

steps; 1) a walking participant hears auditory cues

that change the sound pans among right, center and

left, 2) the participant chases the sound pan direction,

3) congenital, acquired blindness and sighted people

with/without an eye mask perform the same steps of

1) and 2). According to the steps, we observe the re-

sponse time from the beginning of the cues, the rota-

tion amount of body and the walking direction.

We prepared the experimental environment for the

methods above in gymnasiums. Let us explain the

Table 1: Participants attended in the experiment.

Participant Visibility Sex Age

A Congenital blindness Male 27

B Aquired blindness Male 41

C Aquired blindness Female 38

D Sighted Female 37

E Sighted Female 27

F Sighted Female 24

G Sighted Female 20

H Sighted Male 58

I Sighted Female 40

J Sighted Male 62

Y axis

Z axis

X axis

Piggybacking the

sensor system

held in a jogging

holster

Figure 2: Magnetic sensor axis of the integrated sensor sys-

tem piggybacked by participants during the experiment.

materials, the places and the participant selection con-

cept listed below:

• Place for Experiment: We have prepared empty

and silent gymnasiums in Kochi University of

Technology and in the Blind School in Kochi Pre-

fecture in Japan where the walking length by the

participant almost equals to the one of the goal-

to-goal in a basketball court. The air temperature

in the gymnasiums was about 10− 20

◦

C without

any wind during autumn season. The ﬂoor made

of wood keeps ﬂat without inclination.

• Auditory Cues used in the Experiment: The sound

pattern including the auditory cues consists of

three types of tones. The ﬁrst one is the starting

trigger for the participants that indicate the start-

ing point of the experiment. It is constructed by

880Hz sin wave with 16 bit sound resolution. This

trigger sound is outputted in the center pan. The

subsequent sound stimuli organize a right-to-left

or a left-to-right pan change constructedby 440Hz

sin wave with 16 bit sound resolution as illus-

icSPORTS2013-InternationalCongressonSportsScienceResearchandTechnologySupport

The starting cues

Gyro sensor also can

acquire hand up and down.

Acoustic sequence

used for a participant.

Gyro sensor Z axis

Magnetic sensor Z axis

Magnetic sensor X axis

Magnetic sensor Y axis

Approximated response time

synchronized with the video.

These data are

synchronized with the video.

Baseline for

integral of rotation

Rotation timings are matched to the response times.

Rotation amount calculated by

integral based on the baseline

(a) Rotation amount

measurement

(b) Response time

measurement

Figure 3: Measurement methods of the approximated response times and the rotation amounts at the body balance controls

against the acoustic cues. The response time is measured from the beginning of the sound pan recognized by the magnetic

sensor data. The rotation amount is measured by the integral of the gyro sensor data corresponding to the changing timings

of the response time.

trated in Figure 1, which dynamic range varies in

−16384 < s < 16384 where s is the sample value.

These indicate the walking direction to the partic-

ipant. The pan change uses two functions: Hevi-

side and Sigmoid. The Sigmoid function changes

the sound pan smoothly. The changing time from

one side to another takes a second and the same

sound pan is also kept for a second (Figure 1(a)).

Totally during two seconds one of the sound pans

is outputted. But the Heviside one changes the

sound pan suddenly. The length is two seconds

for the same sound pan (Figure 1(b)). The last

type is the center sound pan that is constructed

by 440Hz with 16 bit sound resolution, which the

dynamic range is varied in −8192 < s < 8192 for

the sample value s (Figure 1(c)). The (a), (b) and

as the auditory cues for guiding the direction. The

participant follows the direction of the sound pan.

We just requested them to turn to the direction as

they recognized. In the case of the center sound

pan, we just requested the participants to walk in

straight. We have prepared ﬁve sequences with

the auditory cues for the experiment, in which

the total time of a sequence is 56 seconds. Each

participant tried these ﬁve sequences, where the

sighted participant wears an eye mask for the ex-

periments. Finally the sighted participant also at-

tended in an additional 6th experience without the

eye mask using another acoustic cue sequence.

• Measurement of the Body Balance Control dur-

ing the Experiment: A sensor system originally

equipped with a 400 dps 3D gyro sensor (In-

venSense IDG-400 × 2), a 2.5G 3D accelerome-

ter (Freescale MMA7261QT) and a 3D magnetic

sensor (Aichi Steel AMI602) was introduced to

the experiment. All sensors can achieve up to

200Hz sampling rate and have 12bit resolution

(Texas Instruments TLV2553IPW) for each data.

Therefore, the experiment applies 5 msec to the

sampling time resolution. The data of accelerom-

eter illustrates the timings of legs’ movement dur-

ing the walking of participant. The gyro sen-

sor measures the rotation amount of participant’s

body using the absolute value from the channel of

the vertical axis against body. The magnetic sen-

sor outputs the direction data of the body that rep-

resents the starting or the ending timings of rotat-

ing the body against the auditory cues. Data mea-

surement timings from all axes of all the sensors

are synchronized by the microcontroller (Microb-

SpacePerceptionbyAcousticCuesInfluencesAuditory-inducedBodyBalanceControl

laze at 48MHz) implemented on a Xilinx Spar-

tan3E FPGA. If participant walking in straight,

the data from the magnetic sensor keeps the same

value.

• Participants attended in the Experiment: Totally

ten participants including a mentally healthy con-

genital and two mentally healthy acquired blind-

ness persons and seven healthy sighted persons

attended in the experiment as summarized in Ta-

ble 1. Participant wear cotton thick socks to elim-

inate cognitive effect from the ﬂoor. As shown

in Figure 2, participant piggybacked the sensor

and an iPod nano put into a small jogging hol-

ster, and heard the sound pan changes from the

iPod nano via a noise canceling headphone (Sony

MDR-NC100D). The sound volume was adjusted

as the participant comfortably and clearly can hear

the sound.

We took videos of the experiment with two cam-

eras from the back and the front of each participant.

The video was used to synchronize the data from the

sensors compared with the participant’s movement.

We noted that this method did not have enough ac-

curacy for the response time from an acoustic cue be-

cause it was impossible to record the acoustic cues

heard by the participant. Therefore, regarding the

response time from an acoustic cue to the body bal-

ance control, we request to participants to move their

hands up and down during the starting cues, and then

we synchronize the movement from the video as the

starting point of the acoustic cue sequence. More-

over, because the hand movement was recorded in the

sensor data (especially in the accelerometer data), we

synchronized the movement conﬁrmed by the video

approximately with the sensor data (Figure 3(b)). Fi-

nally we used the rate of the approximate response

time where one of the response times was used as the

base time. We explained the detail of this calculation

in the next section using the realistic cases.

On the other hand, we also focus on the rotation

amount to measure the dynamic volume of body bal-

ance controls induced by the acoustic cues. We in-

tegrate the area more than or less than the baseline

related to the behavior of the response time as shown

in Figure 3(a). This integral regions represent the ro-

tation amount induced by the acoustic cues without

time dependency.

3.2 Hypothesis for the Expected Results

Measuring the approximated response time and the

rotation amount by the experiment, we expect that

seven hypotheses can be approved as enumerated be-

low:

1. The response time differs depending on the

changing function of the sound pan: Changing

the sound pan with the Heviside function induces

faster response to body balance control than the

Sigmoid one.

2. Training the body balance control by the sound

pan change affects to the response time: The num-

ber of sequences in the experiment that a partici-

pant tried affects to his/her response time against

the sound pan change.

3. The response time differs depending on partici-

pant’s age: Age affects to the response time from

the pan change recognition to the body balance.

4. The response time of blindness people becomes

shorter than the one of the sighted: the blindness

promotes keen auditory sense and induces shorter

response time from the sound pan change to the

body balance control than the sighted.

5. There exists left or right bias with respect to the

hypothesis 1: The response time differs depend-

ing on participant’s dominant arm/leg.

6. Rotation amount induced by the sound pan change

differs between the Heviside and the Sigmoid

function: The body balance control induces larger

rotation amount when the Heviside function is

used for the sound pan change than the sigmoid

function or vise versa.

7. The center pan is misunderstood after the previ-

ous sound pan is changed from right or left: The

sound pan sequences in the order of right, left and

center or in the one of left, right and center induce

misrecognition of the center pan perceived by the

participant.

8. Vision inﬂuences the response time against the

sound pan change: When the sighted people use

their vision, their response times against the sound

pan change with the Heviside or the Sigmoid

function differs from the ones without sight.

The hypotheses 1-5 and 8 are related to the re-

sponse time against the sound pan. These hypothe-

ses can be analyzed by the approximate response time

compared with the rates calculated by the division

with the one of the compared response times. The

hypothesis 6 is analyzed by using the rotation amount

measured by the gyro sensor. It is available to eval-

uate the rates calculated by the division of the com-

paring rotation amount with the compared rotation

amount. Thus, we performed the experiment and an-

alyzed the acquired data applying the methods men-

tioned above. Thus the next section will show the re-

sults and investigate the hypotheses.

icSPORTS2013-InternationalCongressonSportsScienceResearchandTechnologySupport

4 RESULTS

4.1 Found Fundamental Knowledge

Regarding the hypothesis 2, 3 and 5, we compared

the approximated response times and the rotation

amounts among blindness/sighted, ages, sex, domi-

nant hand/leg of all participants and the number of

trials. The comparison is performed under these con-

ditions; 1) the approximated response times and the

rotation amounts are compared based on every exper-

iment in ﬁve trials, 2) we compared results among the

sound pan change patterns categorized into four pat-

terns; Heviside or sigmoid function and the sound pan

change of right-to-left or left-to-right, and 3) we com-

pared the normalized rates by the right-to-left pattern

with Heviside function. However, we were not able to

ﬁnd any relation among all data. The data is randomly

spread and independent in each participant. There-

fore, we conclude that the hypothesis 2, 3 and 5 were

not found.

According to this conclusion regarding these fail-

ures of the hypotheses, we apply new conditions be-

low to evaluate other hypotheses 1, 4, 6-8:

1. Due to the failure of the hypothesis 2 regarding

effect for multiple trials, we can use the average

data (mean) from the ﬁve trials of a participant’s

experiment for comparing with other participants’

results.

2. Due to the failure of the hypothesis 3 regarding

effect of age of participant, we can compare data

among different ages of participants.

3. Due to the failure of the hypothesis 5 regardingef-

fect of dominant direction of participant, we can

directly compare both data regarding the sound

pan change patterns of right-to-left and left-to-

right without considering participant’s dominant

hand/leg.

According to the fundamental knowledge dis-

cussed above, the comparisons of the sensor data will

be performed regarding four categories of observa-

tions: 1) the response time, 2) the rotation amount, 3)

the auditory-space perception and 4) the interference

of vision. Let us discuss the observations respectively.

4.2 Observations regarding Response

Time

First, let us focus on the difference of changing sound

patterns. We have found that the difference of chang-

ing patterns regarding sound pan change among the

Sigmoid and the Heviside functions inﬂuences the

0.5

1.5

2.5

3.5

4.5

A B C D E F G H I J

L to R (Heviside)

R to L (Heviside)

L to R (Sigmoid)

R to L (Sigmoid)

Part. L to R R to L L to R R to L

(Heviside) (Heviside) (Sigmoid) (Sigmoid)

A 1.000 0.857 1.909 1.477

B 1.000 1.069 1.716 1.626

C 1.000 1.035 1.655 1.360

D 1.000 1.166 2.238 1.767

E 1.000 0.877 3.300 2.734

F 1.000 1.229 2.393 2.039

G 1.000 0.902 4.039 3.835

H 1.000 1.009 7.717 4.614

I 1.000 1.127 2.997 2.274

J 1.000 0.899 2.228 2.491

Figure 4: Comparison of the response time rates among all

participants normalized by the 1) left-to-right sound pan

change for investigating the hypothesis 1. Any partici-

pant responses against the sound pan changes with Sigmoid

function more slowly than the one with Heviside because

the means of the sound pan change types 2), 3) and 4) be-

come larger than 1.00.

body balance control, which conﬁrms the hypothe-

sis 1. Especially we conﬁrm the truth that the sound

pan change with the Heviside function induces shorter

response time than the one with the Sigmoid.

We ﬁrst categorized the sound pan change patterns

into four types; 1) left-to-right change with Hevi-

side function, 2) right-to-left one with Heviside func-

tion, 3) left-to-right one with Sigmoid function and

4) right-to-left one with Sigmoid function. Applying

the condition 1 in section 4.1, we have calculated the

means of the response times regarding the four sound

pan change types in ﬁve time trials per participant

including all the blindness and the sighted with eye

mask. Figure 4 shows rates normalized by the type 1)

of the means of the response times categorized to each

participant. Note that A, B and C are the blindness

people and others are the sighted. Every participant

begins to turn their body at the sound pan change with

Sigmoid function in a longer response time than the

one with Heviside function. Thus, we conﬁrm that

the smooth pan change induces slow response to the

body balance control.

Next, regarding the response time to the sound pan

change with the Sigmoid function, we have found that

SpacePerceptionbyAcousticCuesInfluencesAuditory-inducedBodyBalanceControl

0.5

1.5

2.5

3.5

L to R (Heviside) R to L (Heviside) L to R (Sigmoid) R to L (Sigmoid)

Blind

Sighted

Part. L to R R to L L to R R to L

(Hev.) (Hev.) (Sig.) (Sig.)

Blindness 1.000 0.987 1.760 1.488

Sighted 1.000 1.030 3.559 2.822

Figure 5: Comparison of the rate of the response times

among blindness and sighted participants to investigate the

hypothesis 4. The means are normalized by the case of the

sound pan change from left to right with Heviside function.

Focusing on the rates of Sigmoid function, we conﬁrm that

the rates of blindness participants are smaller than the ones

of the sighted. This means that the blindness people has a

keen space perception from the sound and also can be in-

duced by the acoustic cue to the fast body balance control.

the blindness people are induced by the sound pan

change in shorter time than the sighted people when

the sighted people do not use their vision (shutting out

the vision with eye mask).

To conﬁrm this hypothesis, as shown in the Fig-

ure 5, we have calculated the means of the rates used

for the analysis in the hypothesis 1 comparing among

the blindness and the sighted participants. Although

both participants have the potential delay against the

sound pan change types with Sigmoid function due

to the hypothesis 1, the rates of the sighted partici-

pants of Sigmoid function become larger than the one

of blindness participants in both cases of the types 1)

and 3) or 2) and 4). This conﬁrms the hypothesis 4.

Thus, we have conﬁrmed that the blindness people are

able to perceive the beginning of smooth pan change

keener than the sighted ones.

4.3 Observation regarding Rotation

Amount

We have found that the rotation amount of body bal-

ance control induced by the sound pan change with

the Heviside function is larger than the one with the

Sigmoid, which conﬁrms the hypothesis 6.

According to the condition 1 and 2 in section 4.1,

we calculated the means of the rotation amounts as

explained in section 3.1 and analyzed the rates of

the Sigmoid cases normalized by the Heviside one as

shown in Figure 7. The values in the columns of the

Sigmoid cases of left-to-right/right-to-left sound pan

changes are normalized by the ones of the same sound

pan change with Heviside function. The almost all

mean values in the graph are less than 1.0. Although

the right-to-left case of the participant E and the left-

to-right case of the participant I show larger values

than 1.000, it is still less than 1.1, which is less than

10% difference. This means that the rotation amounts

of the Heviside cases become larger than the ones of

the Sigmoid. Thus, the sudden sound pan change in-

duces larger rotation at the body balance control than

the smooth sound pan change.

4.4 Observation regarding

Auditory-space Perception

We have a question that if the change of sound pans

can guide the precise direction of the participant un-

der a dynamic body balance control such as walking.

Now, regarding the hypothesis 7, we have found that

the blindness people, especially congenital blindness,

are able to have more precise space perception to de-

tect the sound direction by the acoustic cues than the

acquired blindness and the sighted people without vi-

sion. Moreover, the congenital blindness people are

able to induce the body balance control to the correct

direction.

We focused on the sound pan change sequences

in the order of left-right-center or right-left-center,

and counted the number of times when a) partici-

pant walked straightly followingthe center sound pan,

b) participant turned to wrong direction, and c) partic-

ipant ignored the center sound pan (i.e. he/she turns

to the previous sound pan direction before the cen-

ter pan) as shown in Table 2. The table shows the

percentages of the types a), b) and c) against the to-

tal number of times of the focused sound change se-

quences appeared in the ﬁve trials of the experiment.

In the sound pan change sequences include 4 times

per whole experiment trial. A participant should re-

ceive 20 times of the sound pan change sequence in-

cluding the center sound pan in an experiment trial.

According to the table, we conﬁrm that the congen-

ital blindness has 95% of the correct recognition to

the center sound pan direction and controls their body

balance in front. This shows that they have a very

keen space perception when they construct their space

imagination using the acoustic cues.

On the other hand, the acquired blindness peo-

ple have recognized the center pan correctly only for

20%. The sighted participants without vision have

behaved in just 2% for the correct recognition of the

center sound pan, and also misunderstood the center

sound pan returned from a side as it is changing to the

opposite sound pan direction from the previous sound

icSPORTS2013-InternationalCongressonSportsScienceResearchandTechnologySupport

(a) Congenital blindness

(b) Acquired blindness

False

Correct

Correct Correct

Correct

False False False False

Starting Signals

(Center Pan)

Finalizing Signals

(Center Pan)

Left Sound

(Changing with

Sigmoid Function)

Center Sound

Right Sound

(Changing with

Sigmoid Function)

Right Sound

(Changing with

Heaviside Function)

Left Sound

(Changing with

Heaviside Function)

(a-1)

(a-2) (a-3)

(a-4)

(b-1)

(b-2) (b-3)

(b-4)

(c-1)

(c-2)

(c-3)

(c-4)

Magnetics

Sensor

(X axis)

Magnetics

Sensor

(Y axis)

Magnetics Sensor (Z axis)

Figure 6: Comparison of the three-axis magnetic sensor data synchronized with the sound pan change timings among (a)

congenital, (b) acquired blindness, and (c) the sighted participants to investigate the hypothesis 7. This comparison shows an

experiment trial. The congenital blindness participant perfectly perceives the center pan after the left-to-right acoustic cues.

However, the acquired blindness and sighted participants fails to perceive the center pan.

pan. This means that the sight is somehow involved to

recognize the space perception created by the acous-

tic cues even if the sight has been acquired for a long

time after we have experienced visible.

Figure 6 shows one of the experiment sequence

tried by the participant A (congenital blindness), the

participant B (acquired blindness) and the participant

D (sighted without vision). Lines in the graphs of the

ﬁgure show three axes of the magnetic sensor. While

the participants continue to turn their body direction

during the sound pan change, the sensor data of the

axes are changing. Regarding the sound data in the

graph, there are three starting cues in the center pan at

the beginning, the left sound pan of red part and the

right one of blue part. The purple part in the graph is

the center pan. The triangle shaped purple part in the

graph is the mixing timing of the Sigmoid function.

Let us explain an signiﬁcant case when the congeni-

tal blind participant perceives the center pan precisely

after right-to-left sound pan change shown in Fig-

ure 6(a-2). The corresponding timings of the acquired

blindness and the sighted as shown in Figure 6(b-2)

and (c-2) respectively show that both participants mis-

understood the sound pan direction to their right side.

Thus, we have conﬁrmed that the space perception

by acoustic cues of the participants who have expe-

rienced the sight in their life was weakened due to

somehow inﬂuences of their memory of vision.

4.5 Interference of Vision

Finally, we have found that the sighted people re-

sponses faster against the sound pan change with Sig-

moid function when their vision is available.

After the ﬁve time experiment trials of the sighted

participants without vision, we performed an addi-

tional experiment without eye mask. In the experi-

ment, the participant can use their vision and follows

the same instructions for the experiment as the one

without vision. We measured the response times and

calculated the rates as considered as the evidences of

the hypothesis 1. When we calculated the subtrac-

tions of the rates in the case with vision from the one

without vision, all the subtractions become positive

SpacePerceptionbyAcousticCuesInfluencesAuditory-inducedBodyBalanceControl

values (i.e. rate without vision− rate with vision > 0)

as shown in Figure 8. Thus, the response time rates

normalized by the Heviside case without vision al-

ways become larger than the ones with vision. This

means that the perception of the sound pan direction

change with vision in the case of Sigmoid function

induces body balance control in a shorter time than

the perception without vision. As well as the hypoth-

esis 7, the vision involves somehow the space percep-

tion by the acoustic cues and also inﬂuences auditory-

induced body balance control.

5 DISCUSSION

Let us summarize the results and discuss the ﬁndings

explored during the previous sections. We summarize

the ﬁndings again regarding the hypotheses:

1. A sudden sound pan change induces fast response

to both blindness and sighted people according to

the hypothesis 1.

2. A sudden sound pan change induces large rota-

tion of body balance control to both blindness and

sighted people according to the hypothesis 6.

3. Blindness people perceive smooth sound pan

change keener than the sighted people according

to the hypothesis 4.

4. Congenital blindness people perceive sound pan

direction correctly according to the hypothesis 7.

5. The vision promotes quick response for auditory-

induced body balance control due to the hypothe-

sis 8.

These ﬁndings are considerable to be applied to

engineering application and society design. Due to

the ﬁndings 1 and 2 above, we can use it for engi-

neering application such as the biofeedback system

based on sound guide. The different types of the au-

ditory cue change such as Hevisideand Sigmoid made

by the feedback system will induce different effect to

body balance control. When a fast body balance con-

trol and a large body rotation are required for correct-

ing the balance control by a feedback system, we can

apply the ﬁndings 1 and 2 to induce those potential

body balance controls with different speed. For exam-

ple, during the training of skiing to perform a beau-

tiful Wedeln with the correct timing of the parallel

turns, this feedback system using the ﬁndings 1 and

2, which feedbacks the rhythm of the turns sensed by

accelerometer, will become valid for mastering high

level skill. Thus the outcomes from the experiment in

this paper are clearly applicable and important in real-

istic engineering ﬁeld and the sports sciences applica-

0.2

0.4

0.6

0.8

A B C D E F G H I J

L to R (Sigmoid)

R to L (Sigmoid)

Part. L to R (Sig.) R to L (Sig.)

A 0.832 0.629

B 0.862 0.593

C 0.654 0.505

D 0.793 0.709

E 0.914 1.050

F 0.829 0.860

G 0.878 0.282

H 0.932 0.962

I 1.022 0.904

J 0.791 0.800

Figure 7: Comparison among rates of rotation amounts of

all participants divided by the ones with Heviside functions

of the left-to-right sound pan change to investigate the hy-

pothesis 6. The rate values in the graph show that the Sig-

moid case induces less body rotation than the Heviside one.

tions that can modify the future body balance control

correctly by acoustic cues.

On the other hand, we have found that blindness

people have superior ability to perceive the space than

the sighted people due to the ﬁndings 3 and 4. The

blindness people seem to have their stable origin of

their axes deﬁned at their center of body. Moreover

the origin of the axes is stably ﬁxed to the deﬁned

place even if the body balance control is inﬂuenced by

the acoustic cues. The ﬁnding 4 shows the evidence

of the space perception ability of blindness people,

which can be said as an affordance of blindness. Sur-

prisingly although congenital blindness people have

clearly the ability, the acquired ones do not. We can

draw an inference that if any people have experienced

vision in their life, memory of vision inﬂuences some-

how the origin of the axes placed in the center of body

by the acoustic cues. The space perception of the ac-

quired and the sighted people is dulled by their sub-

liminal vision image, and thus the matching between

their space perceived by the acoustic cues and the per-

ception of center of their body is distorted.

Finally the ﬁnding 5 means that the vision pro-

motes the correct space perception and quick body

balance control induced by the acoustic cues. It has

not been explored the reason why the vision induces

the quick response. However the experimental result

implies that the vision accelerates somehow the quick

response against the auditory cues. The sighted peo-

icSPORTS2013-InternationalCongressonSportsScienceResearchandTechnologySupport

Table 2: To investigate the hypothesis 7, comparison of re-

sponse types and the number of times in each type when

the center pan appears in the experiment trials among con-

genital and acquired blindness, and the sighted participants

without vision. The types are; a) correctly perceiving the

center sound pan and walking without rotation (straightly),

b) incorrectly perceiving the center sound pan and rotating

the body to the opposite direction of the previous sound pan,

and c) continuing the previous rotating movement without

perceiving the center sound pan.

Part. Type a Type b Type c Total times

Congenital 95% 5% 0% 20

Acquired 20% 80% 0% 40

Sighted 2% 94% 4% 140

Part. L to R (Sig.) R to L (Sig.)

D 0.431 0.182

E 1.573 1.060

F 0.879 0.697

G 1.756 1.305

H 5.395 3.027

I 0.530 0.301

J 0.739 0.851

Figure 8: Regarding the sound pan change patterns with

Sigmoid function, comparison among sighted participants

with subtracting the calculated response time rates with vi-

sion (without eye mask) from the one without vision (with

eye mask) to investigate the hypothesis 8. The subtraction

is performed between the Heviside and the Sigmoid cases

corresponding to the same sound pan change patterns that

are the left-to-right and the right-to-left ones respectively.

All data in this graph show positive values. This means that

the participants with vision perceive the sound pan change

with Sigmoid function in shorter time than the one without

vision.

ple would control their body balance with their eye-

ball movements simultaneously to exploit their quick

responses against auditory cues.

Considering the ﬁnding 3,4 and 5, let us discuss

some meanings of the ﬁndings in the society. In the

conventional training on gymnastics, it is not good

method to exploit the gymnast’s potential ability by

disabling his/her vision with eye mask if he/she is not

a congenital blindness because the space perception

by the auditory stimuli around him/her becomes un-

certain. We can say that much better method to ex-

ploit the sighted gymnast’s potential ability should be

the training technique validating their vision with be-

ing aware of sounds in the environment. This method

will induce quickly to master correct body balance

control. Another example is to go to the tunnel with

driving a car. We need to have much attention to de-

tect the driving direction because our eyes are ﬁxed to

the front of the car as the same situation of the exper-

iment mentioned in this paper with eye mask. How-

ever, the sounds are coming all around the car due to

the reﬂections of the sounds from the wall of the tun-

nel. The driver would lose his/her space perception

from the acoustic stimuli surrounding the car without

dynamic vision.

In the society design, we warn troublesome caused

by the situations where the acoustic cues inﬂuence

the body balance control related to the ﬁndings 3, 4

and 5. For example, we can meet the situation where

an old person or a child stands on Median of a large

road. The body balance control of the person is inﬂu-

enced to the opposite side against the vision because

the acoustic cues are moving on the opposite side of

the cars running in front of the person. Thus it is very

important to eliminate this kind of dangerousenviron-

ments where the changing direction of acoustic cue

is opposite to the vision stimuli. Moreover, recently

we can easily see the people who walk with operat-

ing a smart phone in a street. They are trying to per-

ceive the environment surrounding him/her from the

acoustic stimuli. However, as we conﬁrmed by the

experiments in this paper, this activity is very danger-

ous because their vision is disabled by focusing on

the screen on the phone. This accelerates unstable

space perception due to dulling their origins of axes

recognized by the acoustic stimuli. Additionally the

condition to percept the space has become worse in

these days due to wearing headphone to listen to mu-

sic from the phone. While he/she with a headphone

walks in a street with making an email sentences on

a phone, if a car is passing in front of him/her, he/she

should delay to respond to the sound of the car and

will have an accident. Thus, not only the vision, it

should be very important to design the society or the

products considering the auditory effects considering

both the vision and the auditory.

6 CONCLUSIONS

This paper investigates the inﬂuence to the space per-

ception when acoustic cues are inputted to blindness

and sighted people during walking. We performed ex-

periments to investigate eight hypothesis and found

SpacePerceptionbyAcousticCuesInfluencesAuditory-inducedBodyBalanceControl

that the ﬁve hypotheses are conﬁrmed: 1) The speed

for body balance control is affected by the sound pan

change type. 2) The rotation amount for body bal-

ance control is affected by the sound pan change type.

3) The blindness people has keener perception for

the auditory cues than the sighted. 4) The congeni-

tal blindness people has a very stable axes that is not

moved by the acoustic cues. And ﬁnally 5) the sighted

people responses quicker against a smooth sound cue

with vision than without vision.

These ﬁndings outcome the relations among the

acoustic stimuli and the auditory-induced body bal-

ance control during a dynamic movement because the

vection was only the cognitive and psychologicalmat-

ter which is detected internally in the participants on

a stable posture. Our ﬁndings can be applied to en-

gineering applications and the social design to warn

the dangerousness when the sighted people lose the

vision. Recently we meet such situations in many

places. We warn to improve our life style to use vi-

sion with acoustic as much as possible. However, it

is not sure that the absolute reasons such as the re-

lated brain parts and the psychological affects to the

response quickness or the misperception of the center

sound pan. Therefore, expecting an electroencephalo-

graph or a small fMRI used in a walking situation,

we will continuously explore the main reason why the

ﬁndings in this paper occur so far.

ACKNOWLEDGEMENTS

We thank to teachers of blind school in Kochi prefec-

ture, especially to Prof. Iida who accepted our offer

for the experiments. This research is partially sup-

ported by the KAKENHI No. 23650409 of Grant-in-

Aid for Challenging Exploratory Research, and No.

20240060 and 24240085 of Grant-in-Aid for Scien-

tiﬁc Research (A) funded by Japan Society for the

Promotion of Science (JSPS) and also is partially sup-

ported by the JST PRESTO program and Tateisi Sci-

ence and Technology Foundation.

REFERENCES

Aoki, M., Thilo, K., Burchill, P., Golding, J., and Gresty,

M. (2000). Autonomic response to real versus illu-

sory motion (vection). Clinical Autonomic Research,

10(1):23–28.

Blake, R. and Sekuler, R. (2005). Perception. McGraw-Hill

Humanities/Social Sciences/Languages.

Brunelli, D., Farella, E., Rocchi, L., Dozza, M., Chiari, L.,

and Benini, L. (2006). Bio-feedback system for reha-

bilitation based on a wireless body area network. In

Fourth Annual IEEE International Conference on Per-

vasive Computing and Communications Workshops,

2006.

Dozza, M., Chiari, L., Hlavacka, F., Cappello, A., and Ho-

rak, F. (2006). Effects of linear versus sigmoid coding

of visual or audio biofeedback for the control of up-

right stance. Neural Systems and Rehabilitation Engi-

neering, IEEE Transactions on, 14(4):505–512.

Giansanti, D., Dozza, M., Chiari, L., Maccioni, G., and

Cappello, A. (2009). Energetic assessment of trunk

postural modiﬁcations induced by a wearable audio-

biofeedback system. Medical Engineering & Physics,

31(1):48–54.

Hove, M. J., Suzuki, K., Uchitomi, H., Orimo, S., and

Miyake, Y. (2012). Interactive rhythmic auditory stim-

ulation reinstates natural 1/f timing in gait of parkin-

son’s patients. PLoS ONE, 7(3):e32600.

Lackner, R. J. (1977). Induction of illusory self-rotation and

nystagmus by a rotating sound-ﬁeld. Aviation, Space,

and Environmental Medicine, 48(2):129 – 131.

Lackner, R. J. (1978). Handbook of sensory physiology.

Some Mechanisms Underlying Sensory and Postural

Stability in Man, 8:805 – 845.

Lackner, R. J. (1983). Handbook of sensory physiology.

Inﬂuence of Posture on the Spatial Localization of

Sound, 31(9):650 – 661.

Miyake, Y. (2009). Interpersonal synchronization of body

motion and the walk-mate walking support robot.

Robotics, IEEE Transactions on, 25(3):638 –644.

Valjamae, A. (2009). Auditorily-induced illusory self-

motion: A review. Brain Research Reviews, 61(2):240

– 255.

icSPORTS2013-InternationalCongressonSportsScienceResearchandTechnologySupport