Mental Health Self-check System using “Lyspect”
Mayumi Oyama-Higa
1
, Tiejun Miao
2
, Shigeo Kaizu
2
and Junji Kojima
3
1
Department of Systems Innovation, Osaka University, Toyonaka, Osaka 560-8531, Japan
2
Chaos Technology Research Laboratory, Ōtsu, Shiga 520-2134, Japan
3
Rakuwakai Otowa Hospital, Yamashina-ku, Kyoto 607-8062, Japan
Keywords: Plethysmograms, Largest Lyapunov Exponent, Nonlinear Analysis, HFLF Balance, Blood-vessel Balance.
Abstract: The largest Lyapunov exponent (LLE) obtained through nonlinear analysis of plethysmograms contains
information of the cerebral central system, which has been shown by mathematical model and experiments.
Especially mental experiments have shown the significant relationship between LLE and mental status of an
individual. The wave data are obtained by measuring changes in the blood flow of the capillary vessels of a
fingertip, with an infrared sensor. Then the analogue format of the data is changed into digital format for
calculation. Using “Lyspect” software, we next compute parameters including LLE, HF/LF, autonomic-
nerve balance, and blood vessel age. These parameters, as functions of time, can be displayed in graphs.
Notably, while the pulse wave is being measured, changes of LLE and HF/LF can be displayed in real time.
Furthermore, the setup of measuring time and various parameters for calculation is available. The state at
the time of the measurement can be studied by visualization. Additionally, the unsuitable wave data arising
from the accuracy of the sensor and the external noise can be eliminated by the filter of Lyspect. Currently,
a version of Lyspect that is installable on user-friendly smartphones is being developed, which will make
even easier timely self-check of mental status.
1 INTRODUCTION
In chaotic datasets, attractor plots and “divergence”
of attractor trajectories are characterized by
Lyapunov exponents. Previously, we focused on the
Lyapunov exponents of pulse waves in studying
individuals in various situations (Imanishi and
Oyama-Higa, 2006; Miao, Shimoyama and Oyama-
Higa, 2006; Oyama-Higa and Miao, 2006; Oyama-
Higa, Tsujino and Tanabiki, 2006). Our results
showed that to maintain mental health, it is
important for an individual to keep the harmony in
the functioning of the sympathetic nervous system,
which is associated the individual’s qualities such as
flexibility, spontaneity, cooperativeness, and the
ability to interact with the external environment and
society.
We also learned that the values representative of
such harmony are associated with the largest
Lyapunov exponent (LLE) obtained from nonlinear
analysis (Oyama-Higa and Miao, 2005 and 2006).
Essentially, in this research, LLE, which signifies
temporal fluctuations in the attractor trajectory, is
defined as “divergence”. When this value is
continuously low, namely, when there is no
divergence for a long period, adaptability to external
factors in daily life decreases, and thus mental health
cannot be maintained. Conversely, a value that is
continuously high for a long period represents a
continued state of extreme anxiety or stress, and also
mental health cannot be maintained in this case.
For humans, a healthy state can be defined as one
in which high and low divergences constantly
alternate. Normal human life includes a wide range
of emotions, which is probably the cause of such
changes in divergence. Using nonlinear analysis of
pulse waves, an individual’s mental health can be
measured in approximately 1 minute, with the help
of a low-cost pulse wave sensor. This offers the
potential for easy assessment of mental health that
can be performed regularly at home or workplace.
We have created a trial version of the self-check
system.
Mental health changes from day to day and hour
to hour, so it is most important to monitor these
fluctuations closely and intervene quickly when
problematic symptoms emerge. To this end, we
developed a self-check system producing a graph in
which the degree of mental health over time is
visualized as a constellation (Oyama-Higa et al.,
2007).
“Lyspect” software can perform three types of
9
Miao T., Kojima J., Oyama-Higa M. and Kaizu S. (2012).
Mental Health Self-check System using “Lyspect”.
In Proceedings of the Sixth International Symposium on e-Health Services and Technologies and the Third International Conference on Green IT
Solutions, pages 9-18
DOI: 10.5220/0004474600090018
Copyright
c
SciTePress
analysis (chaos analysis, blood-vessel balance
analysis and autonomic-nerve balance analysis)
using finger plethysmograms as input data. Besides
the version that produces a detailed result for
researchers, we also developed a simplified version
of “Lyspect”, called “Lyspecting”, for general users.
2 MAIN FEATURES OF LYSPECT
Lyspect possesses the following functions.
2.1 Analysis
2.1.1 Chaos Analysis
To calculate LLE of pulse waves. Since Lyspect
calculates the spectrum of Lyapunov exponents,
values of Lyapunov exponents of all embedded
dimensions are also found.
2.1.2 Blood-vessel Balance Analysis
To calculate blood-vessel balance equivalent to the
estimated vascular age from a second differential
wave pattern using Lyspect's own method.
2.1.3 Autonomic-nerve Balance Analysis
To calculate the LF and HF ratio from spectral
analysis of variability in the HR cycle.
2.1.4 Others
It can also calculate the balance of sympathetic
nerves and parasympathetic nerves using Lyspect's
own method for estimating HF from heart rate
transitions.
The analysis range and slide width for all
analyses can be specified and adjusted. When the
analysis range is shorter than the wave data length
and the slide width is not 0, multiple results are
generated by sliding along the analysis range for the
amount of the slide width.
2.1.5 Pulse Wave Filtering (Optional)
To analyze pulse waves with a band-pass filter that
removes bass-line transitions of wave data. It is
available only when the filter option is installed.
2.2 Wave Input
Other than normal text files, Lyspect reads CCI's
BACS-Advance
binary wave data files as input
wave data for analysis.
Another way is to dynamically read wave data
from a connected cuff online.
Besides, it can also import wave data records
from the DB of MSAS software, a self-check system
using the database of the pulse waves which we have
developed. In the case the DB contains records of
analysis results, they are also imported and recorded
as clips.
2.3 Analysis Result
There are two types of data to be saved.
1. Save analysis conditions and results to a file in
CSV format, with each analysis on a single row.
2. Save sequences of analysis results in multiple
rows to a file in CSV format when the slide width of
the analysis is not 0 and multiple analysis results are
generated.
2.4 Display and Save Pulse Waves
We introduce the following operations in Lyspect.
1. Save source data of displayed pulse waves
to an external file in text format. This feature can be
used, for example, to load and save BACS-Advance
binary wave data.
2. Display a 3D Takens plot in chaos analysis.
Assess the graph using zoom/pan.
Display Takens plot of wave data applied with a
band-pass filter as source data. (Optional. Available
only when the filter option is installed.)
3. Display second differential wave form of
pulse waves (and markers indicating detected “a”,
“b” and “c” peaks). This wave data can be saved to
an external file in text format.
4. Display a graph of heart-rate transition.
This graph data can be saved to an external file in
text format.
5. Display pulse waveform applied with a
band-pass filter that removes transitions of original
pulse wave data. (Optional. Available only when the
filter option is installed.) This waveform data can be
saved to an external file in text format.
6. Display waveform showing the bass-line
transition. (Available only when the filter option is
installed.)
This waveform is created simply by subtracting
the pulse waveform that has been applied with a
band-pass filter from the source pulse waveform.
The waveform data can be saved to an external
file in text format.
EHST/ICGREEN 2012
10
3 HOW TO USE LYSPECT
Details on the usage of Lyspect are explained here in
several steps.
3.1 Layout of the Operation Windows
The user can freely switch the layout of operation
windows. Listed below are the main operation
windows of Lyspect.
3.1.1 “Parameters”
In this window, analysis parameters are set.
Figure 1: Parameters.
3.1.2 “Wave Form”
In this window, the waveform of pulse waves is
displayed.
Figure 2: Wave form.
3.1.3 “Resultgraph”
In this window, the analysis results are displayed
using graphs.
Figure 3: ResultGraph.
3.1.4 “ClipListTable”
In this window, a log of the analysis results is
displayed in a table format.
Figure 4: ClipListTable.
3.1.5 “LogArea”
This is a message display area for text information.
Figure 5: LogArea.
The above windows appear as sub-windows
within the Lyspect main-window.
The layout is not fixed and can be freely moved
by the user.
3.2 Loading Wave Data
There are mainly two ways to load wave data for
analysis:
1. Text files from a wave data file and
BACS-Advance pulse wave binary files.
2. Loading wave data from a connected cuff.
This operation loads data by sampling the wave data
from a cuff connected to a USB port.
3.3 Setting Analysis Conditions and
Analyzing
Analysis conditions are set in the “Parameters”
window. See Figure 1.
Data length shows the number of sampling
points of the loaded wave data.
In “Sampling sec.”, the sampling rate is set. For
example, we fill in 0.005 for 200Hz and 0.001 for
1,000Hz sampling rate.
There are three tabs that correspond to each type
of analysis.
Figure 6: Three types of analysis.
Mental Health Self-check System using “Lyspect”
11
“Apply filter” is effective only when the band-
pass filter is enabled. The feature is used to perform
analysis of filtered data.
3.4 Graphical Display of Analysis
Result
Analysis conditions are set in the “Parameters”
window.
Figure 7: Summary.
There are a number of tabs on the top of the
“ResultGraph” window.
3.4.1 “Summary”
This tab shows the main analysis result values of the
three types of analyses in constellation graphs.
1. Chaos:
The average value of LLE in sliding analyses is
shown as the angle of the meter and the standard
deviation is shown as a circle where the radius is the
standard deviation value.
2. Estimated vascular age:
The meter is in a range of 0 to 10 from which
one obtain the estimated vascular age by multiplying
the value with 10. For example, the value 3.2 means
the age is 32.
3. Autonomic-nerve balance:
Autonomic-nerve balance is defined as the value
of LF / ( LF + HF ). A value larger than 5 indicates
the sympathetic nerve dominance, namely, a state of
stress, while a value smaller than 5 shows the
parasympathetic nerve dominance, that is, a state of
relaxation.
3.4.2 “Chaos”
This is a graph showing the changes of LLE
resulting from sliding analyses where the number of
slides is in the horizontal axis.
Figure 8: Chaos.
In Lyapunov analysis, the spectrum is found for
the number of Lyapunov dimensions. But in the
initial display, only the first Lyapunov exponent,
namely the largest Lyapunov exponent (LLE), of the
spectrum is displayed on the graph. Click the
checkboxes of the legend on the right of the graph,
and the second, third and fourth Lyapunov
exponents will be displayed. The vertical axis of the
graph is scaled automatically.
One is able to right-click the graph to save the
graph data to an external file in CSV format.
If the chaos analysis is performed with Full range
analysis or Slide pitch is 0, only one set of the
Lyapunov spectrum will be calculated. In this case,
the graph displayed in the “Chaos” tab will be a
graph of the convergent calculation of the Lyapunov
exponents as shown below.
Figure 9: Display of four Lyapunov exponents.
3.4.3 “HFLF Balance”
A graph is displayed when one performs slide
analysis by specifying the slide pitch rather than
full-range analysis and setting the calculation
method to HF-LF method in Autonomic-nerve
balance analysis.
Figure 10: HFLF balance.
The graph shows changes of LF and HF values
for the entire analysis area with the number of slides
EHST/ICGREEN 2012
12
in the horizontal axis.
In the case of slide analysis, the value of
“b10HR” is calculated for the last analysis range
(near the end of data).
Click the checkboxes of the legend on the right
of the graph to enable/disable the display of HF, LF
and “b10HR”.
Right-click the graph to save the graph data to an
external file in CSV format.
3.4.4 “HFLF Power Spectrum”
This spectrum distribution graph is displayed when
the calculation method is set to HF-LF method in
Autonomic-nerve balance analysis.
Figure 11: HFLF power spectrum.
The graph shows changes of LF and HF values
for the entire analysis area with the number of slides
in the horizontal axis.
In the case of slide analysis, the value of
“b10HR” is calculated for the last analysis range
(near the end of data).
Click the checkboxes of the legend on the right
of the graph to enable/disable the display of HF, LF
and “b10HR”.
Right-click the graph to save the graph data to an
external file in CSV format.
3.4.5 “Blood-vessel Balance”
This analysis evaluates suppleness of blood-vessel.
One index value of suppleness is calculated from
every time of pulsation, and shown in the graph.
Figure 12: Blood vessel balance.
3.5 Numerical Display of Analysis
Result
3.5.1 Clip and Clip List
The following is the “ClipListTable” window.
Figure 13: ClipListTable for analysis result.
A clip is a record of a single analysis:
Clip: = analysis data + analysis condition
+ analysis result (+ comment)
(1)
A clip list is a set of clips. Normally, the list contains
multiple clips.
Clip List: = clip + clip + ... + clip (2)
The clip list is displayed in the “ClipListTable”
window in table format. In detail, a clip contains:
Wave data: = wave filename (3)
Analysis conditions: = parameters set in the
Parameters window (4)
Analysis result: = analysis result for each of the
three types of analyses including series of values
such as chaos spectrum, HFLF power spectrum, etc.
(5)
Comment: = any character string user entered.
This item is completely optional, and used to enter
information such as data profile or others. (6)
3.5.2 Setting the Items to Display in the
“ClipListTable” Window
Since a clip contains enormous amount of numerical
data, it is not advisable to display all numerical data
in the “ClipListTable” screen.
The user may select the items to display in the
window. To do so, select menu “Clip” and then
“Clip table display columns” for details about the
item names of analysis results.
3.5.3 Entering Comment
To enter comment, first select “Comments” as an
item to display in the “ClipListTable” window. Next,
click the “Comments” column of the row to enter
Mental Health Self-check System using “Lyspect”
13
comment in the “ClipListTable” window. When the
“EditText” window appears, enter or edit comment.
3.5.4 Displaying Graphics of a Clip
Clicking a clip in the “ClipListTable” window will
display the following:
1. The wave data of the clip in the “Waveform”
window,
2. The analysis conditions of the clip in the
“Parameters” window and
3. The graph of the clip's analysis result in the
“ResultGraph” window.
3.5.5 Saving Clip List to a File
In Lyspect, saving analysis results is performed by
saving Clip list to a file.
Clip list may be saved by selecting “Clip” and
then “Save file” in the menu. In order to prevent
users from accidentally forgetting to save to a file,
however, Lyspect will display a message asking
“The clips have not been saved. Save?” when the
user terminates it without saving Clip list.
3.5.6 Loading a Previous Analysis Result
(Loading a Clip File)
A clip file can be loaded by selecting “Clip” and
then “Read file” in the menu. The number of loaded
clips will be displayed and Clip list will appear in
the “ClipListTable” window.
The last saved clip file can be loaded by
selecting menu “Clip” and then “Read the last file”.
Lyspect saves the names of the last saved clip files
and thus can load them quickly without finding their
paths again.
3.5.7 Exporting Analysis Results in CSV
Format
Although clip files are text files, the format is unique
to Lyspect.
In order to use the analysis results in other
assessment applications such as MS EXCEL, clip
information can be converted and saved in CSV
format with commas separations.
Each clip will be output in a single row.
Since a clip contains enormous amount of data,
information (columns) to output to a CSV file can be
specified in the same way as specifying columns to
display in the “ClipListTable” window.
To save clip files in CSV format, select “Clip”
and then “Export in CSV format” in the menu.
3.5.8 Deleting Clips from Clip List
To delete a clip from Clip list, click the clip to delete
in the “ClipListTable” window, right-click to display
the pop-up menu and then select “Delete”. To delete
all clips in Clip list, select “Clip” and then “Clear
table” in the menu.
3.6 Displaying the Pulse Waveform
The “Wave form” window displays the wave data
loaded from a file or a cuff. See Figure 2.
The wave data displayed in the “Wave form
window is the pulse wave subject to analysis.
The vertical axis of the wave graph is scaled so
that the maximum value and minimum value of the
entire wave data extend to the upper and lower edges
of the graph.
3.6.1 Zooming the Waveform
Drag the mouse in upward direction on the graph to
zoom in the horizontal axis of the waveform. In the
same way, drag the mouse in downward direction to
zoom out.
To reset the scale, right-click on the mouse and
then select the “Reset zoom” in the pop-up menu.
Figure 14: Zooming.
3.6.2 Displaying Secondary Pulse Waves
Waveforms generated by processing the source
pulse waveform in various ways are called
secondary pulse waves. These waves can be
displayed by selecting menu “Display” and
“Secondary waveform”.
To save secondary pulse wave data to an
external file in text format, right-click on each graph
and then select “Save wave file” in the pop-up menu.
3.6.3 Second Differential Waves
Second differential waves are waveforms that
differentiate the source pulse waves twice and are
used in analyses of both blood-vessel balance and
autonomic-nerve balance since they effectively
express the characteristics of pulse wave shapes.
EHST/ICGREEN 2012
14
Figure 15: Second differential waves.
“a”, “b” and “c” peaks used for assessing the
blood-vessel balance are drawn in white, red and
blue lines respectively for each pulse wave.
3.6.4 Heart Rate Transitions
The “a” peak interval of the second differential
waves refers to the pulse-cycle period and the heart-
rate transitions graph shows the momentary heart
rate found by inverse calculation of the pulse-cycle
period.
Figure 16: Heart rate transitions.
3.6.5 Band-pass Filter
This is available only when the filter option is
installed.
3.6.6 Base Line Transitions
This is available only when the filter option is
installed. The graph shows a waveform created by
performing a simple subtraction of original source
waveform - filtered waveform.
It is useful to see the trend of bass-line transitions.
Figure 17: Base line transitions.
3.7 Chaos Analysis Takens Plot
The Takens plot of pulse waves can be displayed by
selecting “Display” and then “Takens display” in the
menu.
Figure 18: The Takens plot.
Although chaos analysis of pulse waves is
performed in the embedded dimensions, the plot
displayed here is a three dimensional plot.
The axis in depth direction is displayed using
line colours. The red lines indicate at the front and
blue lines indicate to the back. (All lines are
displayed in green during in the initial state.)
Besides the source waveform, the filtered
waveform can also be displayed when the filter
option is installed.
As to the attractor displayed by the method of
Takens, the following operations are possible:
1. Rotation of coordinates can be performed.
2. Zoom of coordinates is available.
3. Whether to display the sample point or not is
optional.
4. The starting point of drawing and the number
of the points of drawing can be set up.
5. Delay time and evolution time can be set up.
3.8 Names and Contents of Analysis
Result Items
3.8.1 Items Names and Contents of
Blood-vessel Balance Analysis Result
Figure 19: Output parameters of blood-vessel balance
analysis.
1. “b/a”, “c/a”, “d/a”, and “e/a”
Mental Health Self-check System using “Lyspect”
15
Average values for peak “a”, “b”, “c”, “d” and
“e” ratios for the entire analysis range. These are the
bases for calculating the blood-vessel balance.
2. “Pb”
Estimation of the vascular age (method 1).
3. “age”
Estimation of the vascular age (method 2).
4. “sigma”
It denotes the vascular aging index σ.
5. “Age25%”, “ Age50% and “Age75%”
In contrast with "age" value, AgeXX% values
are based on the distribultion graph of "age" values
calculated from b/a, c/a, d/a and e/a of each pulse
through the analysis area of pulse wave. Age25% is
the value of "age" of one forth counting from
smallest "age" value in the distribution graph,
Age50% is the "age" value of the half counting in
the graph, and Age75% is the "age" value of the
three forth counting in the graph.
3.8.2 Item Names and Contents of
Autonomic-nerve Balance Analysis
Result
Figure 20: Output parameters of autonomic-nerve balance
analysis.
1. “hf” and “lf”
They express each spectrum cumulative value, as
the base of LF/HF spectral analysis. The balance of
autonomic nerves is assessed by the ratio of HF and
LF. LF is in the range of 0.04Hz – 0.15Hz, while HF
is in the range of 0.15Hz – 0.40Hz.
2. “HFa”
It denotes the estimated HF found from the
average heart rate. The formula is
HFa = K1×HR×3 + K2×HR×2 + K3×HR + K4,
(7)
0
HFa
100, (8)
where K1 to K4 are coefficients for approximation
and HR is the average heart rate. These coefficients
can be changed by selecting “System” and then
“Configuration” in the menu.
A value taken between 40 and 60 indicates a
normal balance. HFa
40 and 60
HFa indicate
sympathetic nerve dominance and parasympathetic
nerve dominance, respectively.
3. “HR Avr”
It means the average heart rate within the
analysis range
4. “Ln(LF)/Ln(HF)”
This parameter signifies the autonomic-nerve
balance assessment value found from the log ratios
of LF and HF. The formulas are:
B = Ln(LF) / Ln(HF), (9)
B10 = B × 10 / 3.5, (10)
0
B10
10. (11)
B10 < 5 indicates parasympathetic nerve dominance,
while B10>5 indicates sympathetic nerve dominance.
5. “b10”
b10 shows the autonomic-nerve balance
assessment value found from the ratios of LF and
HF. The formulas are:
hflf = HF / ( HF + LF ), (12)
b10 = ( 1 – hflf ) × 10 = LF / ( HF + LF ) × 10,
(13)
0
b10
10. (14)
b10 < 5 indicates parasympathetic nerve dominance,
while b10>5 indicates sympathetic nerve dominance.
6. “b10HR”
It means the autonomic-nerve balance
assessment value found from the HR value. The
formulas are:
HFa = the HF value ( ×100 ) estimated from HR,
(15)
b10HR = ( 1 – Hfa / 100 ) × 10, (16)
0
b10HR
10. (17)
b10HR < 5 indicates parasympathetic nerve
dominance, while b10HR > 5 indicates sympathetic
nerve dominance.
4 EXAMPLES OF APPLICATION
We give three examples involving the calculation
using Lyspect.
4.1 Example of a Healthy Subject
Below is a mentally healthy person's data. The
subject is measured for three times, and for each
time the duration of measurement is three minutes.
EHST/ICGREEN 2012
16
The blood vessel age of this subject is about 40
years.
LLE and autonomic-nerve balance are also in
good condition.
Figure 21: ResultGraph of a middle-age healthy subject.
4.2 Example of an Elderly Subject
The subject is around 60 years old. From the three
times of measurement, we found the LLE is low, but
the autonomic-nerve balance is normal.
Figure 22: ResultGraph of an elderly healthy subject.
4.3 Example of a Depression Sufferer
In the following we present the result of a 49-year-
old dysthymia-related obstacle (depression) sufferer.
To increase the accuracy, we performed the
measurement for 18 minutes in total. The result
shows that LLE is low and the value of autonomic-
nerve balance is not less than 5 for each time. Thus
the sympathetic nerve is higher in than
parasympathetic nerve. This result agrees with the
general tendency that the LLE of a mental disease
sufferer is low and his or her autonomic-nerve
balance is higher in than 5.
Figure 23: ResultGraph of a mentally ill subject.
5 DISCUSSION
With the help of Lyspect, individuals can perform
self-checks and self-management for mental health.
Mental health can be possibly maintained, if it is
possible to ascertain exactly what kind of day-to-day
conditions bring about high and low divergence. We
are confident that even for those potential mental
disease sufferers who regularly go to consult a
counsellor or a psychiatrist, this system can aid in
the early detection of depression or dementia, and
thus prevent further deterioration of mental health.
Furthermore, we think it highly original to send
and receive data related to mental health indicators
across networks. However, it is essential to take
great care in data management in light of
confidentiality, which has become more important in
recent years.
With the purpose of enabling users to perform
the measurement at any time and anywhere as well
as to compare their past health conditions on the
Internet, we developed the software which can be
used on the smart phone of Android loading . This
software can also display the state of a pulse wave,
LLE, autonomic-nerve balance, and blood vessel age.
A measurement result can be saved via individual
PC or the Internet with security management.
Mental Health Self-check System using “Lyspect”
17
What is more, whenever necessary, the past
measurement results can be retrieved from the
database and investigated, and it therefore becomes
possible to learn the change in one's mental state.
We expect that a small and cheap pulse wave
sensor connectable with a mobile phone or a small
PC will be developed early.
REFERENCES
Imanishi, A. and Oyama-Higa, M. (2006). The Relation
Between Observers’ Psychophysiological Conditions
and Human Errors During Monitoring Task. 2006
IEEE Conference on Systems, Man, and Cybernetics,
2035–2039.
Miao, T., Shimoyama, O. and Oyama-Higa, M. (2006).
Modelling Plethysmogram Dynamics Based on
Baroreflex under Higher Cerebral Influences. 2006
IEEE Conference on Systems, Man, and Cybernetics,
2868-2873.
Oyama-Higa, M. and Miao, T. (2005). Representation of a
Physiopsychological Index Through Constellation
Graphs. Lecture Notes in Computer Science, 3610,
811-817.
Oyama-Higa, M. and Miao, T. (2006). Discovery and
Application of New Index for Cognitive Psychology.
2006 IEEE Conference on Systems, Man, and
Cybernetics, 2040-2044.
Oyama-Higa, M., Miao, T. and Mizuno-Matsumoto, Y.
(2006). Analysis of Dementia in Aged Subjects
Through Chaos Analysis of Fingertip Pulse Waves.
2006 IEEE Conference on Systems, Man, and
Cybernetics, 2863–2867.
Oyama-Higa, M., Miao, T., Sato, K., Tanaka K. and
Cheng, H. (2007). Development of a Self-check
System for Mental Health Using a Pulse Wave Mouse.
ICSOFT 2007, The 3rd International Conference on
Software and Data Technologies.
Oyama-Higa, M., Tsujino, J. and Tanabiki, M. (2006).
Does a Mother’s Attachment to Her Child Affect
Biological Information Provided by the Child? Chaos
Analysis of Fingertip Pulse Waves of Children. 2006
IEEE Conference on Systems, Man, and Cybernetics,
2030-2034.
EHST/ICGREEN 2012
18