978-1-4244-9352-4/11/$26.00 ©2011 IEEE 805
2011 4th International Conference on Biomedical Engineering and Informatics (BMEI)
Physiological Signal Analysis for Patients with
Depression
Yen-Ting Chen
1
, I-Chung Hung
1*
, Min-Wei Huang
23
, Chun-Ju Hou
1
, and Kuo-Sheng Cheng
2
1*
Department of Electrical Engineering, Southern Taiwan University, Tainan, Taiwan
2
Institute of Biomedical Engineering, National Cheng Kung University, Tainan, Taiwan
3
Psychiatry Department, Chiayi Branch, Taichung Veterans General Hospital, Taiwan
E-mail: hminwei@gmail.com
Abstract—Depression is a common and serious mental disorder.
About 1.9 million people in Taiwan are identified as having
depression. There is a trend of increase of the prevalence of
depression, three times more depressed persons within the past
six years. A few patients with depression were in treatment.
Therefore, an available algorithm will be build to measure the
neurophysiology of depression. In this study, the physiological
signals from depressed patients will be compared with those
from normal people. This experiment use different pictures
expresses happiness, sadness fear, and disgust to cause the
emotion of subjects. The physiological signals of the patient is
measured at the same time. The preliminary results show that
the galvanic skin response, heart rate variability, and blood
volume pulse of patients with depression is lower than the
normal people does. More subjects will be evaluated in the
project for investing the clinical significance.
Keywords-Depression; Physiological Signal; Heart Rate
Variability; Galvanic Skin Response
I. INTRODUCTION
There are many kinds of pressure living in present.
Because the pressure become heavy, some people are living
with depression. There are 15% patients with depression
will suicide themselves. And there are 87% suicides
diagnosed with depression during their lifetime. Suicide is
the second cause of death between the age of 15 and 24, and
the third cause of death between the age of 25 and 44, and
the seventh cause of death between the age of 45 and 64.
Moreover, the percentage of depressed patients in treatment
were low. Even if the depressed patients were in treatment,
doctors usually relied on their medical history and clinical
observation [1]. Therefore, the purpose of this study is to
develop an available algorithm to measure the
neurophysiology of depression.
The electrocardiogram, blood pressure, respiration,
electromygram and skin temperature are as measures for
clinical application [2~4]. Depression in Heart Rate
Variability (HRV) measurement, usually shows low HRV
and high ratio of parasympathetic sympathetic LF/HF [5]. In
tension or in relaxation of a people will cause the variation of
finger temperature and the difference of electric potential in
muscle. The other physiology signal in this study for
measurement is Blood Volume Pulse (BVP). BVP is about
the flow of blood when a blood vessel is in vasoconstriction
or vasodilatation in different emotion [6]. Therefore, the
physiology parameters, like Electrocardiogram (ECG),
Electromygram (EMG), Galvanic Skin Response (GSR),
Finger Temperature (Temp), Blood Volume Pulse (BVP)
will be used in estimating depression.
This study will establish a trend analysis to investigate
the relations of physical signals and emotions of patients
with depressive disorder. According to the statistical
inference from subjects to find out the standard for each
parameter.
II. MATERIALS AND METHODS
The physiological signals of two groups of subjects, 16
normal persons and 10 patients, were measured. The
protocol for this study was approved by the Institutional
Review Board (IRB) and the informed consents were
enrolled and agreed by all subjects.
The emotional trigger films including some international
emotional pictures (happiness, sadness, fear and disgust)
from International Affective Pictures System (IAPS) were
applied while physiological signals were measured at the
same time.
The flowchart for physiological signals analysis is as Fig.
1. First, the signals will be separated in different stimulated
time segment. Second, the characteristics of every signals
were evaluated for each part.
Third, the comparisons of evaluated results in different
time segments were performed for investigating the emotion
variation. Finally, the statistical analysis to verify the
experimental results.
Figure 1. The flowchart for physiological signals analysis
Heart rate variability is analyzed from ECG signals. And
the way to calculate heart rate variability is Power Spectral
Density (PSD) algorithm [8]. The algorithm is to find out
the interval T between two adjacent R waveforms. The basic