Detection of a maximum blood flow velocity variation in the carotid artery during physical exercise using BPNN
To know a characteristic of the whole blood cardiovascular system, we need to measure the blood flow respond in relation to giving exercise stress to our body. In this paper, we have developed a noninvasive device to be able to measure blood flow velocity in the carotid during physical exercise by u...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | English |
| Published: |
Springer Berlin Heidelberg
2007
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/44126/ http://irep.iium.edu.my/44126/ http://irep.iium.edu.my/44126/ http://irep.iium.edu.my/44126/1/Detection_of_a_Maximum_Blood_Flow_Velocity_Variation_in_the_Carotid_Artery_during_Physical_Exercise_using_BPNN.pdf |
| Summary: | To know a characteristic of the whole blood cardiovascular system, we need to measure the blood flow respond in relation to giving exercise stress to our body. In this paper, we have developed a noninvasive device to be able to measure blood flow velocity in the carotid during physical exercise by using a Doppler ultrasound method. The purpose of this study is that we focus attention on measured maximum blood flow velocity (Vmax) and analyze the physiological function change of blood flow dynamic system in the carotid during physical exercise by a cicycle-ergometer. Generally biological information (eg. brain wave) is known as a nonlinear signal. Correspondingly, neural networks (NN) have recently been known as one of nonlinear signals analysis tools. For such occasions, we suggest that we identify Vmax (non-linear time series) variation using an MA (Moving Average) model associated with back-propagation (BP) algorithm. After simulating Vmax variation, we found out that the neural network (BPNN) can capture the regularity from that, and the adaptive changes of the system dynamics are evaluated in the connection-weight-space (CWS), a kind of internal representation of the trained network. As a result, we can abstract respective characteristic from that. This technique is expected to estimate a physical exercise quantitatively and rehabilitation training. |
|---|