biosignals_lecture_full_ text.pdf

1 Biological Signals Characteristics of biosignals. Biosignal acquisition. Active and passive sensors. Electrodes Noise. Noise reduction  2 Characteristics of biosignals All living things, from cells...

1 Biological Signals Characteristics of biosignals. Biosignal acquisition. Active and passive sensors. Electrodes Noise. Noise reduction  2 Characteristics of biosignals All living things, from cells to organisms, deliver signals of biological origin. Such biological signals - in short, biosignals - can be of interest for diagnosis, for patient monitoring, and biomedical research. Biosignal can be defined as a description of a physiological phenomenon. Most of the signals we will talk about can be described as some physical quantity changing over time or with position in space. Throughout their lifetimes, living organisms generate an abundant stream of signals, often hidden in a background of other signals and noise components. So it is very important to learn how to separate the biosignal of interest out from other signals and the noisy background. In other words it is necessary to reduce the redundant data stream to only a few, but relevant, parameters. Such parameters must be significant for medical decision making, for example, to solve a medical problem or to increase insight into the underlying biological process. In other words we need to characterize the signal. First we can characterize biosignals according to their origin. By origin we mean the source that generates the signal. Biosignals can be electric (e.g., the depolarization of a nerve cell or the heart muscle), mechanical (e.g., the sound generated by heart valves), or chemical (e.g., the concentration of CO2 in the blood) etc. Then signals could be classified as continuous or discrete. Continuous signals are signals that are defined at any moment in time. Discrete signals are signals that are only defined at discrete points in time. Fig. Continuous (left) and discreet (right) signals. Sampling is the reduction of a continuous signal to a discrete signal. Fig. Sampling  3 Biosignals can also be classified as deterministic or random. Deterministic signals can be described by explicit mathematical relationships, whereas random signals cannot be exactly expressed. Biological processes that show some repetitive character, such as a beating heart or respiration, generate signals that are also repetitive. Such signals often show a more or less deterministic waveform. Deterministic signals can be periodic or aperiodic (non- periodic). Nonperiodic signals can be quasiperiodic, or simply transient. In living organisms, periodic signals (such signals are essentially only mathematically defined, such as a sine wave) are not seen. Therefore, the term quasiperiodic is better used to describe a repetitive biological signal. An example of a quasiperiodic signal is, for instance, ECG. A depolarizing cell, triggered by some stimulus, also generates an electric signal (a depolarization and repolarization wave), which is called a transient signal. Random signals develop during random (or in other words stochastic or statistic) processes. They are generated, for instance, by groups of cells that depolarize in a more or less random fashion, such as muscle cells (generating the electromyogram (EMG)) or nerve cells in the cortex (generating the electroencephalogram (EEG)). The waveshape of such signals bears a nondeterministic character and can be described only in statistical terms. Dependent on the type of biological process, these stochastic signals are stationary or nonstationary. In the case of stationarity, the signal properties do not change over time, for instance, when a patient is in a stable condition. If these properties are changing – signals are called nonstationary. Biosignals are derived from biological processes observed in medicine. Such processes are highly complex and dynamic. Biosignals are usually (not always) a function of time. For example, they can be expressed as s(t), with s being the signal and t the time. Some signals can be described by a few parameters only. A sine wave s(t), for instance, could be defined as s(t) = A sin(t + ). Only three parameters (the amplitude A, the frequency , and the phase ) suffice to fully describe s(t). Once we know these parameters, the  4 signal waveshape is fully determined. So, such characteristics of the biosignal as amplitude and frequency are very important too. Usually biosignals are complex: signal of interest could be a composition of few signals with different amplitudes and frequencies. In this case frequency spectrum (collection of frequencies and corresponding amplitudes or magnitudes) of the signal is analyzed. Examples of biosignals: Electroencephalogram (EEG) – biosignal reflecting electrical activity of neurons. Amplitudes less than 0.1mV, frequency spectrum 3200 Hz.  5 Electrocardiogram (ECG) – reflects electric activity of the heart. Amplitudes less than 5mV, frequency spectrum 0.3-200 Hz. Blood pressure curves – low frequency biosignals of mechanical origin. PCG (phonocardiogram) – represents sounds of heart. Biosignal of mechanical origin. Range of frequencies 30-2000Hz.  6 Biosignal acquisition System for Biological Sensor Electronic signal object interface processing To acquire the biosignal out of the biologic system first we must somehow detect the signal. This is done by sensor. Sensor also converts nonelectric signals to electric so they could be amplified, filtered, and recognized by electronic interface. Biosignals are stored and processed in the system for signal processing. Sensors are special devices that are used to provide an interface between signals from biological systems and the instrument recording them. There are several other words to name sensors: transducers, converters, detectors etc. Sensors can be active and passive. Electric current or potential difference is generated in active sensors. Active sensors are used to convert other kinds of energy to electric power (thermocouple, piezoelectric sensor, etc.). Passive sensors can be used for detection of the biosignal directly or indirectly, when some other parameter associated with the signal is detected. One kind of direct sensors used for detection of electric currents or voltages is called electrodes. It looks like electrodes are the simple kind of sensors. However, electrode is a converter converting ionic currents of biological object to electronic currents. Types of electrodes used for biomedical purposes: 1. Microelectrodes. Conducting small wire, inserted into glass pipette filled with electrolyte solution. Tip of the pipette is open. Such electrodes are used for recording of the potential inside one cell. 2. Small metal wire (diameter of about 1 mm) electrodes can record summary potentials of many cells (about 100 from the cardiac tissue) under the tip of wire. 3. Large metal electrodes as we have used for ECG recording. Noise during recording of biosignals When recording particular biosignal at the same time we can record other signals that are not of our interest. These signals interfere with signals of our interest an are called noise. Noise can be originated from both external and internal sources, for example: Recording devices (sensors, amplifiers...); Electromagnetic interferences (power lines, nearby operating electric equipment); Other signals appearing in the organism, however not related to the recording signal.  7 Here we will discuss mainly noises of electric origin. These noises are induced in our recording system by electric fields. To “clean” recording (or recorded) signal from noise we can use some methods like, for example: Differential amplifier Grounding Electrostatic shield Filtering Differential amplifier is an electronic device with two inputs for voltage (V1 andV2) and one output (Vout). In general its output voltage equals to the voltage difference between input voltages. Usually source of noise is distant from our recording setup. We can connect our electrodes to different inputs of differential amplifier when we record potential difference between these electrodes. In such situation source of noise creates equal potentials at both electrodes. So V1 will be equal to V2. And potential at output induced by noise will be equal to zero.  8 Another method to reduce the noise is grounding. In this case we employ a physical property of the body – capacitance. External electric fields cannot change electric potential of the physical body if the body has a very high electric capacitance. Our planet Earth has for sure very high capacitance. Because of that its potential is considered to be constant. Charges will be able to move freely between our recording object and the ground if we connect them using metal wire. It means, we will have a bit different system (consisting of our object of study and the Earth), however having a very high capacitance. In such case external electric fields (sources of noise) will have less influence to electric properties of the body. Electrostatic shield (or screen) can also be used to reduce the noise (electric). Here we need to know that the total electric field inside an electric conductor is equal to zero. This also applies for the space surrounded by the conductor. More, this surrounding conductor does not have to be solid (continuous). We can use a mesh fabricated from conductor to surround that space. We can reduce the influence of electric fields to the recording  9 signals if we position our biological object inside such a system. So called Faraday cages are used to reduce the noise when acquiring bioelectric signals. Filtering It is possible to reduce the effect of electric noise by using electronic devices – filters. Filters can abolish signals of some particular frequencies an conduct other frequency signals. Filters can be low pass (LPF), high pass (HPF) or band pass (BPF). We can use them if we know frequencies of our recording signals and noises and these are different. If the frequency of noise is higher than frequency of recording signal – lowpass filter could be used. Such a filter will conduct the recording signal and block (filter) the noise. High pass filters could be used if the frequency of noise is lower than the frequency of a signal, because these filters conduct signals with higher frequencies and stop (filter) lower frequency signals. Band pass filters conduct signals inside some range of frequencies.