Bioinstrumentation and Biosensors Notes

Bioinstrumentation and Biosensors

Bioinstrumentation is instrumentation that can be used to measure or interact with biological phenomena happening in the body and/or interact with them.

Types of Bioinstrumentation

Diagnostic: Measures and provides information for the physician to improve medical decision-making. For example, using an ECG to diagnose heart arrhythmias.
Therapeutic: Manipulates physiological processes. Examples include pacemakers and defibrillators.
Assistive: Replaces lost or diminished function, typically in rehabilitative medicine. Examples include prostheses with sensory and functional capabilities.

Advancement of Medical Technology

1903: First ECG (Einthoven)
1929: First EEG (Berger)

Implantable Cardioverter-Defibrillator (ICD)

An ICD detects a rapid heartbeat coming from the bottom of the heart.

Basic Bioinstrumentation System

A bioinstrument consists of similar components as any instrument for measurement, but its signals come from living tissue.

Essential Parts of a Bioinstrumentation System

Measurand: The physical quantity, property, or condition that the system is intended to measure.
- Medically important measurands include: biopotential, pressure, flow, position (imaging), displacement (velocity, acceleration, and force), impedance, temperature, and chemical concentrations.
Transducers: Convert mechanical/chemical/optical properties into electrical quantities.
- A transducer converts energy from one form to another.
Sensor: A primary sensing element (e.g., a capacitance that varies with displacement) is converted to the desired variable, usually voltage.
Signal Conditioning: Transforms the signal into the right range for data acquisition and conversion to digital form (via amplification) and reduces undesirable signal components like noise or interference (via filtering). For example, ECG electrodes.
Output Display: Displays measurement results in a way that readily imparts the most useful information to the human operator.

Optional Parts of a Bioinstrumentation System

Calibration
Feedback control
Data storage
Data transmission

Analog vs. Digital Signals

Analog: Continuous, able to take on any value within the dynamic range.
Digital: Discrete, able to take on only a finite number of different values.
Digital signals are easier to process (e.g., by computers), thus most signals are digitized after acquisition.

Challenges Specific to Biomedical Instrumentation

Measurement is often indirect because the measurand is an internal process.
Physiological quantities are highly variable across and within patients and are interconnected, appearing non-deterministic.
Medical and physiological parameters tend to be very small in almost all physical quantities.
Due to indirect measurement and the complexity of biological phenomena, data needs further analysis to be useful for clinicians.

Process of Measurement

Given a physiological event, what measurable variable reflects it and how:
- Direct vs. indirect measurement.
  - Electromyography is a direct measurement of the electricity that flows in a muscle and an indirect measurement of muscular activation.
  - Electrocardiography is a direct measurement of electricity that propagates from the heart and an indirect measurement of cardiac activity.
What are the possible sources of error?
- Inherently dependent on the physical property that we want to measure.
- Error due to mistakes, hardware limitations, or unavoidable noise.

Definitions

Resolution: The largest change in input $I$ that can occur without any corresponding change in output $O$ .
Accuracy: How close the measured value is to the true value.
Precision: How close two measurements of the same true value are to each other.
Sensitivity: The smallest change in input signal that causes the measuring device to respond.
Specificity: How specific the measurand is to the phenomenon.
Linearity and hysteresis: How linear the output is with respect to the input and how similar the behavior is between increasing and decreasing inputs.
Signal-to-noise ratio: How much of what I am measuring is useful information and how much is noise.

Accuracy and Precision

When trying to measure the true mean of a random variable/process in the presence of noise, the measurement is taken several times and averaged.
Accuracy refers to how close the average is to the true value of the mean.
Precision refers to the variability across measurements (usually the standard deviation).
Poor precision results from random errors.
Poor accuracy results from systematic errors.

Measurement Specifications

What is the amplitude range of the selected variable?
What is the bandwidth? (How rapidly and how slowly can the variable change?)
What is required resolution (smallest change you need to measure)?
What is required accuracy?
Sensitivity, reliability, linearity?

Design Process for Medical Instruments

Factors to consider include:

Measurand signal factors such as sensitivity, range, differential or absolute input, input impedance, transient and frequency response, accuracy, linearity, reliability, and specificity.
Environmental factors such as signal-to-noise ratio, stability (temperature, humidity, pressure, acceleration, shock, vibration, radiation), power requirements, mounting size, shape, invasive or non-invasive nature, and tissue-sensor interface requirements.
Medical factors such as material toxicity, electrical safety, radiation and heat dissipation, patient discomfort.
Economic factors such as cost, production, availability, warranty, consumable requirements, and compatibility with existing equipment.
Regulatory requirements such as FDA (in the US), Medical Device Directive, and European Medicines Agency (in Europe).

SI Base and Coherent Derived Units

The slide lists SI base units and coherent derived units with special names and symbols, covering various physical quantities such as mass, length, time, force, pressure, energy, power, frequency, and more.

Measurement Basics and Interference

Biopotentials are always recorded as the potential between two points in the body.
Electrode-skin interface impedances are to be considered.
The transducer and signal conditioning circuits are designed to remove interference as much as possible.
A differential amplifier amplifies the difference between two signals, thus rejecting interference that is identical between signals.

Calibration

Necessary for measurements that directly relate to physical quantities.
Translates the measured quantity (e.g., a voltage) to a different physical quantity directly related to the measurand.
For example, a force platform measures forces and moments applied in three directions but measures the deformation of the pylons holding the platform as changes in voltage in a Wheatstone bridge circuit.
By applying known forces and moments to the platform, you can map the voltages to the real value and generate a calibration matrix.

Error

How close to the true value of the measurand is the value indicated by the instrument output?
In practice, we can never precisely know the error for any one measurement since the true value is unknown.
Calibration enables us to at least specify the uncertainty associated with a measurement. A useful way to do this is by specifying a 95% confidence interval – the range of values within which the real measurand lies with a probability of 0.95.
Error = Measured - True

Bias (Systematic or Fixed) Error

Bias errors are constant average errors across multiple measurements.
Bias Error = Average of Indicated Values - True Value
Precision error refers to the random variability across successive measurements.

Normal Distribution

By grouping the values of $N$ measurements of the same quantity, you can estimate the distribution of the random variable associated with the measurement.
In most cases, measurements will follow a normal (Gaussian) distribution which is defined by its mean ( $m$ ) and standard deviation ( $s$ ).
Range (min, max)
Standard deviation $s$ is the default, and the 95% C.I. is simply $mean ± (1.96)(s)$ .

Common Sources of Each Error Type

Bias or Systematic Errors
- Calibration errors
- Offset error (zero offset errors, scale errors)
- Non-linearity error
- Certain consistently recurring human errors
Precision Errors
- Random human errors
- Limitation of system resolution (like rounding error)
- Errors caused by random fluctuations of experimental conditions

Electrograms

Electroneurogram (ENG) - Peripheral Nervous System
Electromyogram (EMG) - Muscle
Electrocardiogram (ECG) - Heart
Electrooculogram (EOG) - Eyes
Electroencephalogram (EEG) - Central Nervous System

Differential Amplification

In noninvasive biopotential measurements (e.g., ECG), we are trying to record tiny electrical fluctuations generated inside the body that are swamped in a huge amount of noise.
Selectively amplifying the difference between two electrodes placed on the body gets rid of all of the noise that is common to both electrodes – e.g., electromagnetic interference.
Multiple-electrode recordings are also generally desirable from the point of view of measuring neurophysiological activity comprehensively.
Designing amplifiers that reject common-mode voltages is thus of huge importance in biomedical applications.

Instrumentation Amplifier

First stage: buffer amplifiers to adjust the input impedance
Second stage: differential amplifier

Active/Passive Sensors

Active: Require an external energy source. Example: Petrol Tank Indicator
Passive: Do not require an external energy source. Example: Pressure Indicator

Transducers – Some Examples

Electrodes: From changes in potential on the skin to currents
Potentiometers: Change in resistance
Strain gauges: Deformation to change in resistance
Capacitive sensors: Motion to change in capacitance
Piezo-electric or piezo-resistive materials: Mechanical stress to current/change in resistance

Electrode-Electrolyte Interface

The interaction of metallic ions at the electrode/electrolyte interface generates a current.
The concentration difference driving metal atoms to cross the interface generates a double-layer of charge to achieve equilibrium, modeled by a voltage source and capacitor, respectively.
$Rd$ represents the resistance of the double-layer, and $Rs$ represents the resistance of the electrolyte.
Different layers of the skin act as filters to the current.

Potentiometers

Can be easily made to track angles/positions.
Each angle corresponds with a resistance value.
Easy to calibrate.

Strain Gauge

Strain gauges can track very small movements (e.g., millimeters).
The resistance of a fine wire changes when it is stretched or compressed.
The resistance of a wire is equal to: $R = frac{ρL}{A}$ , where:
- $L$ = length
- $A$ = section
- $ρ$ = resistivity
Change in the parameters lead to change in the resistance.

Wheatstone Bridge

Common circuit that is used to track small changes in the value of a resistor in a circuit.
$V_{out}$ is 0 if the resistance on each side has the same ratio.
A change in one resistance drives a change in $V_{out}$ .

Capacitive Sensors

Capacitance of a parallel-plate capacitor: $C = frac{ε0 εr A}{x}$ , where:
- $ε_0$ is the dielectric constant of free space,
- $ε_r$ the relative dielectric constant of the insulator (air = 1),
- $A$ is the area,
- $x$ the separation of the plates. $x$ is convenient to change.
Thus, reading the voltage drop gives a measure of $x$ .

Piezoelectric Sensors

Piezoelectricity works on the principle that when an asymmetrical crystal lattice is distorted or mechanically strained, positive and negative charges in the material are relatively displaced toward opposite surfaces of the crystal, causing a voltage to be measured by electrodes attached to the surfaces.
Assuming initially that no charge can leak between surfaces (so the piezoelectric element behaves like a parallel plate capacitor), the total induced charge $q$ is directly proportional to the applied force $f$ : $q = Sf$ , where $S$ is a constant.
Thus, by capacitor relations: $V = frac{q}{C}$ .

Inertial Measuring Units (IMU)

Units that include different sensors and integrate their information to estimate movement. Usually include:

Accelerometer (acceleration)
Gyroscope (velocity)
Magnetometer (orientation/relative position)

Accelerometers

Measure acceleration
Mono-axial (one direction of acceleration)
Tri-axial
Example of a simple acceleration measuring system
Quasi-Static Accelerometers:
- $F = kx = ma$ , where
  - $k$ is spring constant
  - $x$ is displacement
  - $m$ is mass
  - $a$ is acceleration
$a = frac{k}{m} ⋅ x = frac{w_0^2}{x}$

Accelerometers

Accelerometers are usually divided in:

Piezoelectric
Piezoresistive
Capacitive

Electrical properties of material change due to mechanical stress.

Accelerometers - MEMS (Micro-Electro-Mechanical Systems)

Currently most used design. Allow for biggest miniaturization. A deformable beam acts as a central plate of a capacitor.