Signals, Systems, and Biomedical Engineering

Background Information

Biomedical signal processing aims at extracting significant information from physiological signals, which includes:

• heart rate
• blood pressure
• oxygen saturation levels
• blood glucose
• nerve conduction
• brain activity

These signals can then be analyzed in order to provide information to physicians about what is going on in the body and allows them to make a diagnosis if any sort of abnormality is detected. This ultimately allows one to determine the state of a patient's health through non-invasive measures.

As technology is improving, engineers are discovering new ways to provide information to clinicians upon which they can make decisions. One of these improvements is through real-time monitoring, which can lead to the better management of chronic diseases, earlier diagnosis of disease, and earlier detection of both heart attacks and strokes. Biomedical signal processing is most useful in the critical care setting due to patient data needing to be analyzed in real-time. By doing complex analyses of the body's signals and having real-time monitoring, we can discover early indicators for how conditions manifest.

For more information about signals, systems, and biomedical engineering, follow these links:

• Biomedical Signal and Image Processing
• Advanced Methods of Biomedical Signal Processing
• Biomedical Signal Processing
• Introduction to Biomedical Signals
• Biosignal Processing

Examples of Biomedical Signals and Signal Processing

Electrical Biosignals: The electrical signals produced by the body that are extremely useful in diagnostics includes:

• Electroencephalogram (EEG) - Monitoring method to record electrical activity of the brain. It is most often used to diagnose epilepsy but can be used to diagnose other conditions such as: sleep disorders, tumors, stroke, coma, encephalopathies, brain death, etc.

For more information on EEG, click here

• Electrocardiogram (ECG) - Monitoring method to record electrical activity of the heart. Indications for performing ECG includes: suspected myocardial infarction, suspected pulmonary embolism, seizures, fainting, cardiac murmur, etc.

For more information on ECG, click here

• Electromyogram (EMG) - Monitoring method to record electrical activity of the muscle. EMG is used as a diagnostic tool for identifying neuromuscular diseases (Parkinson's, multiple sclerosis, Huntington's, etc) or as a research tool.

For more information on EMG, click here

All three signals listed above require filtering of background noise (power source, other biosignals, etc.) and often require conversion from continuous time (CT) to discrete time (DT) for analysis.

Biosensors: A biosensor is defined as a piece of hardware that can interact with either a biological or physiological system to acquire a signal for diagnostic or therapeutic purposes. They are analytical devices that convert a biological response into an electric signal. Biosensor technology incorporates a wide range of devices, which includes:

• Stethoscope
• Thermometer
• Blood Pressure Cuff
• Blood Glucose Device
• Pregnancy Test
• Pulse Oximetry

To learn more about biosensors, click here

Imaging: Biomedical imaging concentrates on the capture of images for both diagnostic and therapeutic purposes. Snapshots of in-vivo physiology and physiologic processes can be collected through advances sensors and computer technology. Biomedical imaging technologies include:

• X-ray
• Magnetic Resonance Imaging (MRI)
• Ultrasonography
• Positron Emission Tomography (PET)
• Single Photon Emission Computed Tomography (SPECT)
• Optical Coherence Tomography (OCT)
• Computed Tomography (CT)

Biomedical imaging processing is very similar in concept to biomedical signal processing due to them both including enhancement, analysis, and display of results.

To learn more about imaging, click here

Wearable and Implantable Technologies: Both wearable and implantable technologies sense parameters of various conditions (diseases, disorders, etc.) and can either transfer the data to a remote center (hospital, physicians office, etc.), direct the patient to take specific action, or automatically perform a function based on what the sensors are reading. Examples of these include:

• Blood glucose - If blood glucose is running high, insulin could be automatically administered from the device
• Cardiac Monitoring - Pacemaker is placed just under the skin to help control abnormal heart rhythms
• Parkinson's - Deep brain simulation sends electrical signals to brain areas responsible for body movement
• Smart Tattoos - Flat, flexible, stretchable electronic sensor placed on the skin to measure various electric signals produced by the heart, brain, and muscles

To learn more about wearable and implantable technologies, click here

Purdue Biomedical Signal Processing Classes:

(Course Descriptions are Provided from Purdue Catalog)

• BME 511 - Biosignal Processing

An introduction to the application of digital signal processing to practical problems involving biomedical signals and systems. Topic include: examples of biomedical signals; analysis of concurrent, coupled, and correlated processes; filtering for removal of artifacts; event detection, analysis of waveshape and waveform complexity; frequency domain characterization of signals and systems; modeling biomedical signal-generating processes and systems; analysis of nonstationary signals; pattern classification and diagnostic decision. MATLAB will be used throughout to provide numerous opportunities for hands-on application of the theory and techniques discussed to real-life biomedical signals.

• BME 495/430 - Biomedical Imaging Modalities

This course covers basic principles and modes of bioimaging methods for biomedical sciences. Topics include interaction of electromagnetic radiation with tissue, basic concepts in imaging and detection, basic modes of imaging modalities (e.g. reflection, transmission, absorption, and emission), and basic image processing/analysis. Model systems to be used to teach the topics include conventional imaging modalities such as optical imaging, optical microscopy, X-ray, computed tomography, ultrasound, magnetic resonance imaging, etc. This course also includes hands-on exercise that reinforces important concepts.

• BME 495 - Deep Learning and Medical Imaging

This course sets the foundation of computational neuroscience: a branch of neuroscience that creates computable models of biological neural systems, in particular large scale neural networks for processing sensory information. The course builds on basic neural modeling, presents computable neuron models and extends to large networks of neurons. Participation in the class will also help students how to use and write efficient software models of mammalian somatosensory systems, with a focus on oriented models, and how to trade one for the other when efficiency is needed in very large networks (> 1 million). Additionally, the course is deeply rooted in machine learning, and supervised and unsupervised learning systems. The application will be centered in synthetic and artificial vision and audition, perception, intelligence for robots and automatic systems. Lecture will include an overview of the state-of-the-art in the field, new opportunities and ideas for innovation and success with such systems. Systems-level lectures will be in the form of recent paper review and discussion.

• BME 521/ ABE 560 - Biosensors: Fundamentals and Applications

An introduction to the field of biosensors and an in-depth and quantitative view of device design and performance analysis. An overview of the current state of the art to enable continuation into advanced biosensor work and design. Topics emphasize biomedical, bioprocessing, environmental, food safety, and biosecurity applications.

• BME 595 - Functional MRI Applications

This course focuses on the principles and applications of various established and emerging technologies for imaging brain activity in vivo across a wide range of spatial and temporal scales. It covers functional magnetic resonance imaging, positron emission tomography, single-photon emission computed tomography, electroencephalography, magnetoencephalography, diffuse optical tomography, intrinsic signal optical imaging, voltage sensitive dye imaging, two-photo calcium imaging, functional ultrasound, and photoacoustic tomography, all in the context of brain functions. Special emphasis is on the pros and cons of individual modalities, and their integration toward more comprehensive understandings of how human sensation, behavior, and cognition emerge from complex network activity. The course will also introduce advanced topics, such as machine learning for functional imaging data, contrast-agent based cellular and molecular imaging in the brain, and portable devices for functional neuroimaging in realistic environments.

• BME 595 - Medical Imaging and Diagnostic Technologies

This new gateway course will provide an introduction to the physics, technologies, and biological considerations associated with modern imaging and diagnostic technologies. Physics underlying image formation and the interaction of biological tissue with associated energies will be emphasized. Specific modalities to be examined will include x-ray, nuclear imaging, ultrasound, optical tomography, MRI and mass spectrometry.

• BME 595 - MRI Theory

This course is an introduction to the theory and design of magnetic resonance imaging systems, with an emphasis on theory from a physics perspective. Mathematical derivations of fundamental principles will be explored. Topics include image acquisition and reconstruction, mechanisms for image contrast and resolution, and an overview of system design, including magnets, gradients, and radiofrequency coils.

• ECE 438 - Digital Signal Processing with Applications

The course is presented in five units. Foundations: the review of continuous-time and discrete-time signals and spectral analysis; design of finite impulse response and infinite impulse response digital filters; processing of random signals. Speech processing; vocal tract models and characteristics of the speech waveform; short-time spectral analysis and synthesis; linear predictive coding. Image processing: two-dimensional signals, systems and spectral analysis; image enhancement; image coding; and image reconstruction. The laboratory experiments are closely coordinated with each unit. Throughout the course, the integration of digital signal processing concepts in a design environment is emphasized.

• ECE 510 - Introduction to Biometrics

Biometrics is an emerging technology for automatic human identification and verification using unique biological traits. Compared to traditional identification and verification methods, biometrics is more convenient for users, reduces fraud, and is more secure. It is becoming an important ally of security, intelligence, law enforcement, and e-commerce. The principle of various biometric technologies and systems is introduced. Especially, students analyze and design fingerprint recognition, face recognition, iris recognition, voice recognition, and multimodal biometric systems. Students have hands-on experience in designing and analyzing biometric systems.