Transcript Document
By Akshay Dalvi Bimal Paul Vikrant Choudhary What Is Cybernetics? Cybernetics began as the science of communication and control in the animal, machine, and society; i.e. special types of systems. It operates on two levels: study of an observed system & study of the people studying a system. Originated from R & D process in the development of the atomic bomb- applied scientific theory & principles in real-world setting. What Is Medical Cybernetics? • Application of systems and communications theory, connectionism and decision theory on biomedical research and health related questions • Investigates intercausal networks in human biology, medical decision making and information processing structures in the living organism What is it used for? • Physiology and Pathology o Diagnosis, therapy, and prevention of diseases • Medical Informatics o Deals with resources and methods required to optimize the transfer and use of information in health and biomedicine. Applied to areas of nursing, clinical care, dentistry, pharmacy, public health and (bio)medical research. Applications in Medical Field Admissions, Discharges, and Transfers Medication Order Entry IV Order Entry Common Order Entry Bar Code Support Physician Order Entry, Profile Review and Management Unit Dose Support It allows the nurse access to review and print the patient profile, generate MARs at the nursing station, as well as physician order updates, discharge order sheets, and multiple nursing worksheets. Clinical Screening for Drug Interactions, Drug Allergies, Food/Drug interactions, Dosing, and Therapeutic Duplications. Drug Utilization Review Diagnostic Related Grouping (DRG) Statistics and Reporting Inventory Control Statistical Reports and Lists Outpatient Pharmacy Module File Generation and Maintenance Topics In Medical Cybernetics • Systems Theory in Medical Sciences Involves searching for and modeling of physiological dynamics in a organism to gain deeper insights into the organizational principles of life and its disturbance • Medical Information and Communication Theory o Aims to mathematically describe signaling process and information storage in different physiological layers • Connectionism o Aims to describe information processing in neural networks • Medical Decision Theory (MDT) o Aims to gather evidence based of foundations for decision making in the clinical setting o A System Theory Approach to an expanded medical model Fig: Multidimensional Healing: The Clinical Process. What is Medical Information and Communication Theory? Medical Information and Communication Theory in Medical Cybernetics mathematically describes the signal transfer processes in different physiological layers. Basics - Information typically managed through a combination of cognitive memory and paper based systems - Technology comes along, aides in tedious tasks such as financing - Computing allows for complex communications - Significantly changes the attitude of communication. - Two forms of communication, horizontal and vertical -Horizontal-> at the process level, linking activities together, supporting front line decision makers, and enabling efficient flow of business -Vertical-> management between top level and the intermediate levels of computing. What Is Connectionism? • Connectionism is a set of approaches that models mental or behavioral phenomena as the emergent process of interconnected networks of simple units. o Most commonly used model of connectionism is the neural network. • The central principle of connectionism is that mental development can be described by interconnected networks of simple units. However the form of connections and units can vary from model to model. Neural Networks A Neural network is a massively distributed processor that has a natural propensity for storing experimental knowledge and making it available for use. It resembles the human brain in 2 respects Knowledge acquired though a learning process Interneuron connection strengths known as synaptic weights store knowledge. Biological Neuron Artificial Neuron Transfer Function Weighted summation Activation Function Threshold Function OUT=1 OUT=0 if NET>T otherwise Squashing Function OUT=1/(1+ ) NET e Hyperbolic Tangent Function OUT=tanh(NET) Recurrent Neural Network Training Objective- Application of a set of inputs produces desired(or at least consistent) set of outputs. Types 1. Supervised Training 2. Unsupervised Training 3. Reinforcement Training Neural Networks + WSN in the medical field What exists? What can be done? What Is Decision Theory? • Aims to identify various issues relevant in a decision and its rationality. • Medical uses include: o Medical Diagnosis o Clinical Decision Support Systems (CDSS) Health Informatics Deals with the resources required to optimize the retrieval and use of information in health and biomedicine Clinical Decision Support Systems • Computer programs designed to assist physicians and other health professionals with decision making tasks. • Two main types: Knowledge Based Make use of huge database and decision trees. o Non-knowledge based Commonly use neural networks or genetic algorithms to find patterns in clinical data. o APPLICATION WIRELESS BIOMEDICAL SENSORS CODE BLUE : HARVARD UNIVERSITY CodeBlue Harvard University in collaboration with various medical facilities introduce CodeBlue CodeBlue - An ad hoc WSN Infrastructure for Emergency Medical Care Goal – “Enhance first-responders’ ability to access patients on scene, ensure seamless transfer of data among caregivers, and facilitate efficient allocation of hospital resources” CodeBlue: VitalDust Wearable wireless pulse oximeter and 2-lead Electrocardiogram Monitor (EKG) Collect heart rate (HR), blood oxygen saturation (SpO2), and hearts electrical activity Devices can be programmed to alert medical personnel when vital signs fall outside normal conditions CodeBlue: VitalDust Implementation Pulse Oximeter Mote-based oximeter connector between Mica2/MicaZ mote platform and the BCI Medical board Measures the amount of light transmitted through a noninvasive sensor attached to the patient’s finger CodeBlue: VitalDust Implementation EKG Mote-based EKG consists of a custom built circuit board attached to a Mica2/MicaZ/Telos mote Measures hearts’ electrical activity through a set of leads attached to a patients heart at a rate of 120 Hz CodeBlue: Pluto Wearable tag wristband Stores patient information and tracks patient location using radio-frequency (RF) signals Mote includes an external push button that can be used by a patient to transmit a one-way alert to medical staff CodeBlue Infrastructure Challenges Communication Challenges Secure, reliable, ad hoc communication among groups of sensors and mobile devices Prioritize transmission of data Computational Challenges Computational power Security and encryption techniques Programming Challenges Level of software services Conclusion Extremely beneficial in disaster response scenarios Real Time Monitoring Requires efficiency and accuracy improvement A step up in saving lives, creating valuable medical research data, and allocation of medical resources References • http://www.medical-cybernetics.de/definition.html (3/7/2010) • http://en.wikipedia.org/wiki/Medical_cybernetics (3/7/2010) • S. Russell and P. Norvig. Artificial Intelligence: A Modern Approach, 2nd ed. New Jersey: Prentice Hall, 2003. pp. 724-748. • A. Hart. “Using Neural Networks for Classification Tasks – Some experiments on Datasets and Practical Advice,” The Journal of Operational Research Society, vol. 43, no. 3, pp. 215-226, Mar. 1992. • Raul Rojas. Neural Networks – A Systematic Introduction, NY, 1996. pp. 337-374 References Continued • M. Stensmo and T.J. Sejnowski. "Automated Medical Diagnosis based on Decision Theory and Learning from Cases," World Congress on Neural Networks 1996 International Neural Network Society, pp.12271231. • http://en.wikipedia.org/wiki/Clinical_decision_support_system (03/14/2010) • Code Blue: An ad hoc sensor network infrastructure for emergency medical care http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.62.5465