Abstract
A sudden unexpected death during epilepsy (SUDEP) can occur to an unattended epilepsy sufferer. The cessation of heart and breathing activity is associated with SUDEP. Therefore, developing a heart and breathing monitoring system that could alarm when an epilepsy patient experiences cessation of heart or breathing activity is necessary. This system will assist healthcare providers to respond to the emergency and prevent sudden death. This collaborative project is to investigate the feasibility of developing a miniature, unobtrusive, and self-contained module that an epilepsy patient could wear in bed and a small bedside alarm unit intended to alert healthcare providers to prevent SUDEP. The wearable unit communicates wirelessly occurrences of breaths and heartbeats to the bedside unit. The bedside unit alerts upon cessation of heart or breathing activity. This project assumed that the monitor receives an input voltage from a small strip of piezoelectric film made of polyvinylidine fluoride (PVDF) attached to the patient’s chest. The PVDF sensor voltage changes in response both to the chest expansion and contraction of breathing and to the chest vibrations caused by heartbeats. In the first stage of this project, human subject recordings of PVDF, strain-gauge plethysmography (SPG),and photoplethysmography (PPG) signals were analyzed to determine how to design the signal processing needed to separate out individual breathing and heartbeat signals. Then, low-power analog, microcontroller, and wireless technologies were used to build a prototype of the wearable module to separate out the breathing and heartbeat signals, detect individual breaths and heartbeats, and telemeter the detected breaths and heartbeats to the bedside unit. A prototype of the bedside unit was built using similar technologies to receive the occurrences of heartbeats and breaths from the wearable module. This unit also calculates heart and breathing rates and determines whether or not to alarm if breaths or heartbeats cease or the wireless link fails to function. For convenience, the wearable module and bedside unit were prototyped using bread-boarded circuitry and commercial development boards containing the miniature electronic components. Once feasibility is shown, the wearable module and bedside unit can be compacted by incorporating the miniature electronic components directly into custom integrated circuits. The performance of the signal processing algorithm to separate out breathing and heartbeat signals and to detect individual breaths and heartbeats was tested by using the algorithm to process PVDF recordings from human subjects and comparing the results with SPG recordings of breaths and PPG recordings of heartbeats from the same human subjects. Miss Swetha Seshu Ayyagari designed the hardware for the signal conditioning circuit and the software for digital signal processing to separate out the breathing and heartbeat signals from the PVDF signal. Mr. Pritam Rajendra Chopda designed the software to detect individual breaths and heartbeats, transmit detected breaths and heartbeats to the bedside unit, calculate heart and breathing rate, and alert upon cessation of heartbeats or breathing or wireless link failure. Prototypes were successfully built of the wearable module and bedside unit. The prototype wearable module was capable of processing an incoming PVDF signal, separating out breathing and heartbeat signals, detecting individual breaths and heartbeats, and wirelessly transmitting them to the bedside unit. The prototype bedside unit was capable of alarming if breaths or heartbeats stopped or if the wireless link failed. On six 22-s PVDF recordings, percentage of peaks correctly detected as breaths (true positive) is 64%, and the percentage of peaks falsely detected as breaths (false positive) is 0%. On six 22-s PVDF recordings,percentage of peaks correctly detected as heartbeats (true positive) is 80%,and the percentage of peaks falsely detected as heartbeats (false positive) is 4%. No attempt was made in the prototype system to adapt to changing PVDF signal amplitudes from subject to subject. Such adaptation is needed before the monitoring system is useful for actual patients. No measurements were made of the power consumption of the wearable module electronics. Therefore, at present, battery size to achieve a given operating time is not known. No measurements were made of whether the wireless link successfully functions when the wearable module was worn by a sleeping subject changing position in bed. Such studies need to be done to determine if the monitoring system will work in actual home settings.