ASIC or PIC? Implantable stimulators based on semi-custom CMOS technology or low-power microcontroller architecture.

To gain a better understanding of the effects of chronic stimulation on mammalian muscles we needed to generate patterns of greater variety and complexity than simple constant-frequency or burst patterns. We describe here two approaches to the design of implantable neuromuscular stimulators that can satisfy these requirements. Devices of both types were developed and used in long-term experiments. The first device was based on a semi-custom Application Specific Integrated Circuit (ASIC). This approach has the advantage that the circuit can be completely tested at every stage of development and production, assuring a high degree of reliability. It has the drawback of inflexibility: the patterns are produced by state machines implemented in silicon, so each new set of patterns requires a fresh production run, which is costly and time-consuming. The second device was based on a commercial microcontroller (Microchip PIC16C84). The functionality of this type of circuit is specified in software rather than in silicon hardware, allowing a single device to be programmed for different functions. With the use of features designed to improve fault-tolerance we found this approach to be as reliable as that based on ASICs. The encapsulated devices can easily be accommodated subcutaneously on the flank of a rabbit and a recent version is small enough to implant into the peritoneal cavity of rats. The current devices are programmed with a predetermined set of 12 patterns before assembly; the desired pattern is selected after implantation with an electronic flash gun. The operating current drain is less than 40 microA.

Development of an accurate portable recording peak-flow meter for the diagnosis of asthma.

This article describes the systematic design of an electronic recording peak expiratory flow (PEF) meter to provide accurate data for the diagnosis of occupational asthma. Traditional diagnosis of asthma relies on accurate data of PEF tests performed by the patients in their own homes and places of work. Unfortunately there are high error rates in data produced and recorded by the patient, most of these are transcription errors and some patients falsify their records. The PEF measurement itself is not effort independent, the data produced depending on the way in which the patient performs the test. Patients are taught how to perform the test giving maximal effort to the expiration being measured. If the measurement is performed incorrectly then errors will occur. Accurate data can be produced if an electronically recording PEF instrument is developed, thus freeing the patient from the task of recording the test data. It should also be capable of determining whether the PEF measurement has been correctly performed. A requirement specification for a recording PEF meter was produced. A commercially available electronic PEF meter was modified to provide the functions required for accurate serial recording of the measurements produced by the patients. This is now being used in three hospitals in the West Midlands for investigations into the diagnosis of occupational asthma. In investigating current methods of measuring PEF and other pulmonary quantities a greater understanding was obtained of the limitations of current methods of measurement, and quantities being measured.(ABSTRACT TRUNCATED AT 250 WORDS)

The accuracy of portable peak flow meters.

BACKGROUND: The variability of peak expiratory flow (PEF) is now commonly used in the diagnosis and management of asthma. It is essential for PEF meters to have a linear response in order to obtain an unbiased measurement of PEF variability. As the accuracy and linearity of portable PEF meters have not been rigorously tested in recent years this aspect of their performance has been investigated. METHODS: The response of several portable PEF meters was tested with absolute standards of flow generated by a computer driven, servo controlled pump and their response was compared with that of a pneumotachograph. RESULTS: For each device tested the readings were highly repeatable to within the limits of accuracy with which the pointer position can be assessed by eye. The between instrument variation in reading for six identical devices expressed as a 95% confidence limit was, on average across the range of flows, +/- 8.5 l/min for the Mini-Wright, +/- 7.9 l/min for the Vitalograph, and +/- 6.4 l/min for the Ferraris. PEF meters based on the Wright meter all had similar error profiles with overreading of up to 80 l/min in the mid flow range from 300 to 500 l/min. This overreading was greatest for the Mini-Wright and Ferraris devices, and less so for the original Wright and Vitalograph meters. A Micro-Medical Turbine meter was accurate up to 400 l/min and then began to underread by up to 60 l/min at 720 l/min. For the low range devices the Vitalograph device was accurate to within 10 l/min up to 200 l/min, with the Mini-Wright overreading by up to 30 l/min above 150 l/min. CONCLUSION: Although the Mini-Wright, Ferraris, and Vitalograph meters gave remarkably repeatable results their error profiles for the full range meters will lead to important errors in recording PEF variability. This may lead to incorrect diagnosis and bias in implementing strategies of asthma treatment based on PEF measurement.

Production of reliability proven physiological data for use in automated monitoring and diagnostic systems.

In recent years manufacturers of intensive care monitoring systems have introduced complex digital processing architectures that theoretically have enormous processing power. This power should allow the realization of many useful processing methodologies that up to now have only been research tools, e.g. the generation of reliable alarms, the implementation of predictive monitoring strategies and reliable diagnostic and treatment guidance to the clinical staff. However, before any of these methodologies can be successfully initiated, each must have accurate and reliable derived physiological data available to them, e.g. beat-by-beat heart rate and blood pressure. From the very nature of monitoring physiological quantities there will be much misinformation or 'noise' superimposed on the raw signal obtained from the patient. The major source of noise (as far as electrocardiogram (ECG) monitoring is concerned) is internal to the body and is electromyographic noise. This results from the contraction of skeletal muscles producing action potentials of similar magnitude and frequency to that of the ECG. Fortunately, nursing staff are very good at 'filtering out' any misinformation before recording any data (on a ward chart for instance). However, in completely automated systems, if this noise is not detected and eliminated or compensated for at an early stage in the processing chain, misinformation will result with potentially serious consequences. The recognition and elimination of such noise cannot be readily achieved using standard filtering techniques without serious degradation of information. This paper discusses the potential of modern digital system architectures developed for ECG monitoring. It analyses the noise that occurs on this physiological variable and demonstrates a novel method of eliminating such noise.

The real time processing of dynamic physiological signals.

A signal processing system has been developed consisting of a microcomputer linked to a minicomputer. The system has been designed to allow ease of development of programs for processing physiological signals in real time. The system also allows extensive non real time processing to be carried out and the generation of graphical displays. Some of the problems involved in producing real time signal processing programs are discussed. The problem of computing the beat by beat mean pressure in real time is described as an example.

Short-term variability of pulse rate and blood pressure in cardiac surgery patients.

The magnitude and mathematical nature of short-term variations of pulse rate and mean arterial blood pressure has been studied in 27 post-operative cardiac surgical patients over a continuous period of 200 min in each case. Similar variability was observed in all patients. Short-term variations were composed predominantly of a series of rhythmic changes ranging from one synchronous with respiration to others between 2 and 5 min in cycle length. There was consistent variance of the beat-by-beat values for both variables about a continuously updating 5 min mean. The average standard deviation was 3.75 beats/min for pulse rate and 0.48 kPa (3.64 mm Hg) for mean arterial blood pressure. For both variables the distribution about the 5-min mean did not differ significantly from a normal distribution in 50 out of the 54 records. These findings have implications for the reproducibility of current methods of estimating mean pulse rate and blood pressure, and the change necessary before two estimates may be regarded as significantly different. The results are applicable both to the interpretation of Ward Charts by medical staff and to automated monitoring systems.

Trend detection of pseudo-random variables using a exponentially mapped past statistical approach: an adjunct to computer assisted monitoring.

The early detection of deterioration in a patient depends on recognising trends: a capability poorly developed in monitor systems. Using exponentially mapped past (EMP) statistical variables, it has been proved that the difference between two EMP means is a function of only one variable, trend of input. Some of the signal variance appears on the difference signal, and so interpretation depends on a statistical approach based on knowledge of the normal variability of the monitored variable. The programs developed and built in a dedicated form utilise parallel hybrid computation of continuous data, but the principles are equally applicable to digital techniques with discrete data.