PubMed HealthSearch

Biomedical subjects

J A Orr

Publications and source records attributed to J A Orr.

At least 19 recordsLinked to original sources

A parallel implementation of the backward error propagation neural network training algorithm: experiments in event identification.

An artificial neural-network-based (ANN) event detection and alarm generation system has been developed to aid clinicians in the identification of critical events commonly occurring in the anesthesia breathing circuit. To detect breathing circuit problems, the system monitored CO2 gas concentration, gas flow, and airway pressure. Various parameters were extracted from each of these input waveforms and fed into an artificial neural network. To develop truly robust ANNs, investigators are required to train their networks on large training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from 6.4 to 1.5 h. By reducing the total run time of the computation through parallelism, we were able to optimize many of the neural network's initial parameters. We conclude that use of the master-worker model of parallel computation is an excellent method for speeding up the backward error propagation neural network training algorithm.

Algorithms

Effects of the thromboxane A2 mimetic, U46,619, on pulmonary vagal afferents in the cat.

Release of thromboxane A2 (TxA2) or infusion of the TxA2 mimetic U46,619 in the cat elicits pulmonary hypertension and rapid shallow breathing (Shams et al., Respir. Physiol. 71: 169-183, 1988). The vagus nerve mediates the observed respiratory, but not the circulatory, effects (Shams and Scheid, J. Appl. Physiol. 68: 2042-2046, 1990). To identify the type of lung vagal afferent fibers involved in this respiratory response to TxA2, we have recorded the functional single-unit activity and its response to infusion of U46,619 in fine strands of the vagus nerve in the artificially ventilated cat and rabbit. The fibers were classified as originating from slowly adapting (SAR) or rapidly adapting (RAR) stretch receptors by their response to sustained pulmonary inflation (intrapulmonary pressure of 20-25 cmH2O) or as C-fibers, by their response to a bolus injection of phenylbiguanide. C-fibers responded variably to lung inflation. U46,619 infusion caused only a small increase in SAR or RAR activity along with increases in end-inspiratory tracheal airway pressure (Paw), but evoked a marked increase in the firing rate of C-fibers, independent of their response to lung inflation. This increase in C-fiber activity was unrelated to the increase in Paw, which accompanied the infusion of U46,619. Since these responses remained the same after indomethacin they appear to be due to a direct action of U46,619, and not to be mediated by prostanoids that might be released by U46,619. These data suggest that C-fibers are indeed involved in the respiratory effects of TxA2. Since the effects exerted on C-fibers by U46,619 were unrelated to increased Paw, TxA2 is likely to stimulate the nerve endings directly, rather than via smooth muscle contraction. On the other hand, the small stimulating effect of U46,619 on SAR and RAR may be mediated by bronchoconstriction.

15-Hydroxy-11 alpha,9 alpha-(epoxymethano)prosta-5

Intelligent alarms reduce anesthesiologist's response time to critical faults.

The proliferation of monitors and alarms in the operating room may lead to increased confusion and misdiagnosis unless the information provided is better organized. Intelligent alarm systems are being developed to organize these alarms, on the assumption that they will shorten the time anesthesiologists need to detect and correct faults. This study compared the human response time (the time between the sounding of an alarm and the resolution of a fault) when anesthesiologists used a conventional alarm system and when they used an intelligent alarm system. In a simulated operating room environment, we asked 20 anesthesiologists to resolve seven breathing circuit faults as quickly as possible. Human response time was 62% faster, decreasing from 45 to 17 s, when the intelligent alarm system was used. The standard deviations in response time were only half as large for the intelligent alarm system. It appears that the computer-based neural network in the intelligent alarm system diagnosed faults more rapidly and consistently than did the anesthesiologists. This study indicates that breathing circuit faults may be more rapidly corrected when the anesthesiologist is guided by intelligent alarms.

Anesthesiology

Failure of pulmonary acidosis to increase respiratory drive.

Experiments were performed to determine whether increases in acidity isolated to the pulmonary circulation would stimulate hypothesized pulmonary chemoreceptors and increase respiratory drive in the anesthetized paralyzed mechanically ventilated cat (n = 9). Respiratory drive was assessed by measuring the frequency and amplitude of the integrated phrenic neurogram. To create an isolated pulmonary acidosis, blood returning to the lung was acidified by infusion of 0.3 M lactic acid (1.91 ml/min) into the inferior vena cava, while systemic arterial pH was restored to near normal levels by simultaneous infusion of base (0.3 M NaOH) into the left atrium. Six minutes after the start of this dual infusion of acid and base, right ventricular (pulmonary) pH decreased from 7.286 to 7.179 and PCO2 increased 7 Torr. Systemic arterial pH and PCO2 were unchanged from measurements immediately before the infusion. This level of pulmonary acidosis failed to increase respiratory drive as assessed by phrenic activity. To test the sensitivity of the preparation to known systemic arterial chemical stimuli, a combined pulmonary and systemic acidosis was elicited by infusion of 0.3 M lactic acid into the inferior vena cava and 0.3 M NaCl into the left atrium. This infusion significantly lowered both systemic arterial and pulmonary arterial pH (7.343 to 7.155 for systemic arterial pH and 7.286 to 7.067 for pulmonary pH) and increased phrenic efferent activity 45%. We conclude that phrenic efferent activity is unaffected by moderate reductions in the pH of the pulmonary circulation in the absence of a significant systemic arterial acidosis.(ABSTRACT TRUNCATED AT 250 WORDS)

Acidosis, Respiratory

Differential features for a neural network based anesthesia alarm system.

We have developed a neural network based alarm system that identifies 19 specific faults in the anesthesia breathing circuit, such as "Inspiratory Hose Leak," or "Y-Piece Disconnection." CO2, pressure, and expired flow waveforms, along with ventilator settings, were sampled by a personal computer. Fifty-two features, such as "maximum CO2" or "minimum pressure", were extracted from each breath, converted to "differential" features, normalized, and used as the inputs of a three layered feed-forward neural network. The network was trained, using backward error propagation with momentum, to classify each breath as normal or containing one of 19 faults. To collect the neutral network training set, seven dogs were anesthetized and ventilated using controlled ventilation. Each of 19 faults were created over a range of ventilator settings and fresh gas flows. The neural network correctly identified 83.1% of 550 events presented to it during testing. These preliminary results are an encouraging example of neural network applications in the field of clinical monitoring.

Anesthesia, Closed-Circuit

Evaluation of a parallel implementation of the learning portion of the backward error propagation neural network: experiments in artifact identification.

Various methods have been proposed in an attempt to solve problems in artifact and/or alarm identification including expert systems, statistical signal processing techniques, and artificial neural networks (ANN). ANNs consist of a large number of simple processing units connected by weighted links. To develop truly robust ANNs, investigators are required to train their networks on huge training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from approximately 6.4 hours to 1.5 hours. We conclude that use of the master-worker model of parallel computation is an excellent method for obtaining speedups in the backward error propagation neural network training algorithm.

Algorithms

Fast neutron treatment for squamous cell carcinoma of the head and neck: final report of Edinburgh randomised trial.

OBJECTIVE: To compare neutron treatment and megavoltage (photon) radiotherapy in locally advanced squamous cell carcinoma of the head and neck. DESIGN: Randomised trial of patients stratified by site of primary tumour and presence or absence of lymph node metastases. Follow up of patients after treatment. SETTING: Department of clinical oncology, Western General Hospital, Edinburgh. PATIENTS: 165 Patients with untreated, histologically proved squamous cell carcinoma of the oral cavity, oropharynx, larynx, or hypopharynx. All patients completed treatment, and no patient was lost to follow up. INTERVENTION: Treatment with either neutrons or photons. MAIN OUTCOME MEASURES: Disease state and morbidity (scored with the system of the European Organisation for Research on Treatment of Cancer) at each visit during follow up. RESULTS: Of the 165 patients, 85 were randomised to receive neutron treatment and 80 to receive photon treatment. Minimum follow up was five years. Local control of cancer remained similar in the two groups, being achieved in 37 (44%) patients after neutron treatment and 36 (45%) after photon treatment. Five year and actuarial 10 year survival rates were 24% (20/85) and 14% respectively in the group treated with neutrons and 34% (27/80) and 30% respectively in the group treated with photons. Five year survival rates without local disease were 19% (16/85) and 30% (24/80) respectively. Necrosis was more common after neutron treatment than after photon treatment. Seven patients in the neutron group who developed necrosis died whereas no deaths were associated with photon treatment. CONCLUSION: Rates of long term local control were similar in the two groups. Necrosis related to radiation was more common in patients treated with neutrons, and the mortality related to treatment was significantly higher in these patients.

Carcinoma, Squamous Cell

Cardiopulmonary function in exercising bar-headed geese during normoxia and hypoxia.

To investigate possible physiologic mechanisms that allow the bar-headed goose to perform strenuous physical activity when flying at high altitude (e.g., above 9,000 m), we measured cardiopulmonary variables during running exercise (treadmill; 0.6 m.sec-1; 2 degrees incline) while the bird breathed either normoxic (21% O2) or hypoxic (7% O2) gases via a face mask. 1. During normoxic exercise, O2 uptake rate doubled and both ventilation and cardiac output increased. Blood gases and pH in arterial, mixed venous and blood from the leg, however, remained virtually unaltered. 2. Hypoxia at rest stimulated ventilation to rise but not cardiac output. The birds reached a steady state with virtually unaltered O2 uptake. 3. Exercise during hypoxia further stimulated ventilation, resulting in elevated arterial PO2 and O2 content compared to hypoxia at rest. However, O2 uptake increased only slightly, and cardiac output did not rise over the resting hypoxic value. The hyperventilation resulted in respiratory alkalosis and increased CO2 output, with R values being as high as 2.0. 4. It is concluded that neither ventilation nor pulmonary gas transfer were the limiting step in supplying O2 to the working muscles during hypoxic exercise in our experiments. It is more likely that muscle blood flow or diffusion from muscle capillaries to mitochondria, or both, determined the aerobic capacity under these conditions.

Animals

Nursing and the process of scientific inquiry.

Nurse education has to confront the problem of developing in students an understanding of the different perspectives put forward in nursing. However, all the perspectives adhere to the scientific method and it is therefore necessary to examine the process of scientific inquiry. As nursing occupies a position at the interface between physical and social sciences, it is also necessary to offer students an examination of the different paradigms which exist within the social sciences.

Nursing

Isocapnic hyperpnea in awake ponies during inspiration of 4% CO2.

Unanesthetized ponies were given 4% CO2 (inspired CO2 pressure = 28 Torr) to breathe at two levels of arterial O2 pressure (PaO2): 1) near 75 Torr and 2) near 200 Torr. During 4% CO2 breathing, at either level of PaO2, the mean arterial CO2 pressure (PaCO2) was unchanged from control measurements (control measurements were made at the same PaO2, but with no CO2 in inspired air), suggesting that awake ponies can "clear" 4% CO2. The ability of individual ponies to clear 4% CO2 was quite variable: some ponies did not clear 4% CO2 and others cleared 4% CO2 on one day but not on the following day. Based on the average of 20 experiments, however, PaCO2 was unchanged from 40 Torr during inspiration of 4% CO2. Direct measurement of chemical stimuli to breathing in arterial blood and cisternal cerebrospinal fluid indicate that ventilation increased during CO2 breathing even though PaCO2, pHa, PaO2, and CSF pH were not changed in a direction that might explain the accompanying change in ventilation. The authors suggest that stimuli to receptors other than peripheral or medullary chemoreceptors may be responsible for the reported "isocapnic hyperpnea."

Acid-Base Equilibrium

Hemodynamic effects of acute hypoxia and minute amounts of endotoxin in awake swine.

Hemodynamic measurements and arterial blood gases were determined from awake swine (n = 8) during acute hypoxia (12% O2, balance N2 for 5 minutes) and following injection of Escherichia coli endotoxin (4 microgram/kg of body weight) into the pulmonary artery. Comparison of baseline data with these treatments indicated: (1) swine exhibited a marked pulmonary pressor response to acute hypoxia (deltaBPpul = 9 to 12 mm of Hg): (2) injection of minute amounts of endotoxin led to a marked increase of pulmonary arterial blood pressure 10 to 20 minutes following the injection (deltaBPpul = 15 mm of Hg); and (3) the pulmonary pressor response to a 2nd exposure of acute hypoxia was unaffected by the intervening endotoxin injection.

Acute Disease

Salivary gland tumors: review of 643 cases.

643 cases of salivary gland tumors constitute two series of histological sections that were studied from hospitals and dental schools in Southeast Scotland and Southern Ontario. The Scottish series represented epithelial tumors of the parotid and intra-oral salivary glands, but the Canadian series also included tumors of the submandibular and sublingual glands. Classification was based on that recommended by the World Health Organisation (Thackray 1972). While direct statiscal comparisons between the two series are not appropriate, the differences between them suggest that malignant tumors are more common in Canada. The Scottish series contains the largest proportion of benign salivary tumors so far reported. In the Scottish series, 88.7% of parotid tumors were benign compared with 51.9% of Canadian series. In the Canadian series from the submandibular glands, 21.2% only were benign. Of the intra-oral salivary tumors, 62.2% from the Scottish series were benign compared with only 34.7% from the Canadian series.

Adenocarcinoma

Hypoventilation in ponies after carotid body denervation.

Seven ponies were subjected to carotid body denervation (CD) and two ponies were sham operated (S). Measurement of arterial blood gases and arterial blood and cerebrospinal fluid (CSF) acid-base balance were made prior to and 1,2,4,9, and 17 wks after surgery in unanesthetized animals. Resting ventilation and ventilatory responsiveness to hypoxia and NaCN infusion were assessed prior to and 2,9, and 17 wks after surgery. Alveolar hypoventilation in the CD ponies was marked 1-2 wk after surgery when VE and VA were reduced 40% and 10%, respectively, from control and PaCO2 was 12-15 mmHg above control. However, the effect was not nearly as great 4, 9, and 17 wk after surgery when the PaCO2 stabilized at approximately 6 mmHg above control PaCO2. Arterial blood pH was normal in the hypercapnic CD ponies, but CSF pH remained acid relative to normal throughout the 17-wk period. Changes in ventilatory responsiveness to hypoxia and NaCN tended to parallel changes in resting ventilation. These findings suggest: 1) the carotid bodies are essential in ponies to maintain normal ventilation: 2) in CD ponies peripheral chemosensitivity is partially regained at some unestablished locus; and 3) pH compensating mechanisms in chronically hypercapnic ponies function relatively better in blood than in CSF.

Acid-Base Equilibrium

Effect of exercise on cardiac output and distribution of uterine blood flow in pregnant ewes.

This study was designed to determine what effect physical training has on heart rate and stroke volume responses to exercise stress and to determine if exercise altered the distribution of uterine blood flow. Measurements were made in ten pregnant ewes at rest and immediately following exercise on a treadmill. Five ewes underwent physical training for 3 wk prior to measurement. An increase in heart rate with no change in stroke volume was observed following exercise in both trained and untrained ewes. Total uterine blood flow was not changed following exercise, but distribution was altered in favor of the placenta. Blood flow was evenly distributed within the placenta before and after exercise. The redistribution of flow to the placenta that occurs after exercise. tphe redistribution of flow to the placenta that occurs after exercise might represent a compensatory mechanism for the fetus.

Animals

Ventilatory control in peripheral chemoreceptor-denervated ponies during chronic hypoxemia.

The present study was designed to provide further insight into the role of the carotid and aortic chemoreceptors in ventilatory (VE) acclimatization during sojourn at altitude. Measurements were made: 1) on 10 ponies near sea level (SL, 740 Torr) under normal conditions, 2) on 6 of these at SL following chemoreceptor denervation (CD), and 3) subsequently on all 10 during 4 days of hypobaric hypoxia (PaO2 = 40-47 Torr). CD resulteo in hypoventilation at SL (deltaPaCO2 = d8 Torr, P less than 0.05), and it prevented hyperventilation normally observed with injection of NaCN and acute exposure to hypoxia (less than 1 h). In contrast, hyperventilation was evident in normal ponies during acute hypoxia (deltaPaCO2 = -6.7 Torr). Ventilation increased in both groups between the 2nd and 8th h of hypoxia (deltaPaCO2 from 1 h = -4 Torr, P less than 0.05). This change, a common characteristic of acclimatization, persisted throughout 4 days of hypoxia in the normal ponies. However, in the CD ponies this change was evident consistently only through the 12th h and after the 44 h hyperventilation was no longer evident. We conclude that the peripheral chemoreceptors are essential in ponies for normal VE acclimatization to this degree of hypoxemia. Two additional findings in CD ponies suggest the presence of a CNS inhibitory influence on the VE control center during chronic hypoxemia. First, acute hyperoxygenation on the 4th day of hypoxemia induced hyperventilation (deltaPaCO2 = -5 Torr, P less than 0.05). Second, again on the 4th day and during hyperoxygenation, VE responsiveness to CO2 and doxapram HCl was greater than at sea level.

Acclimatization

Cardiopulmonary effects of the combination of neuroleptic azaperone and hypnotic metomidate in swine.

Cardiopulmonary measurements were made at given intervals up to 120 minutes on 6 awake, unanesthetized pigs given azaperone and metomidate. Decreases from control values occurred in arterial blood pressure (deltaBPart = 30 mm of Hg), heart rate (deltaHR = 30 to 35 beats/minute), and cardiac index (deltaCI = 1.5 L/minute/m2). Blood gas and pH measurements indicated no severe impairment of pulmonary function or arterial acidosis. Although the drugs led to decreases in the various functions, cardiopulmonary function remained stable and uncompromised.

Animals

Cardiopulmonary effects of mechanical distension of the rumen in nonanesthetized sheep.

The cardiopulmonary effects of mechanical distension of the rumen were studied in nonanesthetized sheep. Results in group I ewes indicated that simply increasing intraruminal pressure by compressed air insufflation did not seriously affect cardiopulmonary hemodynamics. Changes were more serious in group II ewes when water injection was followed by ruminal insufflation with compressed air. Heart rate, total peripheral resistance, and total pulmonary resistance increased, with a concommitant increase in arterial and pulmonary blood pressure and decrease in stroke volume. Infusion of water alone increased heart rate and cardiac output. The blood gas data indicated a definite arterial hypoxemia and hypercapnia, suggesting impaired ventilation during ruminal distension. The increase in total pulmonary resistance was thought to be due to alveolar hypoxia caused by a significant decrease in ventilation.

Animals