PubMed · 5104791
The source did not provide an abstract. Follow the original record for more information.
Explore connections, maps & timelines
M Kuwabara. 1971. [Non-linear vibration].. https://pubmed.ncbi.nlm.nih.gov/5104791/
Cite the original work for its findings. Save a collection to share your selection of sources.
KEEP EXPLORING
Respiratory disturbances in patients with the sleep apnea-hypopnea syndrome (SAHS) may be detected by means of nasal prongs (NP) pressure (PNP). Nevertheless, PNP is nonlinearly related to flow (V). Our aim was to demonstrate the relevance of linearizing P NP for assessing hypopneas and flow limitation in SAHS. V was measured with a pneumotachograph during the hypopneas and flow limitation events in a continuous positive airway pressure (CPAP) titration in six patients with severe SAHS. These flow patterns were reproduced by a flow generator through an analog of the nares and recorded by NP. PNP was linearized [V NP = (PNP)1/2] by a specially designed analog circuit. For each event we used V, P NP, and V NP to compute the hypopnea flow amplitude (HFA) and a flow limitation index (FLI). Owing to NP nonlinearity, PNP considerably misestimated HFA and FLI. By contrast, V NP provided HFA and FLI values that were very close to those obtained from V: HFA (V NP) = 1.098. HFA(V) - 0.063 (r2 = 0.98) and FLI(V NP) = 1.044. FLI(V) + 0.004 (r2 = 0.99). Square-root linearization of NP greatly increases the accuracy of quantifying hypopneas and flow limitation. This procedure, which could be readily carried out in routine practice by means of the analog circuit we developed, is of interest in optimizing the assessment of respiratory sleep disturbances in SAHS.
A belt worn controller and lithium-ion battery pack have been developed for use with the initial clinical trials of the Jarvik 2000 heart. Patient interface considerations, safety, and simplicity were major design inputs for the system. The controller was developed using all analog technology to avoid difficulties with electromagnetic interference (EMI), to minimize susceptibility to electrostatic discharge, and to avoid the need for software validation. Manual control of pump speed is accomplished by a patient operated knob, according to physician instructions for rest and exercise for each individual patient. The system includes alarms and indicators which show the following: the amount of remaining battery charge, if the battery is low and needs replacement, the power in watts being consumed, if the power consumed is above 15 W, if the pump is running below the selected speed setting, and if the pump stops. The control box, curved to be worn on the belt, is only 2.5 inches high for comfort when sitting. The battery pack, also form fitted for patient comfort, weighs just over 1 1/2 pounds and supplies 65 W-h of energy storage, sufficient to run the device for over 8 h at nominal load.
We show that neural networks with three-times continuously differentiable activation functions are capable of computing a certain family of n-bit boolean functions with two gates, whereas networks composed of binary threshold functions require at least omega(log n) gates. Thus, for a large class of activation functions, analog neural networks can be more powerful than discrete neural networks, even when computing Boolean functions.