Cerebrospinal fluid control system

Article Abstract:

An implantable cerebrospinal fluid control system (CFCS) is being developed for the improved shunt control of cerebrospinal fluid (CSF) buildup and pressure in skull of hydrocephalic patients. Major design goals of the system include: non-invasive opening and closing of the shunt to the abdomen, enabling the weaning from the need for a shunt, and improved understanding of the causes of CSF generation and absorption. The CFCS system includes: an implanted control module, consisting of the shunt valve, pressure transducers, command and telemetry systems, memory, microprocessor, and battery; an external patient control unit and modem; and a physician's PC-based control console, communications head, and modem. Both the patient's unit and physician's console can send commands to alter the flow of CSF. Details of the system design are described.

Noncommercial research organizations, Systems analysis, Industrial research, Medical research, Nervous system, Biomedical engineering, Neurophysiology, Systems development, Control systems, Research and Development, System Development, System Design, Applications, Physiological Monitoring

---

Maximum a posteriori decision and evaluation of class probabilities by Boltzmann Perceptron Classifiers

Article Abstract:

Neural networks can efficiently compute 'a posteriori' class probabilities for classifying stochastic patterns. A deterministic feedforward network called the Boltzmann Perceptron Classifier (BPC) can be more useful than Bayesian classifiers because it does not make assumptions about the probabilistic model of the problems. The BPC learns the stochastic properties of a problem by adjusting itself via a learning algorithm. Maximum a posteriori (MAP) classifiers are a special case of BPCs that focuses on the input pattern's most probable class rather than its exact class probabilities. BPC performance is comparable with Bayesian classifier performance, and its complexity is low and it requires no a priori assumptions about a problem's probabilistic model.

Neural networks, Comparative Study, Probability, Bayesian Theory, Classification Systems, technical, Neural Network, Perceptrons

---

Similar abstracts:

• Abstracts: ATC ground communications system optimization techniques. Development and operation of the traffic alert and collision avoidance system
• Abstracts: Center for Advanced Electronic Materials Processing. ULSI reliability through ultraclean processing
• Abstracts: Statistical control of VLSI fabrication processes. Modeling and simulation of VLSI digital systems. Analysis technology for VLSI fabrication
• Abstracts: Dependence of process parameters on planarization isolation and etching of sloped Vias in polyimides for GaAs ICs
• Abstracts: Antenna control and optimization. Spectral self-coherence restoral: a new approach to blind adaptive signal extraction using antenna arrays

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.