eunite

research

applications

knowledge transfer

e-archives



Hierarchical Intelligent Control of Unconsciousness

Submitted by D A Linkens with J S Shieh,29.11.2002,IBA D

Problem

The control of patient unconsciousness during surgical procedures is an extremely challenging task for anaesthetists. Not only is measurement a major problem, but also a rigorous definition of unconsciousness is very difficult. However, human anaesthetists are able to perform this task in routine conditions. The challenge is to produce a computer-based system which will regulate unconsciousness under normal conditions, with the anaesthetist being able to concentrate on the really life-threatening emergency situations which can arise. The system must be capable of utilising a range of inferential measurements (as surrogates for unconsciousness) and be adaptable for a wide range of patient conditions.

Solution

A multi-layer hierarchical architecture has been developed for either intravenous infusion (using the liquid drug propofol) or inhalational breathing (using a gaseous drug isoflurane).The system uses fuzzy merging of the various measurements which embody 3 types of information:1. brain-related signals via evoked potentials,2. vital signs, such as blood pressure and heart-rate monitoring,3. clinical signs, entered manually via anaesthetistís observations of the patient. A Self-Organising Fuzzy Logic Controller (SOFLC) of the Procyk and Mamdani type discovers necessary rules on-line, either from prior simulation or ab initio in the operating theatre. In particular, the scaling factors for the SOFLC are determined via some a priori knowledge of the patient sensitivity obtained during initial sedation prior to commencement of surgical operation.

Status and results

The estimation of inter-patient variability has been successfully determined for both propofol (intravenous) and isoflurane (inhalational) drugs. This is achieved by initial bolus administration of propofol or methohexitine in the pre-operation room and assessment of patient sensitivity into three bands of High, Medium and Low. This information is then used to adjust the parameters in the SOFLC. Successful operations have been achieved using this system, with the anaesthetist acting in a supervisory role. They evaluate the system under normal conditions, and take over control during emergency situations.

Adaptivity and portability

In terms of its multi-layer hierarchy of differing intelligent techniques, the system is hybrid in nature. In terms of adaptivity it achieves Level 1 in the EUNITE definition in that it adjusts to a changing environment. Thus, it uses a form of gain-scheduling to ensure fast stability of the closed-loop control for accommodating the large inter-patient variability of drug sensitivity.

More information

This is part of extensive PhD studies performed by J S Shieh, as described in his thesis:

J S Shieh"Hierarchical fuzzy logic monitoring and control in anaesthesia", PhD thesis, University of Sheffield,UK,1994

Some details are also available in:

J S Shieh,D A Linkens,and J E Peacock (1999),"Hierarchical rule-based and self-organising fuzzy logic control for depth of anaesthesia",IEEE Trans SMC,Part C,29,pp98-109.

technical reports & case studies

executive summaries

scientific papers

tutorials

bibliography

e-dictionary

search eunite with

google



  home