Model-Based Biosignal Interpretation

Abstract:

Two relatively new approaches to model-based biosignal interpretation, qualitative simulation and modelling by causal probabilistic networks, are compared to modelling by differential equations. A major problem in applying a model to an individual patient is the estimation of the parameters. The available observations are unlikely to allow a proper estimation of the parameters, and even if they do, the task appears to have exponential computational complexity if the model is non-linear. Causal probabilistic networks have both differential equation models and qualitative simulation as special cases, and they can provide both Bayesian and maximum-likelihood parameter estimates, in most cases in much less than exponential time. In addition, they can calculate the probabilities required for a decision-theoretical approach to medical decision support. The practical applicability of causal probabilistic networks to real medical problems is illustrated by a model of glucose metabolism which is used to adjust insulin therapy in type I diabetic patients.

Keywords

Model-based Biosignal Interpretation - Parameter Estimation - Causal Probabilistic Networks - Simulation - Physiological Systems

• REFERENCES

• 1 Laursen P. Event detection on patient monitoring data using causal probabilistic networks. In: van’Bemmel JH, Rosenfalck A, Sarannummi N. eds. Biosignal Interpretation. Meth Inform Med (Special issue). 1994. 32 111-5.
• 2 Koski EMJ, Makivirta A, Sukavaara T, Kari A. Frequency and reliability of alarms in the monitoring of cardiac postoperative patients. Int J Clin Monitoring Comput 1990; 07: 129-33.
• 3 Cooper JB, Newbower RS, Kitz RJ. An analysis of major errors and equipment failures in anaesthesia management: Considerations for prevention and detection. Anesthesiol 1984; 60: 34-42.
• 4 Hedlund A, Zaar B, Groth T, Arturson G. Computer simulation of fluid resuscitation in trauma. Description of an extensive pathophysiological model and its first validation. Comput Meth Progr Biomed 1988; 27: 7-21.
• 5 Eykhoff P. System Identification. New York: Wiley; 1974
• 6 Kuipers B. Qualitative Simulation. Artif Intell 1986; 29: 289-338.
• 7 Forbus KD. Qualitative process theory. Artif Intell 1984; 24: 85-168.
• 8 de Kleer J, Brown JS. A qualitative physics based on confluences. Artif Intell 1984; 24: 7-83.
• 9 Irani L, Stefanelli M, Lanzola G. Qualitative models in medical diagnosis. In: Keravnou E. ed. Deep Models for Knowledge Engineering, Medical Artificial Intelligence. (vol. I). Amsterdam: Elsevier Science Publ; 1992: 51-70.
• 10 Coiera E. Monitoring diseases with empirical and model generated histories. In: Keravnou E. ed. Deep Models for Knowledge Engineering, Medical Artificial Intelli- ‘ gence (vol. I). Amsterdam: Elsevier Science Publ; 1992: 71-88.
• 11 Uckun S, Dawant B. Qualitative modelling as a paradigm for diagnosis, prediction and therapy management in critical care environments. Artif Intell Med 1992; 04: 127-44.
• 12 Pearl J. Fusion, propagation and structuring in belief networks. Artif Intell 1986; 29: 241-88.
• 13 Lauritzen SL, Spiegelhalter DJ. Local computations with probabilities on graphical structures and their application to expert systems (with discussion). J Roy Statist Soc B 1988; 50: 157-224.
• 14 Andreassen S, Falck B, Olesen KG. Diagnostic function of the expert EMG assistant -MUNIN. Electroenceph Clin Neurophysiol 1992; 85: 143-57.
• 15 Andreassen S, Hovorka R, Benn JJ, Olesen KG, Carson ER. A model based approach to insulin adjustment. In: Stefanelli M, Hasman A, Fieschi M, Talmon J. eds. Lecture Notes in Medical Informatics. (vol 44). Berlin: Springer Verlag; 1991: 239-419.
• 16 Andreassen S, Benn JJ, Hovorka R, Olesen KG, Carson ER. A probabilistic approach to glucose prediction and insulin dose adjustment – Description of a metabolic model and pilot evaluation study. Comput Meth Prog Biomed. 1994 (in press).
• 17 Dechter R, Dechter A, Pearl J. Optimization in constraint networks. In: Oliver RM, Smith JQ. eds. Influence Diagrams, Belief Nets and Decision Analysis. New York NY: Wiley; 1990: 411-26.
• 18 Kumar V. Algorithms for constraint-satisfaction problems – A survey. AI magazine 1992; 13: 32-44.
• 19 Andersen SK, Olesen KG, Jensen FV, Jensen F. HUGIN – A shell for building Bayesian belief universes for expert systems. Proceedings IJCAI89 1989; 02: 1080-5.
• 20 Groth T, Hakman M, Hedlund A, Zaar B. KBSIM/FLUIDTHERAPY: a system for optimized design of fluid resuscitation in trauma. Comput Meth Progr Biomed 1991; 34: 163-73.
• 21 Andreassen S. Planning of therapy and tests in causal probabilistic networks. Artif Intell Med 1992; 04: 227-41.
• 22 Lauritzen SL. Propagation of probabilities, means and variances in mixed graphical association models. J Amer Statist Assoc 1992; 87: 1098-1108.
• 23 Olesen KG. Causal probabilistic networks with both discrete and continuous variables. IEEE Trans Pattern Anal Machine Intell 1993; 15: 275-9.