Research Outline

The development of technologies to individually predict and prevent the onset of disease is a pressing issue in achieving sustainable healthcare.

In medicine, the symptoms of disease have been explained and controlled by causal mechanisms. However, this approach cannot explain or predict the complex process from health to disease onset. With the development of AI technology, surrogate models that can make highly accurate predictions from large amounts of real-world data have also begun to be developed in medicine. Because surrogate models have a “black box” structure that makes it impossible to know their operating principles, there is a problem that the bias of the data used to construct the model can affect the reproducibility and reliability of the predictions. Surrogate models have not yet been fully accepted in the medical field, where risks are unacceptable.

I have developed a new theory that will be the foundation of life science to realize personalized medicine based on prediction. This theory introduces the concept of force fields used in Newtonian mechanics, electromagnetism, relativity, quantum mechanics, and thermodynamics to biological phenomena. While physical phenomena occur in a non-equilibrium state close to equilibrium, biological phenomena occur in a non-equilibrium state far from equilibrium. The theory I have developed has made it possible to develop a variational principle and governing equations specific to life. I have also developed a theory that integrates different accounts of temporal change, such as causal models, surrogate models, and force field models, using category theory. These two theories form the basis of the Patient Digital Twin.

Reference

Sakurada K, Ishikawa T. Synthesis of causal and surrogate models by non-equilibrium thermodynamics in biological systems. Sci Rep. 10;14(1):1001 (2024)