Complexitybased, personalized, online health monitoring methods and devices

1
Complexity-based, personalized, onlinehealth
monitoring methods and devices
Florin Munteanu, Center for Complexity Studies,
Romania
2
Complicated vs. complex

• A complicated system can be modeled and
  controlled its setup is predetermined its
  behavior is predictable it obeys clear causality
  and operation rules.

A complex system is not represented by a single,
stationary model it is plastic it evolves
within its environment it can self-organize its
internal structure it switches from one
operation mode to another depending on stresses
in the environment (phase transitions) it
shows a loose cause-effect connection.
3
Consequences

• Approaching and harnessing Complexity requires a
  fundamentally different approach
• Identifying patterns of behavior fingerprints
  of each state
• Identifying behavior discontinuities, clues for
  re-organization
• New approaches to data processing, which
  explicitly account for non-linearity, memory,
  scale-dependence, etc
• New experimental protocols

4
Proposal

• Monitor a complex system, and provide an early
  warning of misbehavior

5
Hypothesis

• There is a strong Mental/Affective coupling in
  the brain
• Through hormonal release, mental and affective
  processes influence physiology throughout the
  body!

6
Approach
Subject 1
Integration (e.g., PCA neural nets)
Clinical testing
SignalDatabase
Subject 2
Vectorization A representative set of signal
estimators
Subject n
Modeling Signal processing
7
Approach
Current state of patient is a point in an
abstract parameter space Various regions of this
space correspond to normal and abnormal
behavior
8
(No Transcript)