Page 220 - Artificial Intelligence in the Age of Neural Networks and Brain Computing

4. Multistability in Physics and Biology 211

The concept of metastable patterns and intermittent transients has been extended
to neurodynamics as well [40,41]. Note that in the ongoing discussions we do
not claim that perceptual switching with Necker cube and switches observed in
neurodynamics are the same phenomena. Rather we point out that switches may
occur in different form and at different time sales in cognition and neural processes.
Freeman’s neurodynamics describes the brain state through a trajectory across a
high-dimensional space. The trajectory evolves through a sequence of metastable
states [41,42]. These metastable states may be viewed as intermittent symbolic
representations of the stimuli in the context of the individual past experiences and
future desires. However, these metastable symbols are transient and disintegrate
soon after they emerge [33,42]. The corresponding dynamic view has been crystal-
lized in the concept of the cinematic theory of cognition [32,34]. According to
the cinematic theory, cognition is not a smooth, continuous process in time; rather
it is a sequence of metastable cognitive states, which can be viewed as movie frames.
Such frames exist for about 100e200 ms, and then they briefly collapse.
The collapse of the states signifies the shutter and it takes about 10e20 ms. The
quasi-periodic sequence of the frames (metastable patterns) is the movie of the brain
as cognitive processing evolves in time [33].
Fig. 10.4 illustrates experimental findings on metastability in brain dynamics us-
ing electrocorticograms (ECoGs) from an array of 8 8 electrodes in rabbit sensory
cortex [33]. The top plot with the ECoG signals shows beating patterns of relatively
high synchrony for about 150e200 ms, interrupted by brief desynchronization
periods marked by blue bars. The bottom plots in Fig. 10.4 show the complementary
aspects of microscopic and macroscopic neural processes in rabbit brain, interacting
through phase cones emerging at the mesoscopic scales.

4. MULTISTABILITY IN PHYSICS AND BIOLOGY
Metastable behaviors introduced previously in the context of cognition and brains
have their counterparts in physical and biological processes. Physical systems
consisting of many parts may exhibit various dynamical regimes depending
on the dynamics of their parts and the nature of the interactions among the compo-
nents. If the interaction among the parts is strong, the overall behavior may become
a synchronized regime. Synchronization can be either amplitude or phase
synchrony [43]. In the case of amplitude synchrony, the amplitudes of the various
components are the same across the system, which is a strong case of synchrony. In
some other conditions, the amplitudes of the individual components may differ, but
they are in the same phase, that is, they wake and wane simultaneously. This more
relaxed occurrence of synchrony is called phase synchrony. An even more complex
behavior may occur when different parts of the system show amplitude or phase
synchrony for some time, but the synchrony diminishes for other periods, at least
in parts of the system. This is the case of multistability, when the system is inter-
mittently stable (metastable) for some time and space and it switches to another