header-photo

Claudius Gros

The role of intrinsic plasticity for multistable behavior


The brain disposes of several regulative mechanisms in order  to stabilize distinct working regimes. On the highest level  dynamical regimes suitable for distinct cognitive tasks are
selected via diffusive control, mediated by neuromodulators  such as Dopamine, Seretonin etc (Gros, 2010). On a local level  two distinct adaption processes are known, synaptic plasticity   describing the effectiveness of inter-neural information exchange  and intrinsic plasticity regulating the response of individual neurons upon stimulation. Both adaption processes may work together in  creating two possible types of multi-stabilities, either in terms of  dynamical regime (chaos, intermittent bursting, synchronized firing, ..)  selection or as transient state dynamics (Gros, 2007/2009) within a given  dynamical state. Here we discuss the role of intrinsic plasticity for   stabilizing multistable dynamics.

It is reasonable to assume that a neuron may try to adapt its intrinsic  parameters, like threshold and gain, in order to achieve a firing-rate  statistics maximizing information content in terms of Shannon's information  entropy. Minimization of this objective function leads to slow adaption of  the neurons intrinsic parameters which lead to several types of 
multi-stabilities for the fast dynamics, the neural firing rate. For  networks of rate encoding neurons we show that chaotic, intermittent  bursting and synchronized regimes may be stabilized (Markovic & Gros,  2010/2012). We furthermore show that intrinsic adaption may transiently  stabilize dynamical states corresponding to neural memories in terms of   attractor ruins, with the overall dynamics corresponding to continuous  latching transitions from one memory to the next.

M. Linkerhand, C. Gros,
Self-organized stochastic tipping in slow-fast dynamical systems.
Mathematics and Mechanics of Complex Systems 1, in press (2013).

D. Markovic, C. Gros,
Intrinsic adaptation in autonomous recurrent neural networks.
Neural Computation 24, 523 (2012).

D. Markovic, C. Gros,
Self-organized chaos through polyhomeostatic optimization.
Physical Review Letters 105, 068702 (2010).

C. Gros, 
Cognition and Emotion: Perspectives of a Closing Gap.
Cognitive Computation 2, 78 (2010).

C. Gros,
Cognitive computation with autonomously active neural networks:
an emerging field.
Cognitive Computation 1, 77 (2009).

C. Gros,
Neural networks with transient state dynamics.
New Journal of Physics 9, 109 (2007).