Dynamic systems approaches and levels of analysis in the nervous system

David Parker,Vipin Srivastava

doi:10.3389/fphys.2013.00015

David Parker, Vipin Srivastava

Open Access

https://doi.org/10.3389/fphys.2013.00015

Copy DOI

Abstract

Various analyses are applied to physiological signals. While epistemological diversity is necessary to address effects at different levels, there is often a sense of competition between analyses rather than integration. This is evidenced by the differences in the criteria needed to claim understanding in different approaches. In the nervous system, neuronal analyses that attempt to explain network outputs in cellular and synaptic terms are rightly criticized as being insufficient to explain global effects, emergent or otherwise, while higher-level statistical and mathematical analyses can provide quantitative descriptions of outputs but can only hypothesize on their underlying mechanisms. The major gap in neuroscience is arguably our inability to translate what should be seen as complementary effects between levels. We thus ultimately need approaches that allow us to bridge between different spatial and temporal levels. Analytical approaches derived from critical phenomena in the physical sciences are increasingly being applied to physiological systems, including the nervous system, and claim to provide novel insight into physiological mechanisms and opportunities for their control. Analyses of criticality have suggested several important insights that should be considered in cellular analyses. However, there is a mismatch between lower-level neurophysiological approaches and statistical phenomenological analyses that assume that lower-level effects can be abstracted away, which means that these effects are unknown or inaccessible to experimentalists. As a result experimental designs often generate data that is insufficient for analyses of criticality. This review considers the relevance of insights from analyses of criticality to neuronal network analyses, and highlights that to move the analyses forward and close the gap between the theoretical and neurobiological levels, it is necessary to consider that effects at each level are complementary rather than in competition.

Highlights

Life is said to occur at the border between order and chaos (Macklem, 2008): it requires stability, traditionally expressed in terms of homeostatic principles, with flexibility and adaptability at the micro and macro levels. Schrodinger (1944) defined life as the passage of encoded material from parent to offspring, and the spontaneous emergence of self-organized order
Significant mechanistic insight has been obtained by reducing systems to their components. While this approach has driven some major advances in genetics, developmental biology, and cellular physiology, it is unlikely to be successful at the systems physiology level where outputs depend on non-linear interactions between components parts
The highlighting of variability in physiological systems; that correlations over different time scales may reflect the degree of complexity, control, and adaptability of a system; the move away from the caricature of homeostasis as equal to negative feedback driven clamping of a set point; the implications of fractal effects to the mean and variance in data sets; and the loss of complexity not regularity as a marker in several disease conditions are all important features derived from dynamical systems approaches that should inform experimental analyses of neuronal networks

Summary

INTRODUCTION

Life is said to occur at the border between order and chaos (Macklem, 2008): it requires stability, traditionally expressed in terms of homeostatic principles, with flexibility and adaptability at the micro (molecular and cellular) and macro levels (networks, organisms). Schrodinger (1944) defined life as the passage of encoded material from parent to offspring, and the spontaneous emergence of self-organized order. Schrodinger (1944) defined life as the passage of encoded material from parent to offspring, and the spontaneous emergence of self-organized order The former aspect is being dealt with quite successfully by biochemistry and molecular biology; the latter aspect is for physiology to address and it remains open. Significant mechanistic insight has been obtained by reducing systems to their components (i.e., assuming proportionality and superposition, that systems are sums of their parts) While this approach has driven some major advances in genetics, developmental biology, and cellular physiology (obvious examples in the nervous system are the Hodgkin–Huxley analysis of the action potential and Katz’s quantal model of synaptic transmission), it is unlikely to be successful at the systems physiology level where outputs depend on non-linear interactions between components parts.

Parker and Srivastava

Findings

CONCLUSIONS