Catalyzing Inquiry at the Interface of Computing and Biology

174 CATALYZING INQUIRY

Box 5.16

Simulators for Computational Neuroscience

The nervous system is extraordinarily complex. A single cubic centimeter in the brain’s cerebral cortex con-

tains on the order of 5 billion synapses, and these differ in size and shape. The transmission of chemical

signals is very complex, with many molecules involved, and is an area of intense study. With the introduction

of more powerful computer hardware and advances in algorithms, quantitative modeling and realistic simu-

lation in three-dimensions of the interplay of biological ultrastructure and neuron physiology have become

possible and have provided insight into the variability in signaling and plasticity of the system.

To deal with the complexity, multiscale range of space and time, and nonlinearity of neural phenomena, a

number of specialized computational tools have been developed.

MCell (a Monte Carlo simulator of cellular microphysiology) simulates individual connections or synapses

between neurons and groups of synapses. MCell simulations provide insights into the behavior and variability

of real systems comprising finite numbers of molecules interacting in spatially complex environments. MCell

incorporates high-resolution physical structure into models of ligand diffusion and signaling and thus can take

into account the large complexity and diversity of neural tissue at the subcellular level. Monte Carlo algo-

rithms are used to simulate ligand diffusion using three-dimensional random walk movements for individual

molecules. Effector sites and surface positions are mapped spatially, and the encounters during ligand diffu-

sion are detected. Bulk solution rate constants are converted into Monte Carlo probabilities so that the diffus-

ing ligands can undergo stochastic chemical interactions with individual binding sites such as receptor pro-

teins, enzymes, and transporters.

GENESIS (the General Neural Simulation System) is a tool for building structurally realistic simulations of

biological neural systems that quantitatively embed what is known about the anatomical structure and phys-

Recent theoretical work has suggested that it is possible to generally characterize the dynamics,

representation, and computational properties of any neural population (Figure 5.13).^101 Applications of

these methods have been used successfully to generate models of working memory, rodent naviga-

tional tasks (path integration; see Figure 5.14), eye position control, representation of self-motion, lam-

prey and fish motor control, and deductive reasoning (Figure 5.15).

Box 5.17 illustrates the use of computational modeling to understand how dopamine functions in

the prefrontal cortex. The box also illustrates the often-present tension between those who believe that

simple models (in this case, advocates of a connectionist model) can provide useful insight and those

who believe that simple models cannot capture the implications of the complex dynamics of individual

neurons and their synapses and that the addition of considerable biophysical and physiological detail is

needed for real understanding. Many of these models require large numbers of individual, spiking

neurons to be simulated concurrently, which results in significant computational demands. In addition,

calculating the necessary connection weights requires the inversion of extremely large matrices. Thus,

high-performance computing resources are essential for expanding these simulations to include more

neural tissue, and hence more complex neural function.

(^101) C. Eliasmith and C.H. Anderson, Neural Engineering: Computation, Representation and Dynamics in Neurobiological Systems,
MIT Press, Cambridge, MA, 2003.