Skip secondary menu

Computational and Theoretical Neuroscience: From Synapse to Circuitry

Table of Contentes:

Sponsored by the National Institute of Neurological Disorders and Stroke

The Neuroscience Building, Room A
6001 Executive Blvd., Rockville, MD
April 28, 2000

General Report
by Jason L. Pyle

I. Introduction

The National Institute of Neurological Disorders and Stroke (NINDS) recently conducted a workshop to explore applications of computational and theoretical neuroscience. This discipline is loosely defined as the application of rigorous mathematical and physical principles to the solution of neurobiological problems. Scientists from public and private institutions, representing many fields within neurobiology, assembled to discuss the use and future direction of quantitative techniques. A series of brief presentations highlighted the contributions that computational and theoretical approaches have made to our understanding of human neurological processes. The combination of rigorous quantitative science and experimental neurobiology was shown to have solved several previously intractable neuroscience problems. Despite the apparent advantage of increasing computational and theoretical research in neuroscience, economical and cultural restraints have restricted efforts to increase emphasis on computational and theoretical approaches in neurobiology. In order for the NINDS to further its goal of funding basic research pursuant to the improvement of public health, a change in funding structure, education, and institutional culture was recommended to support further computational and theoretical research.

II. Background

The explosion of experimental research techniques has dramatically increased our understanding of the nervous system and its function. The application of molecular biology, genetics, and cell biological techniques to the problems of neurobiology have drawn considerable and well deserved attention. Despite the insight these approaches afford, the business of the nervous system is to gather, process and transmit information; thus, neuroscience is an inherently computational discipline. A mature science of the brain ultimately requires a mathematically based understanding of the information and processing that occurs within the structures of the nervous system. To capitalize on several decades of rewarding experimental research, the NINDS, in partnership with the neuroscience community, has recognized the growing need for rigorous, quantitative research. This workshop marks a new effort by the NINDS to champion computational and theoretical neuroscience.

III. Discussion

Distinguished scientists from many fields within neurobiology opened the workshop with brief presentations highlighting current computational and theoretical efforts. Applications to several medically relevant neurological systems were presented and future implications to research, medicine and public health were discussed as follows. For example, a number of physiological processes, including motor control and sleep, are dependent upon neural circuitry that generates rhythmic oscillations. Rhythmic systems are particularly suited to exploration by quantitative techniques, as the study of physical sciences has already established a theoretical foundation. Investigations that model the delicate balance of oscillatory circuits are likely to increase our understanding of neuropharmaceutical and neurosurgical interventions. A complete quantitative understanding of motor control and sensory circuitry will ultimately lead to fully functional prosthetic devices for the handicapped.

Other types of brain circuitry are also amenable to quantitative analysis. Scientists, using a combination of experimental science with computational and theoretical modeling, have explored the visual system for several decades. Experimentation has determined that the retina of the eye transforms visual images into electrical information in the optic nerve, a one million-channel information conduit. Comprehension of such a system is simply impossible without the benefit of quantitative and computational analysis. While great progress has been made, workshop participants agreed that the full utility of computational and theoretical neuroscience has not yet been reached.

The origin of many neurological disorders is still unknown. Our current understanding of synaptic transmission, the process by which information is passed from one neuron to the next, is not sophisticated enough to resolve how minor alterations might result in major diseases, such as movement disorders, mood disorders or schizophrenia. While great strides have been made in elucidating the molecular physiology of synapse function, how information is encoded by the electrical firing of neurons in the brain is still poorly understood. Quantitative analysis of neural systems, from the level of the synapse to the behavior of the awake animal, is likely to be the only means of deciphering the neural code. The goal of NINDS, to understand normal and diseased neurological function, depends on elucidating the principles of information processing that occur in biological neural networks. This is fundamentally a computational and theoretical endeavor.

Two important aspects of current scientific culture inhibit expanding computational and theoretical work in neuroscience: education and career opportunity. Two separate issues concerning education were raised. Scientists who have "crossed over" to biology from the physical, engineering and mathematical sciences have traditionally performed a considerable amount of the quantitative work done in biological sciences. Thus, continued support and development of programs that attract such students to neuroscience is of obvious benefit. Workshop participants suggested that a more important goal might be to train biological scientists in quantitative techniques. The current collegiate level biological science education, aimed primarily at pre-medical students, does not have the adequate computational and theoretical emphasis needed to develop quantitatively minded neuroscientists. Institutions, such as the Sloan Centers, where funding and organizational emphasis has been placed on integrating computational and theoretical studies into biological curricula have been largely successful and rewarding. The experience of several scientists who had developed programs where computational and theoretical studies where integrated into traditional biological studies highlighted career opportunity as the second major impediment to improved computational and theoretical research. Students in quantitative fields such as engineering and computer science are largely discouraged from entering the biological sciences due to the extremely low postdoctoral compensation. Furthermore, computational and theoretical scientists find few available faculty positions in biological departments. Without adequate compensation and career advancement opportunities, neuroscience will continue to lack the students and faculty needed to support a true computational and theoretical research endeavor.

IV. Recommendations/Conclusions

The workshop distilled its discussion into several recommendations. There was clear agreement concerning the need to expand funding opportunities for computational and theoretical projects. This recommendation was tempered with the acknowledgement that additional funding with out additional direction will not be maximally effective. The situations deemed most likely to be immediately profitable were collaborations that contain an interface between experimental and theoretical scientists. Moreover, expanding the scope of current programs to encourage theoretical work on issues relating to disease is particularly suited to NINDS goals and the benefit of the public at large. Further recommendations on funding suggested that the current review process for grants be modified so that the merit of computational and theoretical project proposals can be more accurately considered.

Participants agreed that emphasis needs to be placed on expanding and developing programs designed to alter the current system of biological science education. Larger awareness and appreciation of computational and theoretical research by the general biological scientist is guaranteed to increased collaborations and "crossover" between the biological and the physical sciences. While funding and education would provide the necessary short-term boost, an institutional change was deemed necessary for a true base of quantitative research to flourish in neuroscience. Thus, scientists recommended that NINDS work with educational institutions to establish faculty career opportunities within biological departments for computational and theoretical scientists.

In summary, NINDS support of basic discovery applicable to human neurology is best supported by a two-tiered approach. The first, a change in the current funding structure, allows a judicious increase in computational and theoretical research largely directed at the interface of experimental and theoretical science. The second, partnerships with current institutions aimed at changing the culture of biological research, establishes a lasting base of computational and theoretical research within the neuroscience community.

V. Participants

Richard W. Tsien, D.Phil (Co-Chair)
Department of Molecular and Cellular Physiology
Stanford University School of Medicine

Eve E. Marder, PhD (Co-chair)
Volen Center
Brandeis University

Ruzena Bajcsy, PhD
Directorate for Computer and Information Science and Engineering (CISE)
National Science Foundation

Tom M. Bartol, Jr., PhD
Computational Neurobiology Laboratory
The Salk Institute

William Bialek, PhD
NEC Research Institute

John P. Donoghue, PhD
Department of Neuroscience
Brown University

Gwen A. Jacobs, PhD
Center Computational Biology
Montana State University

Stephen G. Lisberger, PhD
Department of Physiology
University of California School of Medicine

Wade Regehr, PhD
Department of Neurobiology
Harvard University School of Medicine

John M. Rinzel, PhD
Center for Neural Science
New York University

H. Sebastian Seung, PhD
Department of Brain and Cognitive Science
Massachusetts Institute of Technology

Karen Ann Sigvardt, PhD
Center for Neuroscience
University California at Davis

David W. Tank, PhD
Bell Laboratories
Biological Computation Research
Lucent Tech

Charles Wilson, Ph.D.
Cajal Neuroscience Research
Division of Life Science
University Texas at San Antonio

Last updated April 15, 2011