SYLLABUS
NSC
4368 Spring 2007
COMPUTATIONAL NEUROSCIENCE
Room GR4.708
Thursday evenings 7:00 – 9:45PM
Instructor: Larry Cauller, Ph.D., Associate Professor of Neuroscience in the School of Behavioral and Brain Sciences
Office hours: JO4.214 on
Tuesdays and Thursdays 1:00PM or by appointment (via e-mail).
e-mail: lcauller@utdallas.edu
website: www.utdallas.edu/~lcauller
In this course, you will learn how to use
state-of-the-art computer software to construct realistic neural models and
simulate the dynamic behavior of realistic neurons and networks of neurons
connected with realistic synapses.
Computer simulation provides an unprecedented method to study the complex interplay of the biophysical mechanisms that govern the biologically realistic behavior of neurons and synaptic neural networks. The tremendous growth of computational neuroscience over the past three decades has paralleled the development of powerful neuron-simulation software. Two sophisticated neuron-simulation environments have been developed with major funding from the NSF and ONR, and have now been released for public distribution. By providing a standardized forum for controlled experimentation and critical scientific analysis, this state-of-the-art software has unified the computational neuroscience community and spurred the further growth of active research employing these advanced methods for neuro-simulation.
These new tools have been developed specifically to make these powerful
neuro-simulation methods available to many more
students, educators and researchers who lack the specialized training and
advanced computer programming skills previously needed to perform the
sophisticated neuro-simulaiton operations embodied in
these new tools. These new tools provide a friendly graphical user interface
(GUI) environment that enables all neuroscientists, novice and expert alike, to
quickly build 3-D computer models of neurons and load them with neural
mechanisms from an extensive built-in library by means of simple
point-drag-and-click operations. Unlike the simulators of the past, when users
had to wade through the obscure codes of text-based programs, with the advent
of friendly new GUI environments neuro-simulation
research today is more like drawing objects with popular applications such as
Illustrator or mathCAD, and playing video games
(especially if you’re a neuro-nerd).
Students will gain the essential skills to use these new tools by
working hands-on with the software in the classroom to complete a series of
increasingly sophisticated Exercises. Students will learn how to use the
software in class where they can directly observe the instructor step through
each Exercise on a large-screen computer monitor. The instructor will introduce
each new Exercise by openly demonstrating new operations to the class, and will
follow-up after students have completed the Exercise, by personally completing
the Exercise in class to discuss the significance of the results and personally
address any specific problems students may have encountered. This incremental
training method, with personal supervision in the classroom, is specifically
intended to maximize the opportunity for every student to acquire the essential
skills to take full advantage of these sophisticated neuro-simulation
tools.
Students will receive advanced instruction on topics spanning
neuroscience, from cellular to systems, specifically related to the preparation
and interpretation of the computational models presented as part of the course.
The course will work on models specifically chosen to demonstrate how students
may employ these new tools to gain a deeper understanding of specific neural
mechanisms (e.g. voltage-sensitive channels or synaptic plasticity) with
respect to the roles such mechanisms may play in neural behavior and the
functions it serves. Computational neuroscience research recognizes the
essential complexity of biological systems and embraces the intractable
complexity generated by simulations constructed with realistic mechanisms.
Students will become familiar with many of these realistic mechanisms including
the non-linear kinetics and multi-factor dependence of ionic channels, spatial
and temporal summation of synaptic and neuromodulatory
inputs, synaptic plasticity and neural interactivity, oscillations, pattern
generation and dynamical network behavior. Neuro-simulation
makes it possible to access every variable at every moment and control every
parameter to examine in detail the role of each mechanism in the overall
behavior of the system. Students will discover they may use their simulations
to realize their understanding and express their explanation of how specific
mechanisms influence neural behavior.
It is important for students to recognize that the methods of
computational neuroscience are fundamentally different from the methods of
‘artificial intelligence’ or artificial ‘neural networks’. Such artificial
systems are founded upon extreme mathematical simplification and are based upon
top-down designs that implement theoretical models. While the simplification of
artificial systems has facilitated progress through the era of relatively slow
computer power, the limitations of such simplified systems are now widely
acknowledged. In contrast, ‘biological
realism’ is the foremost concern of contemporary research in computational
neuroscience which emphasizes reconstruction of realistic neural structures
assembled from the bottom-up out of realistic neural elements. The new neuro-simulation software is poised to leap ahead with the
accelerating pace of computational power which is rapidly approaching the point
where the full potential of computational neuroscience methods will produce
significant breakthroughs, both with the realistic simulation of increasingly
large-scale models of complex neural behavior, and with the identification of
the neural mechanisms whose importance for higher functions is otherwise lost
in the process of artificial simplification.
This course will culminate in Final Student Projects that will give
students the opportunity to demonstrate their ability to successfully employ
these new tools. Students will construct neural models and run simulations
related to specific neural functions chosen from suggestions offered in class
(e.g. listed below) or developed by students with the help of the instructor to
address particular student interests.
This final exercise is primarily intended to strengthen student
confidence in the use of these new tools and kindle their enthusiasm to employ neuro-simulation methods as part of their independent
exploration into the neural basis of the higher functions (e.g. perception or
learning) or dysfunctions (e.g. Alzheimers or Parkinsons Disease) that engage their personal interests
and motivate their long-term effort to become professional neuroscientists,
researchers and physicians.
Specifically, this course will train students how to use the NEURON
simulation environment developed at Duke and Yale Universities (east coast), one
of the two popular neural simulation applications (versus GENESIS from Caltech,
west coast), both of which are freely available over the internet (see NEURON
link below). The authors of NEURON have played a central role in the growth of
neural simulation methods for research. The National Science Foundation has
funded the development of NEURON specifically as a teaching tool for medical
and neuroscience educators, and we will take advantage of the NEURON lessons
available from their website. NEURON provides a great range of tools that are
simultaneously simple enough for users to quickly learn and make significant
progress, and yet powerful enough to meet the demands of even the most advanced
researchers (i.e. customized mechanisms and simulation objects may be written
in MODL and compiled into the NEURON environment). NEURON is rather fun to play with, but it is
no toy... it is a sophisticated, state-of-the-art neuro-simulation
environment used worldwide.
While some familiarity with computers is required for this course, it
is not necessary for students to be computer programmers or to have technical
knowledge of programming languages or computer hardware. Students that have
taken the Neurophysiology core course will be better prepared to grasp the significance
of the Exercises and tackle more sophisticated Projects. However, a basic
understanding of action potentials (i.e. membrane potential and
voltage-dependent sodium and potassium channels) is sufficient to complete the
Exercises, learn how to use NEURON and successfully complete the course.
Ø Grading will be based upon weekly participation (10%), completion of NEURON Exercises (40%),
Ø Plus a final Student Project (50%) consisting of a working NEURON simulation related to some neural function such as:
· the role of lateral inhibition for contrast enhancement as in the ‘mach band’ illusion;
· central pattern generation such as that responsible for walking or other complex movements;
· cellular mechanisms of spontaneous patterned or bursting behavior;
· networks that switch their behavior such as sleep-wakefulness, or resting-seizure;
· any simulation project that addresses some neural function related to a student’s particular interests
· (additional suggestions and examples from previous courses will be presented).
Ø The final simulation Project may be completed by a team of up to 3 students with a corresponding increase in expected sophistication.
Ø The final simulation Project must be orally presented to the whole class and follow-up questions must be conscientiously answered (for teams – each student must make an individual contribution to the presentation and follow-up answers).
Ø The final simulation Project must be submitted as a final paper that includes
1. The rationale for the simulation
2. Cited references to any external sources of information used to design the model
3. Printed copies of screen shots with captions demonstrating
a. the model structure and
b. the results of simulation
4. A brief discussion of
a. the simulation results and
b. conclusions about the significance of the simulation.
This course will employ the latest version of the Neuron
simulation software which may be downloaded for free from…
http://www.neuron.yale.edu/neuron/install/install.html
Previous versions should be uninstalled before launching this installer. After
installation, the banner version should be
10/Nov/2006 Version 5.9 release 10
Students are encouraged to install NEURON on their personal computers.
All students will be
given sufficient opportunity to complete all course requirements during class
time and/or by appointment using the computer systems available in the
Neuroscience Teaching Lab classroom (GR4.708).
CLASS SCHEDULE:
January 11
What
is Neural Simulation and What is it Good For?
The
Difference Between Simulation and Mathematical Analysis
Fundamentals
of Neural Simulation: Numerical Analysis
Fundamentals
of Computer Simulation: Step-wise Accumulation
NEURON:
One of the two leading neural simulation environments.
January 18
NEURON
DEMO
NEURON
EXERCISE 1: Cable Properties
January 25
NEURON
EXERCISE 1: Cable Properties
NEURON
EXERCISE 2: Action Potential Generation
February 1
NEURON
EXERCISE 2: Action Potential Generation
NEURON
EXERCISE 3: Axonal Conduction (Introduction)
February 8
NEURON
EXERCISE 3: Axonal Conduction (Review)
NEURON
EXERCISE 4: Synaptic Integration
February 15
NEURON
EXERCISE 4: Synaptic Integration
Exploring
Parameter Space: What Computational Neuroscientists do most.
NEURON EXERCISE 5:
Exploring the Parameter Space of the Action Potential
February 22
NEURON EXERCISE 5: Exploring
the Parameter Space of the Action Potential
NEURON
EXERCISE 6: Synaptic Interaction Between Two Neurons
March 1
NEURON
EXERCISE 6: Synaptic Interaction Between Two Neurons
PREPARATIONS for the FINAL
STUDENT PROJECT:
·
Up to 3 students may work in a team on a
common Project.
·
Students may choose from suggested examples
or develop their own.
March 8 – SPRING BREAK
March 15
PREPARATIONS for the FINAL
STUDENT PROJECT:
·
All teams must be identified by this date
·
Chosen student Project topics reviewed and
related simulation methods openly discussed in class.
NEURON EXERCISE 7: Dynamic
Behavior of Small Neural Networks
March 22
NEURON EXERCISE 7: Dynamic
Behavior of Small Neural Networks
Specific topics for every Student Project must be finalized by this date
Example Student Project
March 29
Guided Progress on Student
Projects
Advanced instruction to help
with specific topics chosen for Student Projects
Guided Progress on Student
Projects
Examples
of Published Simulation Studies (with relevance to Student Projects if found)
April 12
Guided Progress on Student
Projects
First Presentations of Student Projects
April 26 – FINALS WEEK
Remaining Presentations of Student Projects
All Student Project Papers Due (no
Final exam)
Larry Cauller, Ph.D.
lcauller@utdallas.edu
(972)883-2436 FAX (972)883-2491
Back
to Larry's Home Page