CN570 class wiki is located here
CN570: NEURAL AND COMPUTATIONAL MODELS OF CONDITIONING, REINFORCEMENT, MOTIVATION, AND RHYTHM
Spring 2009
Instructor: Can Ozan Tan, Ph.D. Email: tanc [AT] cns [DOT] bu [DOT] edu Office Phone: 617-573-2723
Office: Rm. 113, 677 Beacon Street Office hours: By appointment only. Send an e-mail with two suggested alternative times.
Teaching Fellow: Jesse Palma Email: jpalma [AT] cns [DOT] bu [DOT] edu Office Phone: 617-353-6426
Office: Rm. 110, 677 Beacon Street Office hours: By appointment.
Class meeting time: Wednesdays at 5-8 pm.
Last Day to ADD Classes: Wednesday, January 28, 2009
Last Day to DROP Classes (without a 'W' grade): Thursday, February 19, 2009
Last Day to DROP Classes (with a 'W' grade): Monday, March 30, 2009
Purpose: The goal of this course is to survey traditional and contemporary psychological and behavioral approaches to the study of conditioning and reinforcement, and to integrate behavioral phenomena with the current knowledge of the physiological, neurobiological and computational bases of these phenomena. The primary focus of the lectures is to review experimental literature and to introduce the most influential theoretical and computational theories in the field. In particular, the course will focus on neural and computational models of how humans and animals learn to successfully predict environmental events and generate behavioral actions that satisfy internally defined criteria of success or failure. Classical and instrumental learning and related homeostatic (drive, arousal, rhythm) and non-homeostatic (reinforcement) modulators will be analyzed in depth to survey how several processes cooperate to generate successful goal-oriented behavior. Maladaptive behaviors and certain mental disorders will be analyzed from physiological, theoretical, and computational perspectives. Applications of these concepts to the solution of basic technological problems will also be noted.
Grading and assignments: Course grades will be based on the following:
- (25%) Attentive participation in class sessions, and submission of 8 ~1 page (single-spaced, 12 pt font) critical review of one of the selected session's recommended readings; due at the beginning of whatever 8 of the 15 sessions are preferred by the student.
- (25%) April 8th in-class mid-term exam.
- (20%) In class workshop (see below)
- (30%) Term project due by
May 10th at 12pm May 8th at 11:59pm midnight (no exceptions).
Policy on late work and incompletes: An incomplete will be given only in cases of documented personal catastrophe or illness. Late work will be accepted only if late submission is approved by the instructor prior to the due date, and only if it is submitted by the agreed make-up date.
Policy on absences: Please warn the instructor by email if you need to miss a class session.
In-class workshop (March 25th): Workshop session will focus on the theme “Application of Pavlovian and Instrumental Learning Concepts to Technological Problems”. Students are required to sign up for a topic at least three weeks prior to the workshop. Students may work as individuals or as a small group (if the proposed topic is deemed to be sufficiently thorough to justify collaborative work). Topics are expected to focus on either (a) critical treatment of recent concepts/proposals in the pertinent literature, or (b) development of a conceptual/theoretical framework for application of learning concepts to technological problems. All students/groups are expected to (1) develop and deliver an approximately 15 minutes-long talk on the chosen topic, (2) to meet with the instructor and/or the teaching assistant within a week before the workshop to review the material, and (3) to lead a brief (~10 minutes) discussion following their presentation. Grades will be based on the participation in the discussion as well as on the presentation. Criteria of evaluation: detailed evidence of intellectual mastery of the chosen topic, the empirical data base, and/or the model; clarity of exposition, including use of graphics and mathematical formulation where appropriate; adequacy of the mathematical/procedural description; thoroughness of scholarship; tightness of reasoning; creativity.
Term project: The term project may consist of
- A proposal, development, analysis, and implementation of a computational model relevant to the topics covered during the course. Students are expected to provide a detailed literature review encompassing empirical data, a review and critique of prior relevant models, a problem/computational analysis, and at least initial steps toward a neural network simulation, with at least a detailed plan for how the simulation results would be compared with selected pertinent data. Two types of simulations appropriate as part of the term project are: 1) a simulation that uses an already developed neural network architecture, but further probes its properties and limitations, or 2) a simulation based on a novel neural network architecture, and designed to assess its explanatory/predictive adequacy in some domain of conditioning, learning, or motivation. Or,
- A detailed analysis of one or more behavioral tasks that are used to probe different aspects of Pavlovian/Instrumental learning. Students are expected to provide a detailed literature review encompassing behavioral and physiologic data, to identify shortcomings/advantages of the selected behavioral task(s), and to critically review studies that deploy the selected behavioral task(s). Or,
- A novel proposal for an experimental (behavioral and/or neurophysiological) study related to the topics covered during the course. It is expected that the proposed experiment(s) are hypothesis-driven, novel, and feasible. Students are expected to provide a detailed review encompassing prior pertinent empirical data and/or models, to formulate sound hypotheses based on prior research, and to provide a detailed research proposal (including, but not limited to, experimental design, data collection, data analysis, and interpretation).
For the first alternative, the formal write-up should mimic the format of a modeling research article, with sections on: literature review, problem framing, methods, projected or actual simulation results, and a discussion. Each student/group is also expected to provide a working code (Matlab, C/C++, or Fortran) if the topic includes a computational treatment. For the second alternative, the term paper should be in the format of a review article, with sections on: introduction and task description, motivation for the task analysis, a thorough review of behavioral and physiologic data pertinent to the task(s), a comparison of the primary task(s) with alternatives, a discussion and recommendations. For the last alternative, an acceptable format is to prepare your report in the form of a grant proposal.
Collaboration: Students are encouraged to collaborate on preparation for the mid-term exam. Collaboration is encouraged for the workshop topics and term project as long as the topic/project is sufficiently detailed to justify the collaboration.
The College of Arts and Sciences Academic Conduct Code: All students entering Boston University are expected to maintain high standards of academic honesty and integrity. It is the responsibility of every student in the College and Graduate School of Arts and Sciences to be aware of the Academic Conduct Code's contents and to abide by its provisions. See: http://www.bu.edu/cas/academics/programs/conductcode.html. See also: http://www.bu.edu/grs/academics/resources/adp.html
Reserve readings, if any: CNS Department Library.