|
Instructor: Yael Niv
Email: yael at princeton dot edu
Office hours: by appointment at 3-S-11
AI: Nathan Parker
Email: nfp at princeton dot edu
Office hours: Fridays 2-3pm at 2-C-18
Format and content:
The purpose of this course is to provide a modern,
integrative view of classic animal learning phenomena from
experimental psychology, through the lens of contemporary learning
theory, computational models of learning and decision making, and
current neuroscientific knowledge. Each week we will focus
on one topic in experimental psychology. We will first discuss the major
behavioral findings and the learning theory that pertains to them. Following
this, we will discuss computational
theories pertaining to the topic and what is currently known about
the neural basis of these behaviors or computations.
Course requirements:
Active class and precept participation (20%), reading of assigned
literature, 4 out of 5 problem sets (40%), final Exam (40%; open
book/notebook; no laptops by university rules).
Syllabus topics and presentations*:
Note: handouts are printed before class and are a tentative idea of
what we will talk about. They also have intentional gaps where I will ask
you questions and prefer to not reveal the answers yet. Slide
presentations are put up after the class and account for what we actually
covered, including corrections of errors in the handouts which we noticed
in class.
- Week 1 - Introduction - motivation,
structure, requirements of the course; defining learning and decision making
slides
- Weeks 2-4 - Classical conditioning -
basic phenomena, error driven learning and the Resorla-Wagner
model, second order conditioning and the Temporal Difference model,
dopamine in the basal ganglia, opponent processes: appetitive vs.
aversive, inhibitory vs. excitatory, opponent process model, fMRI of
temporal difference errors
slides I
slides II
slides III
slides IV
slides V
slides VI
- Weeks 5-7a Instrumental conditioning -
Thorndike, basic procedures, Skinner & free operant behavior,
modeling action selection, Actor/Critic framework, neural substrates of
Actor/Critic model, learning Q values, comparing models using neural data,
modeling free operant behavior, actions and habits, S-R versus R-O,
devaluation, dual neural pathways, uncertainty based arbitration
slides I
slides II-III
slides IV
slides V
slides VI
- Week 7b: Classical vs. instrumental conditioning
- comparison of instrumental and Pavlvovian procedures and processes,
omission (we did not cover Pavlovian to Instrumental transfer this year)
slides
- Weeks 8-9: Too much dopamine can be bad for
you: Learning as a model of disorders - drug abuse, latent
inhibition and schizophrenia
addiction I
addiction II
latent inhibition
- Week 10: Extinction - What do animals
really learn? - renewal, reinstatement and spontaneos recovery,
Bayesian inference and infinite capacity latent cause models
slides
- Week 11-12: Generalization and discrimination -
phenomena & challenges, configural vs. elemental theories
slides
- Week 12: Summary & a brief touch on topics we
did not cover - attention and associability, other neuromodulators,
amygdala, summary - the big picture, advice for a young investigator
slides
* Many thanks to all my colleagues and fellow teachers who have (mostly
unknowingly) contributed much of the material that is included
in the above slides
Readings: (for PDFs go to blackboard)
- Textbook: Learning and Memory: From
Brain to Behavior by Gluck, Mercado & Myers (2nd ed) (See also
the companion website for chapter outlines, summaries and online quizzes.)
Note that this is not our textbook in the traditional sense, in that
it covers more than we will be discussing (eg, memory) and does not
cover everything that we will be discussing (eg, computational
models)
Some books I recommend:
Interesting (and
useful) Links:
|