Statistical and dynamical models of brain function

UNIZH code: INI-416 / ETHZ code: 402-0818-00

Instructors: Stefano Fusi and Hans Scherberger

T.A.: Jason Rolfe
E-mails: {fusi, hjs, rolfe}@ini.phys.ethz.ch

SS 2007, ETH Physics (14 weeks)
Time: DI 16:00-18:00 V / DI 18:00-19:00 U
Room: Y23 G04
TA office hours available on request - email rolfe@ini.phys.ethz.ch

 

This web page will be updated through the whole period of the lecture. Usually the lecture slides will be posted the day after each lecture and will appear along the outline of the lecture as .pdf attachments.

     SPECIAL ANNOUNCEMENT: Lecture 5 will be given from 1-3pm on Monday April 23    
     in the ELK ROOM OF THE INI   (Building 55, ground floor)
     NOTE THAT THIS IS A DIFFERENT ROOM THAN USUAL

Recomended Books:

  • Dayan and Abbott, Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems.

  • Hertz, Krogh, and Palmer, Introduction to the Theory of Neural Computation

  • Kandel Schwartz, and Jessell, Principles of Neural Science.

  • Martin, Wallace, Fuchs, and Nicholls, From Neuron to Brain: A Cellular and Molecular Approach to the Function of the Nervous System.

  • Rieke, Warland, de Ruyter van Steveninck, and Bialek, Spikes: Exploring the Neural Code
 
1.Outline of Lecture      2.Exam       3.Homework
 
  
I. Outline of lecture
  1. Introduction: Brain hardware
anatomy - physiology - differences of different species - Why do we need a brain?
  2. Brain-machine-interfaces
history, sources of information, state of the art.  
    Paper, main video, supplementary video 1, supplementary video 3, supplementary video 4, supplementary video 5 
  3. Principles of neural data analysis
single unit analysis - methods of analysis: PSTH, ISI, spike spectrum, spike correlations
  4. Analysis of the LFP
nature of the LFP; methods of analysis: evoked potentials, LFP spectrum, coherency, multi-taper analysis
  5. Decision making - MEETING APRIL 23 FROM 1-3PM in the Elk room of the INI (building 55, ground floor)
perceptual decisions (Newsome experiment), controlling behavior with microstimulation,
decisions with free-choice
  6. Reading the mind: decoding brain functions
classifier and learning algorithms: Bayesian decoding; linear discriminant analysis; Kernel methods;   
  7. Modeling brain functions
    the brain as a physical dynamical system - attractor neural networks and working memory - 
    Hopfield networks: digging valleys in an energy landscape - Memory capacity.
  8. Mean field approach (and ppt)
    simple techniques from statistical mechanics, working memory in inferotemporal cortex:
    experiments and interpretation 
  9. Towards more realistic networks
    integrate-and-fire neurons - Fokker-Planck approach.
  10. Learning and memory
The basic concepts of learning - learning and synaptic plasticity
learning as a stochastic Markov process - The necessity of biological complexity
  11. Learning to decide
- Combining decision making with learning - Experiments of E.K. Miller (MIT) in prefrontal cortex
  12. Attention models of biased competition
    models of attention with learning
  13. Closing lecture
    tba.
 
II Exam
The exam is passed if 50% of the homework is done correctly throughout the course period.
																			
III Homework    
  
Homework is given on a weekly basis,  it is supposed to be handed-in (or sent by e-mail) within a week.  
Homework will not be accepted more than two weeks after it is assigned.  The corrected homework will be 
handed back a week later either at the lecture or standing at Jason's desk.  The homework will be posted 
on this page every week.
Homework list

  1. 03/20/2007

  2. 03/27/2007 (doc) (pdf) 

  3. 01/04/2007 (pdf) - due next lecture at the very latest

  4. 10/04/2007 (pdf) - due in 2 weeks at the very latest - color copies in Jason's mailbox

  5. 23/04/2007 (doc) (pdf) - (paper) 

  6. 24/04/2007 (doc) (pdf) 

  7. 08/05/2007 (pdf) (hopfield.m)

  8.  

  9. See lecture slides 

  10.  

  11.  

  12.  

  13. no homework