INF5470 Fall 2012 Lecture 4: Coding in the Nervous System Content - - PowerPoint PPT Presentation

inf5470 fall 2012
SMART_READER_LITE
LIVE PREVIEW

INF5470 Fall 2012 Lecture 4: Coding in the Nervous System Content - - PowerPoint PPT Presentation

INF5470 Fall 2012 Lecture 4: Coding in the Nervous System Content Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions Lecture 4: Coding in the Nervous System 2


slide-1
SLIDE 1

INF5470 — Fall 2012

Lecture 4: Coding in the Nervous System

slide-2
SLIDE 2

Content

Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions

Lecture 4: Coding in the Nervous System 2

slide-3
SLIDE 3

Content

Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions

Lecture 4: Coding in the Nervous System 3

slide-4
SLIDE 4

Definition: Code

A code is simply a function from one signal domain onto another.

Lecture 4: Coding in the Nervous System 4

slide-5
SLIDE 5

Definition: information on a Code (1/2)

We do have knowledge about a code if we can assign unencoded and encoded signals to each-other for at least a subset of all possible signals. In other words if we can reconstruct some of the inputs to a system by observing internal encoded signals or vice versa, or if we can predict

  • utputs from a system by observing an internal signal or

vice versa.

Lecture 4: Coding in the Nervous System 5

slide-6
SLIDE 6

Definition: information on a Code (2/2)

An even less stringent requirement: we do have knowledge

  • f a code even if we can assign unencoded/encoded

signals to members of limited set of encoded/unencoded signals with given probability. I.e. if we can reconstruct a partial input from looking at a output code, in other words limit the possible inputs. Or correspondingly, predict an approximate output from looking at an internal/input signal.

Lecture 4: Coding in the Nervous System 6

slide-7
SLIDE 7

Content

Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions

Lecture 4: Coding in the Nervous System 7

slide-8
SLIDE 8

Muscle Output

Static Charges make Frog Legs Twich (Galvani 1780)

Lecture 4: Coding in the Nervous System 8

slide-9
SLIDE 9

Rate Response of Orientation Selective Cells in V1 (Hubel and Wiesel)

Lecture 4: Coding in the Nervous System 9

slide-10
SLIDE 10

Segmentation by Synchrony

Lecture 4: Coding in the Nervous System 10

slide-11
SLIDE 11

Content

Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions

Lecture 4: Coding in the Nervous System 11

slide-12
SLIDE 12

Precise Spike Patterns as Response to High Entropy Stimuli (Bair and Koch)

Lecture 4: Coding in the Nervous System 12

slide-13
SLIDE 13

Synfire Chains (Abeles)

Lecture 4: Coding in the Nervous System 13

slide-14
SLIDE 14

Phase Shift in Hippocampal Place Cells (O’Keefe and Recce)

Lecture 4: Coding in the Nervous System 14

slide-15
SLIDE 15

Spike Timing Dependent Plasticity, Spike Based Learning

Lecture 4: Coding in the Nervous System 15

slide-16
SLIDE 16

Reaction Time In Psychophysical Recognition Tasks (Simon Thorpe) (1/2)

150ms to 200ms reaction time in visual recognition task

Lecture 4: Coding in the Nervous System 16

slide-17
SLIDE 17

Reaction Time In Psychophysical Recognition Tasks (Simon Thorpe) (2/2)

Lecture 4: Coding in the Nervous System 17

slide-18
SLIDE 18

Content

Definition: Knowledege of a Code Experiments: Rate Codes Experiments: Temporal Codes Coding Categories Overview Weekly Questions

Lecture 4: Coding in the Nervous System 18

slide-19
SLIDE 19

Coding Categories Overview

Lecture 4: Coding in the Nervous System 19

slide-20
SLIDE 20

Example Coding-Function: Latency Encoding

Lecture 4: Coding in the Nervous System 20

slide-21
SLIDE 21

Weekly Questions

  • 1. How is a reconstruction from a rank order code of a

black and white image impaired? Imagine 2x2 pixels, two of them black, the other two white. What information is lost in the encoding?

  • 2. Can you again explain why a rate-, population code

can also lead to overall fast response times and might be an alternative explanation of the Thorpe reaction times experiment?

Lecture 4: Coding in the Nervous System 21