0% found this document useful (0 votes)
10 views16 pages

Cnuw 5.4

The slides of UW's Computational Neuroscience course on Coursera

Uploaded by

handong ji
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
10 views16 pages

Cnuw 5.4

The slides of UW's Computational Neuroscience course on Coursera

Uploaded by

handong ji
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 16

Where

 to  from  here?  

Hodgkin-Huxley

Biophysical realism Simplified models


Ion channel physics Fundamental dynamics
Additional channels Analytical tractability
Geometry
Dendritic computation
Neurons  have  complicated  spatial  structures  

Real Neurons Our Model


Geometry  matters!  

Input

Input

Inject  current  at  the  cell  body  and   Inject  current  in  a  dendrite  and  
record  effect  in  a  dendrite   record  effect  at  the  cell  body    

Voltage decays with distance in passive membranes (How?)

Dayan and Abbott


Linear  cables  

This  problem  sounds  familiar!  

Lord Kelvin
Voltage  V  is  a  func8on  of  both  x  and  t   (1824-1907)
Developed cable
  theory for
undersea cables

rm  and  ri  are  the  membrane  and  axial  resistances,  i.e.  


the  resistances  of  a  thin  slice  of  the  cylinder  
Johnston and Wu
The  cable  equation  

Before  we  had:  

Now  we  also  have  to  consider  ii,  the  current  down  the  cable,  
due  to  voltage  changes  in  x.    
That  current  works  against  internal  resistance,  ri    

or

where Time constant


Space constant
How  does  voltage  decay  in  space?  

⎛ | x| ⎞
⎜ − ⎟ Potential  decays  
v( x) ∝ e ⎝ λ ⎠
exponentially  
from  x  =  0  

InEinite  Cable,  
Constant  current  at  x  =  0  

Dayan and Abbott


How  does  voltage  decay  over  space  and  time?  
InEinite  Cable,  
Current  pulse  at  t  =  0,  x  =  0  

Potential  peaks  later  (and  


at  lower  values)  for  points  
further  away  from  input  
General  solution:  Eilter  and  impulse  response  

Diffusive

spread

Exponential

decay

OK:  now  what?    
1.  The  geometry  can  be  extremely  complicated  

Cable Equation
?
∂V 1 ∂ ⎛ 2 ∂V ⎞
cm = ⎜ a ⎟ − im + ie
∂t 2arL ∂x ⎝ ∂x ⎠

2.  And,  um,  ion  channels?  

è  Quickly  becomes  intractable  to  solve  analytically  for  


realistic  neurons  
 
Solution:  Divide  and  Conquer  
Compartmental  models  

Decreasing  number  of  “compartments”  


Each  compartment  =  one  dV/dt  equation    
(usually  no  dependence  on  x)  
The  gory  details  

And now you


see why
Genesis and Coupling
NEURON were conductances
developed—
thank you!

Dayan and Abbott


What  do  dendrites  add  to  neuronal  computation?  

London and Hausser, 2005


Delay  lines  in  sound  localization  

Spain;  Scholarpedia  
Direction  selectivity  
V(t)

V(t)

Rall; London and Hausser, 2005


Enthusiastically  recommended  references  
• Johnson  and  Wu,  Foundations  of  Cellular  Neurophysiology,  Chap  4  
The  classic  textbook  of  biophysics  and  neurophysiology:  lots  of  
problems  to  work  through.  Good  for  HH,  ion  channels,  cable  theory.  
 
• Koch,  Biophysics  of  Computation  
Insightful  compendium  of  ion  channel  contributions  to  neuronal  computation  
 
• Izhikevich,  Dynamical  Systems  in  Neuroscience  
An  excellent  primer  on  dynamical  systems  theory,  applied  to  neuronal  models  
 
• Magee,  Dendritic  integration  of  excitatory  synaptic  input,    
Nature  Reviews  Neuroscience,  2000  
Review  of  interesting  issues  in  dendritic  integration  
 
• London  and  Hausser,  Dendritic  Computation,    
Annual  Reviews  in  Neuroscience,  2005  
Review  of  the  possible  computational  space  of  dendritic  processing  

You might also like