Dieter Jaeger

1
How neurons integrate thousands of synaptic
inputs each second
Dieter Jaeger Department of Biology Emory
University djaeger_at_emory.edu
2
(No Transcript)
3
The textbook view
4
KSJ 4th ed., Fig. 10-7
5
Kandel, 4th edition
6
(No Transcript)
7
In vivo input levels
100 mm
100 mm

• GP neuron
• surface area 17,700 mm2
• number of synapses (ex/in) 1,200 / 6,800
• number of inputs / s 12,000 / 6,800

• Ca3 pyramidal neuron
• surface area 38,800 mm2
• number of synapses (ex/in) 17,000 / 2,000
• number of inputs / s 170,000 / 20,000

8
In vivo recording from striatal medium spiny
neuron
9
Coding with thousands of inputs each second
10
5,000 AMPA and 500 GABAA Synapses at 10 Hz
Ein -70 mV
Eex 0 mV
11
dynamic current clamp
DCN neuron
patch pipette
Isyn Iex Iin Gex(Vm-Eex)
Gin(Vm-Ein)
Isyn
Vm
AxoClamp 2B
Vm
slice, 32 C
Isyn
To apply in vivo like input
12
Dynamic current clamping of GP neuron
13
current versus conductance source
Vm
Esyn
- 40 mV
5 mV
Isyn
outward
Iexp
0 nA
inward
0.2 nA
100 msec
14
spike triggering events
Input frequency
1.0
Input conductance
input synchronization
10 groups
100 groups
50 ms
15
Small conductance KCa current (Sk)
16
The effect of Sk block on synaptic integration
17
Space! The next frontier
18
(No Transcript)
19
(No Transcript)
20
Shunting by somatic conductance
21
Shunting by distributed conductance
22
(No Transcript)
23

• Functional Implications
• synaptic conductance stabilizes Vm through
  shunting
• spikes can only be triggered from transients
• spikes reflect inputs correlated on the order of
  1-10 ms
• spike rate reflects correlation as well as input
  rate
• inhibition has equal access to the control of
  spiking

24
More complexity to come

• gap junctions
• short term plasticity (history dependence)
• calcium signaling
• dendritic spike initiation

25
Acknowledgements
Contributors Volker Gauck Svetlana Gurvich Lisa
Kreiner Mayuri Maddi Kelly Suter
Other Lab Members Alfonso Delgado-Reyes Jesse
Hanson Chris Roland Simon Peron
26
(No Transcript)
27
(No Transcript)
28
Current models of basal ganglia function
determine spike rates based on simple summing of
synaptic inputs
Normal Parkinsons Disease
29
cerebellar circuit
cerebellar cortex
Cerebellar cortex
!?
deep cerebellar nuclei
DCN
mossy fibers
climbing fibers
from Paxinos Watson, "The rat brain', Academic
Press
30
The effect of synchronization
100 independent inputs
10 independent inputs
20 mV
-50 mV
200 msec
200 msec
31
precision rate
spike timing precision
spike frequency

rel.
gain factor
gain factor
synchronization
high
intermediate
none
32
spiking in vitro and in vivo
in vitro
20 mV
200 msec
in vivo, awake (from LeDoux et al. 1998,
Neuroscience, 86(2)533)
500 msec
10 msec
time scale for coding
rate code
temporal code
33
Constructing in-vivo like synaptic input
gmax 2.1 pS - 69 pS gain 0.5 -
gain 16
Gin 1 nS at gain 1
Gex
30,100 UCs/s
0.5
inhibitory unitary conductance
0
Esyn
- 40 mV
10 mV
100 ms
34
(No Transcript)
35
Shink and Smith, J. Comp. Neurol. 358 119-141
(1995)
36
100 mm
100 mm

• DCN neuron
• surface area 11,056 mm2
• number of synapses (ex/in) 5,000 / 15,000
• number of inputs / s 25,000 / 750,000

• Purkinje cell
• surface area 261,000 mm2
• number of synapses (ex/in) 175,000 / 5,000
• number of inputs / s 350,000 / 10,000

37
100 mm

• Cerebellar Stellate cell
• surface area 2,305 mm2
• number of synapses (ex/in) 1,000 / 100
• number of inputs / s 2,000 / 200

38
(No Transcript)
39
(No Transcript)
40
-70 mV Eleak
41
(No Transcript)
42
(No Transcript)