Transcript Dynamical Neuroscience: A Viewpoint

“Can we predict synchrony and
asynchrony in networks coupled by
multiple dendritic gap junctions?”
Frances K. Skinner
Toronto Western Research Institute
University Health Network and
University of Toronto
New York University
April 13, 2008
From Scholarpedia:
Mathematical Biology article of
Frank Hoppensteadt
• “highly interdisciplinary nature”
• “barriers to collaborations between mathematicians and biologists”
• “a shift from mathematical analysis to computer simulation due mostly
to improvements in computer power and accessibility.. With the shift
being made possible to include more information in models and still
derive useful insights from them.”
Especially in neuroscience with all the details being uncovered,
increasingly sophisticated techniques etc. these comments are very
timely. With increasing specialization and interdisciplinarity and
potential moving apart of mathematical and biological sciences (or
separation of viewpoints) we organized a Theoretical Neuroscience
Minisymposium at 2006 Society for Neuroscience meeting, one of the
aims being to help counteract this.
Interneuron Heterogeneity
Domain-specific innervation of hippocampal interneurons
apical
basal
purple –laminae
where axonal arbor
typically extends
turquoise indicates
that other interneurons
rather than principal
cells are targets
Different types of interneurons containing calcium-binding proteins and neuropeptides
McBain and Fisahn NRN 2001
• General challenge – how to best consider
various neurobiological details.
• Specific challenge – understanding the
contribution of electrical coupling in
different contexts.
• Outline of Talk: background, discussion of
some of our previous work, and then get to
question posed for this talk.
“Can we predict synchrony and
asynchrony in networks coupled by
multiple dendritic gap junctions?”
Frances K. Skinner
Toronto Western Research Institute
University Health Network and
University of Toronto
New York University
April 13, 2008
Acknowledgements
Tariq Zahid
Fernanda Saraga, Leo Ng
NSERC of Canada
Computing support – RIS of UHN
The hippocampus (part of medial temporal lobe) is
an intensely studied region of the brain because:
• It is associated with memory and learning (i.e., LTP,
LTD), epileptic seizures, and neurogenesis.
• It exhibits a wide range of population rhythmic activity
patterns (<1 to >200 Hz) that are associated with
various behavioural states.
• It is amenable to experiment, retaining its synaptic
circuitry and thus population activities in the slice.
EEG activities of
mouse hippocampus
Electrode location
Theta-Gamma
SPW-ripples
Sharp wave-ripples
Spontaneous Rhythmic Field Potentials (SRFPs)
(Liang Zhang’s lab; Wu et al. J Physiol 2002, J Neurophysiol 2005)
Gillis et al., J. Neurosci. Meth.
(2005)
Blockade of field rhythms and pyramidal IPSPs
by GABA-A receptor antagonist
rhythmic activities also dependent on electrical coupling (gap junctions)
Background
Electrical coupling (i.e., gap junctions) is present
in much of the mammalian brain (e.g., inferior
olive, striatum, neocortex, hippocampus, retina,
thalamus).
In particular, gap junctions occur between
inhibitory cells, often of the same type,
and can be located at sites quite distant
(> 200 μm) from the soma.
Interneurons
represent 10-20%
of the neuronal
population but may
provide the precise
temporal structure
necessary for
ensembles of
neurons to perform
specific functions.
- Buzsáki and Chrobak, 1995
Interneuron Heterogeneity
Domain-specific innervation of hippocampal interneurons
apical
basal
purple –laminae
where axonal arbor
typically extends
turquoise indicates
that other interneurons
rather than principal
cells are targets
Different types of interneurons containing calcium-binding proteins and neuropeptides
McBain and Fisahn NRN 2001
Background
Gap junctions located far
from cell bodies, at
non-proximal sites
(basket cells in hippocampus)
Gap junctions can be
modulated
Inhibitory cells have active
dendrites, spikes can be
generated in dendrites
From Fukuda & Kosaka, J Neurosci 2000
Dendrodendritic Gap Junctions
Fukuda & Kosaka, J Neurosci 2000
Model (Hippocampal Basket Cell)
Passive dendrites
50 mV
372-compartment
model developed
in NEURON
100 ms
Morphology from
Gulyas et al.(1999)
Saraga et al., J Neurophysiol 2006
WB used for kinetic model basis, Martina and Jonas (1997), Martina et al (1998)
used as conductance value basis and spike characteristics and electrophysiological
responses from Morin et al. (1996) and van Hooft et al. (2000).
“Reduced” 3-compartment model
based on matching electrotonic
length from soma (Vout)
s
d
d
Electrotonic distance to soma (L)
Vin
1.2
0.4
0
100
300
500
Anatomical distance to soma (mm)
Electrotonic distance from soma (L)
Vout=0.14
Vin =0.25
Vout
0.12
0.04
0
100
300
500
Anatomical distance from soma (mm)
(Distal) phase
response curves (PRCs)
Voltage along
dendrite
s
d
s
d
d
d
Using the reduced model geometry
0.5%
Phase Shift
1%
1.5%
10%
50 mV
10 ms
Phase
PRCs calculated using
XPPAUT (Ermentrout, 2002)
LOW
MEDIUM
HIGH
75
26
26
24
24
50
% Phase Lag
23
22
20
20
25
18
18
16
16
14
14
0.1
0.1
11
10
10
% Active
% Active
0
0
100
100
% Phase Lag
28
28
Intrinsic Frequency (Hz)
Intrinsic Frequency (Hz)
Predicted Network Dynamics
Intrinsic
Frequency
% Phase
Lag
Weakly coupled oscillator
theory used to define three
different dynamic regions
-LOW, MEDIUM, HIGH
that refer to PRCs with
particular characteristics
(e.g., negative PRCs
for MEDIUM)
Phase lags determined from
interaction functions calculated
using XPPAUT (Ermentrout, 2002)
Results
Simulations confirm
theoretical predictions
28
LOW
MED
HIGH
75
75
50
50
22
% Phase Lag
Intrinsic Frequency (Hz)
24
25
20
25
18
% Phase Lag
Intrinsic Frequency (Hz)
% Phase Lag
26
Intrinsic
Frequency
% Phase
Lag
0
16
14
0.1
0.1
11
10
10
0
100
100
% Active
% Active
•“Weak coupling” is about 10 pS (comparing predicted and simulated)
• Compare full and reduced model phase lag values to
“define” synchronous and asynchronous
• Synchronous is 10% or less phase lag, asynchronous otherwise
Cell 1: 15% basal attenuation, 2% apical attenuation
Cell 2: 8% basal attenuation, 8% apical attenuation
Cell 3: 6% basal attenuation, 14% apical attenuation
apical
apical
CELL 3
basal
basal
Cell 1, apical coupling
(multistability)
Beyond weak coupling
Beyond weak coupling
Beyond weak coupling
Discussion and Conclusions
• PRC skewness quantifications can be used to predict whether
synchronous or asynchronous modes occur in electrically coupled
basket cells.
• Averaged PRCs can be used to predict modes for coupling at multiple
sites.
• Predictions cannot be made under all circumstances and multistability
can occur.
• Different apical and basal attenuation (due to different channel
densities) allow more ‘robust’ asynchrony to occur with coupling on
the more attenuated dendritic side.
• Network couplings that produce asynchrony (as compared to
synchrony) with weak coupling encompass more dynamic richness
(i.e., range of possible phase lags) with gap junction conductance
changes.
• Thus, gap junction coupling may be able to tune networks in and out of
synchronous activities if asynchrony with weak coupling is predicted.
the end
• Thank you!