3rd dendrites workshop
official CNS*2015 workshop
|CNS*2015 workshop, Prague, July 22 in room RB211
3rd workshop on
Dendrite function and wiring: experiments and theory
|(please note that there were some last minute changes!)|
|Giorgio Ascoli (George Mason University) keynote|
|Claudia Clopath (Imperial College London)|
|Dieter Jaeger (Emory University)|
|Peter Jedlicka (Goethe University, Frankfurt)|
|Greg Jefferis (University of Cambridge)|
|Daniel Justus (German Center for Neurodegenerative Diseases DZNE, Bonn)|
|Athanasia Papoutsi (IMBB-FORTH, Heraklion-Crete)|
|Arnd Roth (University College London)|
|Balázs Ujfalussy (Institute of Experimental Medicine HAS, Budapest)|
|Katharina Wilmes (Humboldt University Berlin)|
|Giorgio Ascoli: Reconstructing dendrites: from development to computation|
A quantitative systems-level understanding of how multiple
local interactions of cytoskeleton components during differentiation
define mature dendrite arbor shape is still missing.
How are these processes regulated in a class-specific manner
to give rise to the characteristic arbor shapes of different neuron classes?
I will describe a closed-loop modeling approach to address these issues
with computational simulations constrained by and validated
against custom-designed experimental data.
First, focusing on genetically modified Drosophila larval sensory neurons,
we simultaneously acquire temporal sequences of whole dendrite arbor
structures and multiple co-registered subcellular components
by high-resolution time-lapse multi-channel confocal microscopy in vivo.
We then digitally reconstruct both the neurites and the cytoskeletal
distributions by quasi-automated tracing,
thus enabling high-throughput identification of spatial-temporal
associations between key molecular dynamics (actin and tubulin polymerization)
and quantitative morphometric features.
Lastly, we use the extracted statistics to stochastically simulate
the growth of anatomically realistic virtual neuronal analogues
and underlying developmental processes.
Comparing simulated and real arbors as well as their subcellular components
helps assess the plausibility of the model
(that is, of the hypothesized biological interactions)
or reveal additional necessary mechanisms.
This strategy requires integration of different disciplines,
including genetics and developmental neurobiology,
in vivo multi-parameter confocal imaging, algorithmic tracing,
computational modeling, quantitative morphological analysis,
Most importantly, this research would be impossible without free
and public community sharing of both experimental data sets
(through NeuroMorpho.Org) and corresponding (animal and computational) models.
|Claudia Clopath: Synaptic plasticity across dendritic location|
Dendrites are not just passive cables but exhibit many active mechanisms.
These mechanisms result in a highly non-linear behaviour,
which depends on the location of the synapses on a dendrite.
We investigate how synaptic plasticity depends on the dendritic location.
In particular, we implemented a voltage-based spike-timing dependent learning rule,
combined with different dendritic properties.
We found that synaptic plasticity shapes different connectivity patterns depending
on the synaptic location, yielding different functions.
This work demonstrates that single neurons can perform complex computations
shaped by dendritic synaptic plasticity.
|Dieter Jaeger: Modeling Dendritic Function in Globus Pallidus Neurons|
|Peter Jedlicka: Biologically realistic models of dendritic and synaptic plasticity in the hippocampus|
LTP and LTD that occur in hippocampus are widely accepted to be synaptic mechanisms
involved in learning and memory.
It remains uncertain, however, which particular activity rules are utilized by hippocampal neurons
to induce LTP and LTD in behaving animals and what is the role of dendritic morphology in this context.
In the first part of my talk,
I am going to present simulations which indicate that the interplay of STDP-BCM plasticity rules
and ongoing neuronal activity is able to account for experimentally observed amplitudes of LTP and LTD
in the dentate gyrus of awake rats.
In the second part of my talk, I will be describing anatomically detailed models of hippocampal neurons
which predict that changes in dendritic morphology are able to selectively modulate local synaptic plasticity.
|Greg Jefferis: NBLAST: Rapid, sensitive comparison of neuronal structure and construction of neuron family databases|
Neural circuit mapping efforts in model organisms are generating
datasets of 10,000s of labelled neurons, demanding
new computational tools to search and organise them.
We present a general, sensitive and rapid algorithm, NBLAST,
for measuring pairwise neuronal similarity.
NBLAST considers both position and local geometry,
decomposing a query and target neuron into short segments;
matched segment pairs are scored using a probabilistic scoring matrix
defined by the statistics of matches and non-matches.
We validated NBLAST on a published dataset of 16,129 single Drosophila neurons.
NBLAST distinguishes two images of the same neuron and neuronal types
without a priori information.
Cluster analysis of extensively studied neuronal classes
identified new types and unreported topographical features.
NBLAST supports additional query types including matching neurite tracts
with transgene expression patterns.
We organise all neurons into 1,052 clusters of highly related neurons,
simplifying exploration and identification of neuronal types
including sexually dimorphic and visual interneurons.
|Daniel Justus: Locomotion-speed dependent disinhibition of inputs to CA1 pyramidal neurons is mediated by a medial septal glutamatergic circuit|
Before the onset of locomotion the hippocampus undergoes a transition
into an activity-state specialized for the processing of spatially related
information. This brain-state transition is associated with increased
firing rates of CA1 pyramidal neurons and
the occurrence of theta oscillations,
which both correlate with locomotion velocity.
Moreover, during locomotion place related input from CA3
and entorhinal cortex is integrated by CA1 pyramidal neuron dendrites
and contributes to the formation of spatial memories.
However, the neural circuit by which locomotor activity is linked
to hippocampal oscillations and neuronal activity is unresolved.
Here we reveal a mechanism that, via glutamatergic septo-hippocampal
projections onto alveus/oriens interneurons, regulates
feed-forward inhibition of Schaffer collateral
and perforant path inputs onto CA1 pyramidal neurons
in a locomotion-speed dependent manner.
It allows these inputs to be integrated on CA1 pyramidal neuron
dendrites more efficiently during faster locomotion
and thus leads to increased firing rates and enhanced processing
of spatial information.
|Athanasia Papoutsi: Modeling the interplay of dendritic spikes and network connectivity in persistent activity.|
Pyramidal cells are embedded in networks of thousands of neurons and process
information in order to mediate cognitive functions.
This information processing depends on the mode (linear or non-linear)
of synaptic integration that takes place in their dendritic compartments
and on the non-random local circuitry profile.
In the prefrontal cortex
(PFC) in particular, which mediates working memory and attentive tasks,
the interaction between the dendritic and network properties
(size/connectivity) enables the emergence of short-term memory
through persistent spiking activity.
However, the ways in which highly reciprocally connected PFC microcircuits
give rise to this prolonged spiking activity are not well understood.
To address this question, we undertook a modeling approach
whereby we simulated both small- and
large- scale networks of L5 PFC circuits.
We show that sufficient synaptic drive to induce dendritic spikes
reduces the minimum network size required for persistent activity induction.
Even though relaxation of connectivity and synaptic delay constraints
eliminate the gating effect of dendritic spikes,
this is done at the cost of much larger neuronal networks.
Overall, weprovide evidence for a tight interaction between
dendritic non-linearities and network properties in mediating
the short-term memory function.
|Arnd Roth: Untangling cerebellar circuits with scanning electron microscopy and focused ion beam milling|
Understanding how neural circuits work requires detailed knowledge
of the connectivity patterns between different circuit elements,
and ultimately the complete wiring diagram.
To obtain such information, a number of strategies are currently
being developed for the automated acquisition of
three-dimensional image data of neural tissue at ultrastructural resolution.
Serial sectioning combining a focused ion beam and
a scanning electron microscope (FIBSEM; Knott et al., 2008),
is a particularly promising strategy for achieving high spatial resolution
in all three dimensions, which is essential to facilitate
the automated tracing of dendrites, axons and glial processes
in the neuropil as well as the identification of synapses.
We have developed new approaches for extending the volume of tissue
that can be imaged using FIBSEM, and are using this approach to establish
the rules governing the connectivity of parallel fibres to
interneurons and Purkinje cells in the molecular layer of
the cerebellar cortex.
|Balázs Ujfalussy: Discovery of presynaptic ensembles by structural and intrinsic plasticity in dendritic branches|
The dendritic tree of cortical neurons is the locus
of a wide variety of powerful plasticity mechanisms
but little is known about the functional principles
underlying these processes.
Based on the principle that dendritic plasticity tunes
sub-cellular connectivity and dendritic nonlinearities
so that dendritic branches can best detect the activation
of presynaptic ensembles, we derived optimal rules
with biological approximations for structural and intrinsic plasticity.
We show that the optimal structural plasticity rule efficiently
identifies ensembles by clustering synapses along the dendritic tree,
and its input timing dependence matches experimental data
from cortical neurons. The same principle further predicted
an intrinsic plasticity rule to fine-tune the nonlinear properties
of dendritic branches to the dynamics of their presynaptic ensembles,
reproducing experimentally observed forms of branch-strength potentiation.
Our approach provides a novel framework for studying plasticity
in individual dendrites from the perspective of circuit-level computations.
|Katharina Wilmes: Local dendritic inhibition as a simple pathway-specific switch for Hebbian synaptic plasticity|
Synaptic plasticity is the foundation of learning in the brain’s neural network.
To ensure the stability of established network structures, however,
plasticity should also be regulated. Here, we propose that
compartmentalized dendritic inhibition could modulate synaptic plasticity
by gating dendritic signals. Hebbian forms of learning require a signal
of postsynaptic activity at the synapse.
So far, the only known mechanism mediating such a signal is the backward propagation
of action potentials into the dendrite and, indeed, forms of LTP and LTD in various places
of the dendritic tree have been found to depend on the back-propagating action potential.
The rules for induction of plasticity vary with synaptic location, potentially due to
the characteristics of the backward-directed signal arriving at the synapse.
In the apical tuft of pyramidal neurons, for example,
back-propagating action potentials may fail to reach the synapse and hence to trigger LTP.
When paired with coincident dendritic input, however, they can evoke a non-linear calcium event
that serves as a signal for coincident pre- and postsynaptic activity and a trigger for plasticity.
Using a conductance-based multi-compartment model,
we show that inhibition can gate these backward-directed signals in an all-or-nothing manner,
allowing for a rapid and profound change in the learning rule of downstream synapses.
Moreover, the switch-like mechanism enables a differential modulation of plasticity at synapses
in different dendritic compartments, potentially contributing to a pathway-specific regulation
of plasticity. Our quantitative analysis reveals that inhibition of the
back-propagating action potential is feasible if inhibition is precisely timed
on a millisecond timescale.
Inhibitory decay time constants close to those reported for GABA-A mediated inhibition
are sufficient for regulation.
In summary, we propose that the regulation of back-propagating dendritic signals
through localized dendritic inhibition constitutes a useful mechanism in the context
of learning and the formation of memories.