| Friday, May 4 |
Thackeray Hall
Room 704 |
| 8:45 am - 9:15 am |
Coffee, Bagels & Doughnuts • Opening remarks |
|
| 9:15 am - 10:15 am |
Hidden structure in neocortical networks revealed by activity-dependent
gene expression
Abstract
- Unbiased methods to assess the firing activity
of individual neurons in the neocortex have revealed that a
large proportion of cells fire at extremely low rates (<0.1
Hz), both in their spontaneous and evoked activity. Thus,
firing in neocortical networks appears to be dominated by a
small population of highly active neurons. Here we use a
fosGFP transgenic mouse to identify and characterize the
properties of cells with a recent history of elevated activity in primary somatosensory cortex. Layer 2/3 neurons expressing fosGFP fired at higher rates compared to
fosGFP- neurons, both in vivo and in vitro. Elevated activity could
be attributed to increased excitatory and decreased
inhibitory synaptic drive to fosGFP+ neurons, in particular
from layer 4, the input layer of the cortex. Paired-cell
recordings indicated that fosGFP+ neurons had a greater
likelihood of being connected to each other. Thus,
heterogeneous firing rates can be attributed to
differential wiring of neocortical neurons.
|
Alison Barth
Carnegie Mellon University |
| 10:15 am - 10:30 am |
Coffee Break |
|
| 10:30 am - 11:30 am |
Runaway consolidation of network activity mediated by metaplasticity.
Abstract
- Hebb proposed that patterns of “reverberation” in neuronal
networks could be consolidated by synaptic strengthening, and vice versa.
This implies a positive feedback loop which is potentially epileptogenic
and requires compensatory mechanisms to keep both network activity and
synaptic strength within favorable operating regimes. In small networks
of hippocampal neurons, we observe that synaptic potentiation consolidates
network reverberation, but that this reciprocal positive feedback stops
quickly after a few episodes of reverberation, yield networks which are
just connected enough to reverberate. However, following chronic
inactivation of such networks, we observe a three-stage response: prior
to network reactivation, both macrosynaptic and network properties are
similar to controls; shortly after reactivation, both reverberatory
activity patterns and synaptic efficacy were increased; and at variable
latency after these increases, chaotic spontaneous discharges emerge and
persist for the duration of the experiment. These data argue that ongoing
synaptic metaplasticity constrains positive feedback and network behavior,
but that this feature may be maladaptive under pathological conditions of
inactivity.
|
Rick Gerkin
Carnegie Mellon University |
| 11:30 am - 12:00 pm |
Optimizing memory using heterogeneity in neuronal networks.
Abstract
- Spatially localized bumps of neural activity have long
been proposed as a mechanism of short term memory in neuronal networks
(Amari 1977). Neural field equation models of bump formation are often
translationally invariant so that the bump can be initiated at any location
in the network. One consequence of this degeneracy is that noise causes
the bump to diffusively wander such that the memory of its original location
is quickly lost. One remedy to this problem is to break the translation
invariance of the network so a stable bump can only lie on one of a chain
of discrete states, rather than on a continuum of marginally stable states
(line attractor). We show this can be accomplished by including periodic
inhomogeneities in the synaptic weight kernel of the neural field. This
ultimately helps stabilize bumps in the presence of noise so the memory
of their initial condition is much better retained.
|
Zachary Kilpatrick
University of Pittsburgh |
| 12:00 pm - 1:30 pm |
Catered Lunch |
|
| 1:30 pm - 2:30 pm |
Phase transitions and information processing in the brain.
Abstract
- The cerebral cortex is a highly complex network comprised
of billions of excitable nerve cells. Dynamic interactions among these cells
underlie our thoughts, memories, and sensory perceptions. A healthy brain must
carefully regulate its neural excitability to optimize information processing
and avoid brain disorders. If excitability is too low, neural interactions are
too weak and signals fail to propagate through the brain network. On the other
hand, high excitability can result in excessively strong interactions and, in
some cases, epileptic seizures. While it is commonly supposed that healthy neural
excitability must lie between these extremes, the optimal degree of excitability
is not known. In this talk, I will present experimental evidence that brain
dynamics undergo a phase transition as neural excitability is tuned from low to
high. Importantly, the critical excitability at which the phase transition occurs
also results in optimal information processing. These results suggest that the
optimal excitability is that which places the brain closest to the phase transition.
Moreover, many mental disorders such as epilepsy, Down syndrome, and autism may be
caused by deviation from this optimal excitability.
|
Woodrow Shew
University of Arkansas |
| 2:30 pm - 3:30 pm |
Stochastic dynamics on networks.
Abstract
- Dynamical systems defined on networks have applications
in many fields, including computational and theoretical neuroscience. In
particular, it is important to understand when networks exhibit synchronous or
other types of coherent collective behaviors. Other questions include whether
such coherent behavior is stable with respect to random perturbation, or what
the detailed structure of this behavior is as it evolves. We will examine several
models of networked dynamical systems and present a mixture of results that range
from rigorous theorems for abstract models to quantitative comparisons of models
and data.
|
Lee Deville
University of Illinois-Urbana Champaign |
| 3:30 pm - 4:00 pm |
Coffee Break |
|
| 4:00 pm - 5:00 pm |
Agent-Based Modeling of the Dengue Virus Vector in Realistic Communities.
Abstract
- The WHO estimates that about half of the world's population is at
risk of contracting Dengue, a vector-borne virus for which there is
no cure. Researchers are hopeful that current clinical trials may
yield an effective vaccine in the next two years. Current
interventions emphasize control of the disease vector, Ae. aegypti.
We have created an Agent-Based Model (ABM) that explicitly
represents both hosts and vectors at virtually all life stages as a
means to simulate disease propagation at the village/regional level
and evaluate intervention mechanisms. This model, called CLARA, has
been calibrated using a variety of datasets, with in-depth attention
given to the timing and effects of infecting mosquitos with various
strains of Wolbachia bacteria, thereby rendering them
vector-incapable. We will discuss the core framework of the model,
performance metrics, the calibration processes, some validation
exercises, and guidance presented to field entomologists based on
results to date.
|
Nathan Stone
Pittsburgh Supercomputing Center |
| Saturday, May 5 |
Thackeray Hall
Room 704 |
| 9:00 am - 10:00 am |
Epidemic spread in adaptive social networks: Effects of avoidance behavior.
Abstract
- Many infectious diseases spread along a network of person-to-person
social contacts. We consider the case of adaptive social networks, in
which the network structure changes adaptively as people adjust their
social contacts to avoid exposure, and the changes in network geometry
affect subsequent spreading dynamics. The form of adaptation most
frequently studied is avoidance rewiring, where susceptible nodes rewire
their connections away from infectives and toward other susceptibles.
Two new models are presented, showing effects of different forms of
network adaptation. In the first, not all individuals in the population
are aware of the need for self-protective behavior. We model
simultaneous spread of an epidemic and information about the epidemic.
The effects of adaptation, external information sources (e.g., media),
and node-to-node communication on the dynamics of epidemic and
information spreading are explored. In the second model, we study an
adaptation mechanism in which people temporarily deactivate social
contacts with infected neighbors but reactivate the connection once it
is safe. We study the interaction between the infection spread and the
geometry of the active subnetwork. For both models, we derive a mean
field system of equations to predict the epidemic dynamics.
| Leah Shaw
College of William and Mary |
| 10:00 am - 10:30 am |
Ensemble modeling of symptoms to human immune response of Influenza A virus infection.
Abstract
- Deterministic models of a host-level response to influenza A virus
(IAV) infection assume a perfect prediction, while an ensemble
approach may account for patient and strain variability, and
uncertainty in data used to calibrate the models. We generate an
ensemble of parameter sets that represent a calibration to
experimental data of viral titers and symptoms measured in humans with
IAV infection to a host-level model with innate and adaptive
immunities. Systemic, upper respiratory and lower respiratory symptoms
are mapped to model interferon levels, and extent of upper and lower
respiratory cells damage. In order to differentiate between upper and
lower symptoms, we compartmentalize the respiratory tract into upper
and lower compartments. We measure clinical factors such as onset
and severity of symptoms across our ensemble distribution and obtain
biologically relevant distributions while also achieving variability
in host responses. Sensitivity analysis across the parameter
ensembles is employed in order to characterize population-scale
relevant clinical phenotypes (severity of infection, immunogenicity)
to model kinetic parameters.
| Sarah Lukens
University of Pittsburgh |
| 10:30 am - 11:00 am |
Coffee Break |
|
| 11:00 am - 12:00 pm |
Modeling cholera dynamics in Haiti.
Abstract
- Cholera was introduced to Haiti in October 2010. As of March 2012, the resulting
epidemic has caused more than 530,000 cases and 7,000 deaths. I will describe some
ongoing efforts to model cholera dynamics in Haiti. This will include some
mathematical questions which naturally arise, including oscillations, transmission
delays, and waterborne disease dynamics on networks.
| Joseph Tien
The Ohio State University |
| 12:00 pm - 1:00 pm |
Epidemic dynamics on social networks.
Abstract
- In this talk I will give a brief overview of infectious disease dynamics
by using equation-based and social network models. I will discuss the relationship
between the prevalence of the disease and basic reproduction ratio and generation
time. Basic reproduction ratio is defined as the number of secondary cases generated
by a single infected individual when introduced into a totally susceptible
population. Another important parameter in epidemic dynamics is the generation time
which can be defined as the average time between the initially infected single
individual's infection time and its secondary infections. I will try to illustrate
this relationship not only by using equation-based and social network models but
also in our large-scale census-based agent-based simulation software, FRED.
| Hasan Guclu
University of Pittsburgh |
| 1:00 pm - 2:00 pm |
Catered Lunch |
|