Overview
The brain uses information gathered from the body and the surrounding world to build internal representations and to control behaviour. Understanding how this is achieved at a mechanistic level is one of the most exciting and challenging problems in science today. Sensory information is processed as it flows from the periphery through various networks of neurons, and computations are thought to be performed at the synaptic, neuronal and network levels.
Our lab works on how synapses, single cells and neural networks carry out these computations. We are particularly interested in the contribution of synapses, which are thought to be one of the most important and elementary computational units in the brain. Synapses are not static, but highly dynamic, and exhibit a wide range of facilitating and depressing behavior over many timescales. The vast number of synaptic connections ( ~1014 ) and their activity-dependent properties are thought to enable the brain to process and store huge amounts of information. Our main goal is to identify new synaptic mechanisms and to understand how they contribute to information processing at higher levels. It is hoped that by studying the properties of brain function at various different levels, using a combination of experiments and modeling, we will be able to develop a mechanistic understanding that links molecular interactions to signal processing at the network level.
Preparations
We study synaptic mechanisms and signal processing in the following two brain regions:
Cerebellar Cortex
Much our work examining the mechanisms underlying synaptic transmission has been carried out on the mossy-fibre granule cell synapse in the cerebellum (Figure 1). We are also investigating the properties of synaptic integration in granule cells and the network behaviour of the granule cell layer (see figure at top of page).
 |
|
Figure 1
Two-photon image of Synaptophluorin labelled cerebellar mossy fibre terminals. |
Barrel Cortex
The barrel cortex is an attractive sensory system to work on because of the well defined one-to-one mapping of the whisker inputs to specific barrels. We are examining the properties of synaptic integration and network activity in layer 4 and layer 2/3 of the barrel cortex (Figure 3).
Figure 3. Reconstruction of layer 4 spiny stellate and layer 2/3 pyramidal cell in barrel cortex (Silver et al. 2003).
Techniques
We are currently developing a number of new optical and computational methods for investigating brain function at the synaptic, neuronal and network levels.
Optical Methods
We use a number of different types of laser-based microscopy to measure neuronal calcium, vesicular release (Figure 1), and network activity in acute brain slices. These include 2-photon microscopy, conventional confocal microscopy and high temporal resolution spot-confocal microscopy. We also use single-photon and two-photon photolysis methods to release biologically active compounds. We have recently used glutamate uncaging to investigate receptor properties at synapses and are presently using this technique to study synaptic integration in dendrites. We are currently working on a new type of 2-photon microscope that allows high speed data acquisition from multiple points in 3-D space (e.g. 30 points at 1 KHz/point). This will allow us to monitor the activity of many cells within a brain region on a time scale relevant for signal processing.
Electrophysiological Methods and Quantal Analysis
Whole-cell patch-clamp methods are widely used in the lab to record the electrical behaviour of single neurons in acute slices. We have also developed a quantal analysis method (multiple probability fluctuation analysis) that allows the functional properties of synapses to be quantified from synaptic responses (Silver 2003). Much of our data acquisition and analysis is carried out on a suit of programs ( NeuroMatic ) that we have developed in the IGOR PRO programming environment.
Computational Methods
We use a number of computational approaches. At the lowest level, we have developed a simulator, D3D, for modeling diffusion and reactions in arbitrary 3D space. This has been used to simulate spot confocal measurements of calcium in presynaptic terminals, glutamate diffusion in the synaptic cleft (Nielsen et al. 2004), and glutamate uncaging at the synapse. At the single cell level we have used integrate-and-fire and conductance-based compartmental models to investigate inhibition-mediated gain modulation (Mitchell and Silver 2003). While at the network level we have developed a new software tool, neuroConstruct, for constructing, visualizing and analyzing the behaviour of networks of neurons in 3D space. We are presently using this application to investigate information processing in the granule cell layer of the cerebellum.
Research supported by the Wellcome Trust, Medical Research Council, European Commission, Biotechnology and Biological Sciences Research Council and Engineering and Physical Sciences Research Council.
Software
Data Acquisition and Analysis
We have developed an acquistion and analysis software package for electrophysiology called NeuroMatic. This application is freely available, but runs within the IGOR Pro environment.
Neural Network Modeling in 3D
We are presently developing a software package called neuroConstruct, for the construction, visualization and analysis of biologically realistic neural networks in 3D-space. This application automatically generates scripts for running simulations within the NEURON and GENESIS simulation packages.
Diffusion Reaction Modeling in 3D
We are also developing a Java version of a 3D-diffusion reaction simulator called D3D. This software package will allow users to simulate diffusion within an arbitrary 3D geometry using an explicit finite difference method. We are currently using D3D to simulate calcium entrance/buffering in axonal boutons and glutamate uncaging within the synaptic cleft.
Recent Publications:
Rothman JS, Cathala L, Steuber V, Silver RA (2009)
Synaptic depression enables neuronal gain control.
Nature 457(7232):1015-8
link
Publications:
|
