Professor Stephen Williams

Prof. RF in Learning and Memory

Queensland Brain Institute
+61 7 334 66352


Professor Stephen Williams joined the Queensland Brain Institute in 2010 to establish a laboratory studying synaptic integration in neuronal networks. He moved to QBI from the MRC Laboratory of Molecular Biology, Cambridge, UK where he was a tenured programme leader.

Research Interests

  • Synaptic Integration
    Synaptic Integration: A single neuron in the central nervous system may receive thousands of synaptic inputs distributed widely across its dendritic arbor. A fundamental operation of neurons is the integration of such time varying input signals to form an output signal, termed the action potential, which is communicated to other neurons and/or effector systems such as muscles. We investigate key questions of single neuron computation by generating diverse spatial and temporal patterns of synaptic inputs in central neurons maintained in vitro. We use advanced electrophysiological and optical techniques that allow simultaneous recording from multiple dendritic sites of a single neuron, the uncaging of neurotransmitters at identified dendritic sites and the optical activation of neuronal pathways expressing light-activated channels to explore synaptic integration strategies in neurons of diverse morphology.
  • Network dynamics
    Network dynamics: Neocortical pyramidal neurons are embedded in active neuronal networks that fire spontaneous and stimulus-evoked patterns of action potentials in the working brain. How effectively are physiological firing patterns transmitted through excitatory synaptic connections in the neocortex? We have shown that the synapses of the output neurons of the neocortex code information on an action potential-by-action potential basis. These synapses showed increased neurotransmitter release in response to short high frequency trains of action potentials, suggesting that bursts of action potentials are an important unit of information in the neocortex. Recently, we have precisely recreated action potential firing patterns recorded from neocortical neurons in vivo, to investigate how physiological patterns of activity are transmitted through different classes of excitatory synaptic contacts in vitro. We find that excitatory synaptic transmission between neocortical pyramidal neurons in response to physiological firing patterns is pathway-specific. Layer 2/3 to layer 5 responses were dominated by synaptic depression, transmitting only the onset of physiological firing patterns. In contrast, layer 5 to layer 5 excitatory synapses faithfully transmitted each action potential of the physiological train. We suggest that the intra-cortical synaptic output of layer 5 pyramidal neurons is ideally suited to dynamically control the cortical network on an action potential-by-action potential basis across a wide range of frequencies and for sustained periods of time, providing a reliable internal representation of neocortical output.
  • Single neuron computation
    Single neuron computation: Multi-site whole-cell recordings from the dendrites of principal neurons of the rodent neocortex, thalamus, substantia nigra and cerebellum have allowed insight into the role of dendritic mechanisms in neuronal computation. For example, in cortical pyramidal neurons, the most abundant neuronal class in the neocortex, work has shown that synaptic conductance is uniform for excitatory synapses positioned throughout the apical dendritic tree. Excitatory postsynaptic potentials (EPSPs) generated from distal dendritic sites therefore have a negligible impact at the level of the soma, a site close to the axonal site of action potential initiation. We find, however, that distal dendritic EPSPs can be locally integrated in the dendritic tree, to trigger powerful dendritic spikes that robustly forward propagate through the dendritic arbor to drive neuronal output. This form of compartmentalised synaptic integration was found to be robust under conditions that mimic active states in the working brain. Moreover, we find that under conditions of ongoing action potential firing, the recruitment of dendritic sodium and calcium voltage-activated channels greatly amplify the impact of barrages of distal dendritic EPSPs, providing a conditional mechanism for the normalization of synaptic efficacy. Taken together, these findings reveal that dendritic mechanisms extend the computational repertoire of single neurons.
  • Ion channel targeting
    Ion channel targeting: We have identified that classes of voltage-activated ion channels are non-uniformly distributed through the dendritic tree of neurons and that specific somato-dendritic expression patterns of ion channels are a signature of neurons from different brain regions. For example, we were the first to describe the highly polarized apical dendritic distribution of Hyperpolarization-activated Cyclic Nucleotide gated (HCN) channels in neocortical pyramidal neurons and show the important function of this channel type in controlling the spatio-temporal integration of dendritically generated synaptic potentials.


  • Doctor of Philosophy, University of Wales
  • Bachelor of Science, University of Wales


View all Publications


  • Doctor Philosophy

  • Doctor Philosophy

  • Doctor Philosophy

View all Supervision


Book Chapter

  • Williams, Stephen R. (2010). Dendritic dynamic clamp – a tool to study single neuron computation. In Alain Destexhe and Thierry Bal (Ed.), Dynamic-clamp : From principles to applications (pp. 31-48) New York , NY, USA: Springer.

  • Etherington, Sarah J., Atkinson, Susan E., Stuart, Greg J. and Williams, Stephen R. (2010). Synaptic integration. In Encyclopedia of life sciences (pp. 1-12) Chichester, England, U.K.: Wiley Library. doi:10.1002/9780470015902.a0000208.pub2

Journal Article

Grants (Administered at UQ)

PhD and MPhil Supervision

Current Supervision

  • Doctor Philosophy — Principal Advisor

  • Doctor Philosophy — Principal Advisor

  • Doctor Philosophy — Associate Advisor

Completed Supervision