Cybernetics of thalamic and cortical networks
Dir. Thierry Bal

Our research is centered on the cellular and network mechanisms by which thalamic networks and cortico-thalamic interactions gate the processing of sensory information depending on the state of arousal. We are using standard electrophysiological techniques and hybrid networks interfacing neural cells to computers to explore neuronal circuits that produce organized population activity in vitro. These combined techniques and novel approaches using voltage sensitive dye imaging and magnetic recording/stimulating techniques are being developed in collaboration with other teams in the lab and through external collaborations.

Recent results:


In collaboration with G. Le Masson (INSERM, Bordeaux), we have constructed "hybrid" thalamic circuits by connecting biological thalamocortical cells to silicon- or software-generated nRt neurons and retinal inputs. Using such hybrid circuits, we have shown that feedback inhibition from nRt to TC neurons can be considered as a variable gain control mechanism able to modulate the efficiency of sensory signal transfer through the thalamus (Le Masson et al., Nature 2002). These results have implications for pathological disorders such as absence epilepsy (petit mal). Using a thalamic slice preparation that spontaneously generates spindle waves (10 Hz sleep-like synchronized oscillations), we studied the synchronizing effect of corticothalamic inputs on thalamic cell population activity. We found that corticothalamic inputs can force a temporal redistribution of thalamic cell activity, mediated through GABAb receptor activation, leading to a 3 Hz slower and hypersynchronized oscillation resembling absence seizure episodes (Bal et al., J. Neuroscience 2000).

During waking and attentiveness, it is known that cortical and thalamic neurons are under constant bombardment from synaptic input. The role of this synaptic bombardment has recently gained renewed interest. In spontaneously active cortical network in vitro (Shu et al., J. Neurosci. 2003) and more recently in thalamic neurons, by mimicking the activity of thousands of synapses using dynamic clamp, we have shown that background synaptic "noise" influences the integrative and electrophysiological properties of neurons, controlling important parameters such as the gain and the sensitivity of their input-output function. In the thalamus in particular, there is a mixture of single-spike and burst responses at all membrane potentials, leading to probabilistic and linear encoding of inputs (Wolfart et al., Nature Neuroscience 2005).


Previous work in invertebrates and vertebrates has shown that the understanding of the responsiveness of central neurons requires a detailed knowledge of their intrinsic properties, which are mediated by various ionic- and voltage-dependent conductances (Llinas, Science 242 :1654, 1988). Our results and those obtained in vivo (Fregnac’s team) and in computo (Destexhe’s team) suggest that background activity alters this responsiveness fundamentally. We therefore suggest that a complete characterization of the properties of central neurons requires the knowledge not only of intrinsic but also of synaptic background conductances, as well as the amount and statistical structure of conductance fluctuations ("background or ongoing noise").


Interactions between multidisciplinary teams of the UNIC lab (electrophysiology, computer science, modeling) and external collaborations (G. Le Masson, INSERM Bordeaux France, R. Brette, INRIA Paris France, José Gomez, Canary islands Spain ; Leonel Gomez-Sena, Montevideo Uruguay) has fostered the development of neuron-machine interface and AEC, a novel recording approach based upon Active Compensation of the Electrode (AEC; R. Brette, Z. Piwkovska, M. Rudolph-Lilith, T. Bal and A. Destexhe).

Our most recent development is therefore based on the dual use of dynamic clamp and AEC. Dynamic clamp using RT-NEURON (Designed by G. Le Masson INSERM Bordeaux, G. Sadoc UNIC), our real time version of the NEURON simulator (Hines and Carneval) allows to insert artificial conductances in the membrane of a biological neuron via current injections calculated by the computer. These current injections mimic the electrical activity of ionic channels. However, injection of high-frequency currents through a sharp microelectrode or through a high impedance patch electrode (such as those used in vivo) is known to produce signal distortions in the recording. The AEC eliminates such distortions by subtracting in real time a model of the electrode contribution from the voltage recording (Brette, Piwkowska et al., submitted). The dual use of AEC and RT-NEURON allows high temporal resolution dynamic clamp in vivo.

Selected Publications

Gilles Ouanounou, Baux Gérard and Thierry Bal, A novel synaptic plasticity rule explains homeostasis of neuromuscular transmission, eLife 5: , (2016) [pdf] [abstract]

Sébastien Béhuret, Charlotte Deleuze and Thierry Bal, Corticothalamic Synaptic Noise as a Mechanism for Selective Attention in Thalamic Neurons, Frontiers in Neural Circuits 9: 11633, (2015) [pdf] [abstract]

Amanda Casale, Amanda Foust, Thierry Bal and David A. McCormick, Cortical Interneuron Subtypes Vary in Their Axonal Action Potential Properties, Journal of Neuroscience 35: 15555, (2015) [abstract] [PubMed]

Jakob Wolfart, Damien Debay, Gwendal Le Masson, Alain Destexhe and Thierry Bal, Synaptic background activity controls spike transfer from thalamus to cortex, Nat Neurosci 8: 1760-7, (2005) [pdf] [abstract]

Gwendal Le Masson, Sylvie Renaud-Le Masson, Damien Debay and Thierry Bal, Feedback inhibition controls spike transfer in hybrid thalamic circuits, Nature 417: 854-8, (2002) [pdf] [abstract]