The overall goal of the department is to understand cortical computation at the level of individual spiking neurons with an emphasis on the role of precise spike timing in neural coding and the role of neural coherence in attentional modulation. To achieve this goal, two separate research tracks are pursued.
First, models are constructed to connect data obtained from in vivo experiments at the single neuron level with those obtained at the systems level using fMRI, EEG or MEG techniques in order to generate predictions for local circuit dynamics. The guiding hypothesis is that inhibitory neurons play a pivotal role in generating coherent neural activity and can thereby selectively modulate the flow of information. New genetic approaches in combination with multi-electrode recordings or two-photon imaging make it possible to identify specific classes of interneurons and record their activity, thus allowing for an explicit test of model predictions. At the single neuron level we utilize models at various levels of detail ranging from leaky integrate-and-fire neurons to complex multi-compartment models with Hodgkin-Huxley type channels, whereas network models are studied ranging in scale from local circuits of a few hundreds of neurons to large-scale models of visual cortical areas with a few hundred thousands of neurons.
Second, data analysis techniques are developed in order to analyze multi-neuron spike trains recorded from multi-electrode arrays and imaged using two-photon microscopy and compare them to the outcome of model simulations. To this purpose, the latest state-of-the-art machine learning techniques are adapted for use in neuroscience. Both research tracks take place within the context of experimental research conducted within the Donders Center and also involve collaborations with various national and international research groups.