Neurophysiology of behaviour

Lead Research Organisation: The Francis Crick Institute


We are trying to understand how the brain processes information. To this end, we develop technology to record from large numbers of neurons, the cellular building blocks of the brain. This technology will then not only be used in our lab but could provide the basis for new generations of neuroprosthetics, devices that might be able to help to restore for example the ability to see, hear, touch, walk in patients with a range of accident- or disease-related disabilities.

In our efforts to understand how information is processed in the brain we use the mouse sense of smell as a model system. In this somewhat reduced system we combine behavioural analysis with electrical and optical recordings to assess how the neurons in the smell regions of the brain represent information. We then analyse the structure of these brain regions using electron microscopes and combine these structural and functional information to build accurate models. We test these models using sophisticated “optogenetic” modifications that allow us to implement “light switches” in individual cells, allowing us to switch those cells on and off during electrical recording and behaviour, probing the predictions of the models and further improve them.

Together this will give us a mechanistic understanding how these brain regions processes information – a basis not only for understanding how the brain in general computes but also what could go wrong in debilitating neurological diseases.

Technical Summary

This work was supported by the Francis Crick Institute which receives its core funding from the UK Medical Research Council (FC001000), the Wellcome Trust (FC001000),and Cancer Research UK (FC001000)

Understanding how complex behaviour emerges from the properties of molecules, cells and ensembles of cells is one of the key challenges in neuroscience. The first part of the Neurophysiology of Behaviour laboratory tackles this question employing the olfactory system of mice as a model system.
To understand how smells are processed we modify specific selected brain areas - in particular the olfactory bulb - using transgenic mice, pharmacological tools or targeted virus injections. We then probe how these specific modifications alter the neural networks and the resulting cellular function and physiology in vivo during behavioural tasks using whole-cell patch-clamp recordings, 2-photon microscopy and unit recordings. Ultimately, we perform quantitative behavioural tasks in such modified mice. We combine volume electron microscopic analysis of the olfactory bulb (using serial block-face scanning electron microscopy of columes of up to 0.2 mm^3) with functional analysis in vivo in order to delineate the underlying anatomical substrate and provide a mechanistical basis for our understanding of physiological responses.
Combining these genetic, physiological, anatomical and behavioural techniques with computational modelling approaches we aim to elucidate the cellular basis of olfactory behaviour and ultimately more general complex behaviours.
A second part of the Neurophysiology of Behaviour Laboratory aims to develop tools to enable large-scale electrophysiological recordings to benefit both basic neuroscience research and neuroprosthetics. Current electrical recording technologies in neuroscience are held back by the difficulty of fabrication. An electrode is connected to a plug, which is connected to a pre-amplifier, which is connected to digitization equipment. These components are fabricated separately and then assembled by hand, limiting their minimum size. For that reason, the number of simultaneously recorded neurons have in the past not scaled with Moore’s law (doubling every 2 years), as it arguably should, but far slower (doubling every 7 years). While there are some systems under development aimed at locally multiplexing of neural signals, they are still generally based on specialized hardware unlikely to scale efficiently.
To overcome this, we develop technology around easily tailored bundles of glass-sheathed metal microwires. These mechanically connect to the surface of a camera chip readout integrated circuit (ROIC). By harnessing the already highly-developed microscale electronic readouts and amplifiers for each pixel in the ROIC, we aim to record from up to 106 individual wire elements. These pixels will be mated to glass-sheathed gold microwires 5-30 µm in diameter which can be inserted into the brain and individually addressed. This part of the lab strongly benefits from collaborations with material scientists, engineers, and computational neuroscientists, enabling us to develop all necessary components that will make it possible to obtain unprecedented amounts of data about neural activities in vivo.


10 25 50