Developing imaging of action potential times in arrays of cerebral cortical neurons expressing transgenic calcium-sensing fluorescent proteins.

The ultimate aim of the project is to develop an imaging technique for accurately recording the times of nerve impulses ('spikes') in most of the neurons (brain cells) in a 3-dimensional volume of cerebral cortex in an awake, behaving animal. (The cerebral cortex is the biggest part of our brains, which does most of the thinking and perceiving). The point of this is to discover what kind of neural code the brain actually uses - do the exact times of spikes matter? We will establish the feasibility of doing this by means of monitoring flashes of light along branches of neurons called apical dendrites, which rise most of the way towards the surface of the brain. Whenever a brain cell fires a nerve impulse, this also travels backwards along part of the apical dendrite. Calcium enters the dendrite during the nerve impulse. In specially genetically-engineered mice this can briefly make an artificial calcium-sensing protein in the dendrite fluoresce more brightly (flash) when illuminated by strong, fast pulses of infra-red laser light focused down a microscope. This '2-photon' microscope can also be used to build up beautiful 3-dimensional pictures of brain cells in the grey matter of the cerebral cortex,despite it being non-transparent. One of the biggest technological hurdles in developing a technique to record spike times from all neurons at once in a chunk of grey matter, is the difficulty of focusing a microscope up and down through the thickness of the grey matter rapidly enough. We propose to test whether we can use the spike-related calcium flashes along the apical dendrites to get around this problem: if we can record spike times from dendrites of all the neurons we are interested in at a single depth in the cortex, there is no need for rapid up and down focusing. In practice, we will probably need to repeat the process at 3 or 4 different depths, since the flashes typically don't run along the full length of a given apical dendrite. But we can 'splice' together spike times from different depths to get a picture of the total network activity in that volume of brain for a given stimulus, repeated carefully while focusing at the different depths. If we do this for enough stimuli, we should be able to understand whether the brain uses the precise times of spikes across large networks of neurons to encode important information.

Does the brain use an ensemble spike time code? Our ultimate aim is to develop a technique to simultaneously image action potential (AP) times to millisecond resolution in a high fraction of neurons in a 3-dimensional module of cerebral cortex such as a barrel (whisker) column, in a behaving animal, using a 2-photon microscope. Methods for rapid focusing are being developed, but it is not clear if these will be able to perform a 3-D scan through the entire volume of a cortical module in 1-2 ms with enough signal-to-noise to pick out spike times. We propose to exploit 3 fortuitous features of neocortical anatomy/physiology to circumvent the need for rapid focusing: each pyramidal cell body sends a relatively thick apical dendritic trunk towards the cortical surface; APs back-propagate some way along the trunk; each AP is associated with a fast-rising Ca2+ transient over a significant portion of the proximal trunk. Using line scans, AP times can be determined from these transients to millisecond accuracy, in brain slices. A new strain of mouse has become available, expressing a transgenic fluorescent Ca2+ sensing protein cerTN-L15 in most cortical neurons, including their apical trunks. This sensor is fast and sensitive enough for detecting AP pairs and often single APs, in cell bodies, which have smaller Ca2+ signals than proximal apical trunks. We will test, using whole-cell recording and 2-photon imaging in brain slices, if single AP times can be imaged in apical trunks of these mice. We will repeat in lightly anaesthetised mice, calibrating with electrical recordings of APs. For cells of different layers, we will establish what fraction of the apical trunk allows AP time imaging, with motion correction. We will find how many optical sections are needed to image APs from all pyramidal cells in a cortical module in overlapping subsets, using repeat trials with same stimulus time-locked to the EEG, splicing subsets together to get the full network spiking pattern.