We implant rats and mice with tetrodes. A tetrode is four 12.5µm wires twisted together to form a single probe.
    Each tetrode can be independently lowered into the brain, and we record from multiple brain areas simultaneously.
    We record voltage changes at the tips of all the wires continuously, with minimal filtering, as the animals perform behavioral tasks and during sleep. This typically requires measuring 3,000,000 samples every second, for hours at a time, while maintaining tight synchronization with behavioral events.
 
    For some experiments we use genetically modified mice. We are currently working with mice whose NR1 gene (which encodes an essential component of the NMDA receptor that is vital for certain forms of learning and memory) is flanked by small loxP sequences. This does not affect the animal during development, but when we subsequently inject the AAV-Cre virus into particular parts of the brain we can obtain a focal inactivation of the NMDA receptor, and then use tetrodes to record from cells in the affected region.
 
Behavioral Electrophysiology
21-tetrode rat implant
    To the right is an example of raw signals obtained from the four wires of one tetrode (in this case from mouse hippocampus). Each trace is 200ms in duration. Note the broad fluctuations that seem very similar between the four channels - this is the local field potential that reflects the synchronized activity of large populations of neurons and synapses. In addition, there are abrupt spikes - caused by the action potentials of individual neurons.
    To extract the spikes, we filter the signals off-line to keep just spike-frequencies, and apply thresholds (grey dashed lines). We take a snippet of data (1-2ms; orange boxes) surrounding each occasion that the signal crosses this threshold.
   A closer look at some of the threshold-crossing events (right) shows that not all spikes look the same.
6-tetrode mouse implant
    In particular, they vary in relative size on each channel (compare the first two spikes to the third spike). We make use of the peak voltage on each wire (and other properties) to assign each spike to a corresponding neuron (this is called spike-sorting). In the plot below right, each point represents one spike; the spikes form distinct clusters that can be marked with different colors (click on the plot for video). This is because each neuron generating spikes has a distinct physical location relative to the tips of the four tetrode wires, and so its spikes consistently appear larger on some wires than others (the spikes for each cluster are shown below left). While not perfect, the use of tetrodes greatly improves our ability to distinguish between different neurons. For example, it would not be possible to tell which spikes belong to the “yellow” neuron and which to the “green” neuron if one only had the first of the four wires.
Eight clusters isolated from one tetrode in rat striatum
    The average shape of the spike waveforms provides important information about the responsible neurons. For example, in many brain areas “fast-spiking interneurons” generate frequent, short-duration spikes (right, top row), while projection neurons have longer-duration spikes (right, 2nd row).
    The ability to distinguish between different types of neurons allows us to learn about how local microcircuits within brain structures operate. For example, fast-spiking interneurons have a key role in controlling the precise timing of spikes within the larger population of projection neurons, and this appears to be important for the striatum’s role in action selection.
Distinct neuronal waveform shapes in rat striatum
    We use many different analysis techniques to examine the relationships between neural activity and behavior. Below we describe just one step in the process - identifying when individual neurons are active.