Alex D. Reyes - US grants

DESCRIPTION (provided by applicant): The primary goal of this project is to determine how acoustic signals are represented, processed, and transmitted in cortex. Signal processing depends on several variables including: 1) the membrane properties of neurons; 2) the patterns of anatomical connections between the many types of neurons in cortex; and 3) the dynamic properties of synaptic transmission. An in vitro slice preparation of the mouse auditory cortex will be used to examine how these variables affect signal processing in a small network of neurons. Simultaneous whole-cell recordings from 2-5 interconnected neurons, each of which can be identified visually. The experiments of Aim 1 will determine the patterns of connections between the pyramidal neurons and the different types of excitatory and inhibitory neurons in cortex and characterize the properties of synaptic transmission between each neuron. The experiments of Aim 2 will determine which of these cells are innervated by thalamic afferents, a major source of input into the auditory cortex. At the completion of these two sets of experiments, the circuitry that process signals will be determined. The link between the input neurons (those that receive thalamic afferents) and the output neurons (pyramidal neurons) will be established. The experiments in Aim 3 will then examine, using a novel iterative procedure, how computergenerated signals delivered to the thalamus affects the firing of successive neurons in circuit. These experiments will provide both basic information about circuitry as well as provide insights as to how auditory signals are processed in cortex. This information may elucidate mechanisms of deafness as well as lead to better designs of cochlear implants.

DESCRIPTION (provided by applicant): Dendrites receive the majority of synaptic inputs and therefore play a central role in integrating incoming information. The transfer of excitatory post-synaptic potentials along the dendritic tree represents an excellent opportunity to transform this information before it reaches the axon where action potentials are initiated. Due to the difficulty of recording from the dendrites, our understanding of these signal transformations remains rudimentary. Thus, we rely on computational models of dendrites for predictions and insights into their biophysical mechanisms and functional ramifications. In this project, we will develop a novel electrophysiological technique, Dendritic Replacement Dynamic Clamp (DRDclp), that combines somatic dynamic clamp with computer modeling to replace biological dendrites with virtual ones. To perform DRDclp: 1). native dendrites will first be uncoupled from the soma to minimize their electrical impact on somatic current injection; 2). virtual dendrites will then be attached by injecting the axial current between a real-time simulated cable model and the soma via dynamic clamp. As the native spike generating mechanisms will remain functionally intact, synaptic integration will be studied by observing the effects of varying dendritic properties on neuronal output. Two broad classes of virtual dendrite will used: Duplicate virtual dendrites will be parameter fitted so that their properties match those of the native dendrites they replace; alternative virtual dendrites will be based on dendrite models for the neuronal class in question however their properties will be judiciously chosen to test some hypotheses. In Aim 1, we will use physical uncoupling with alternative virtual dendrites to probe an example problem of synaptic integration in two model systems: gerbil medial superior olive (MSO) neurons and rat cortical pyramidal neurons. By using two neuronal classes that possess different dendritic and electrical properties, we plan to demonstrate the universality of the DRDclp technique. In Aim 2, we not only propose to use DRDclp as a means of attaching duplicate virtual dendrites to gerbil MSO neurons, but also to perform electrical uncoupling by cancelling the current entering or leaving native dendrites. Real-time parameter fitting of our dendrite model will be necessary to implement electrical cancellation and virtual dendritic attachment successfully in this aim. Experimental data generated by targeting synaptic stimulation and channel block along individual dendrites will be used to facilitate this calibration. Simultaneous recordings from dendrites and soma will also help to confirm the success of electrical cancellation. We expect DRDclp can be used to enrich our understanding of dendritic computations while also making sense of the variability in dendritic properties in neuronal populations.

Detailed description of the proposed use of the animals, including species, strains, ages, sex, and number to be used; Dissociated, primary cultures will be prepared from the cortex of new born mice of either sex (mus musculus, Postnatal day 0-1). These experiments will be performed using pups obtained from Vglut1-IRES2-Cre strains mated with Floxopatch (Lou et al., 2016) strains so that pyramidal cells will express channel rhodopsin CheRiff and voltage indicator QuasR2. Up to 5 cultures can be grown, which can be used for 5 experiments. Justification for the use of animals, choice of species, and numbers to be used; Description of procedures for minimizing discomfort, distress, pain, and injury; and Method of euthanasia and the reasons for its selection. Transgenic mice (category C) will be used because the experiments rely on the expression of CheRiff and QuasR2 for optical stimulation/recording. There are no alternative species. All experiments are terminal. Newborn mice will be cooled in ice prior to cervical dislocation and excision of the brain for culture preparation. All procedures are in accordance with guideleines of the NYU Animal Welfare Committee. The estimated number of animals below ensure that we have a continuous supply of cultures to perform experiments on a daily basis. Based on our experience in the past 3 years, each postnatal day 0-1 (P0-P1) mouse will generate at least 5 culture preparations. A viable culture will have a have a high density of neurons that express the Floxopatch construct, which varies from mouse to mouse. Thus, if one culture is not viable, all cultures derived from that mouse are also unusable. The success rate is approximately 50%. Because cultures can be made only form P0-1 day old mice and because it is not possible to determine whether expression is sufficient until the day of the experiment 2-3 weeks later, we will need to make 2 sets of cultures per week. Thus, we will need to maintain 2 breeding pairs per week or 8 breeding pairs per month to have a continuous supply of cultures. The breeding pairs will either consist of wildtype mice (for viral injection) or Vglut1-IRES2-Cre - Gt(ROSA)26Sor (floxopatch) mice. Breeding pairs will be replaced every six months. Total animals for experiments: 32 mice/year (=16 mice (8 breeding pairs) x 2/year replacement of breeding pairs). To maintain the 3 lines, 2 males and 2 females each of Vglut1-IRES2-Cre, Gt(ROSA)26Sor, and wildtype mice will be kept in separate cages (4+4+4=12 mice total). Every 6 months (twice per year), the mice of each type will be bred. From their offsprings, 4 males and 4 females from each line will be weaned and kept in separated cages as above (4+ 4 + 4 = 12) and the 4 original breeding pairs (12 mice) euthanized. Overall, 24 mice/year will be used to maintain the lines. Total mice/year: 56 (=24 mice/year to maintain lines + 32 mice/year for experiment). Total mice over 3 years = 168 mice. Reference Lou et al., (2016) Genetically Targeted All-Optical Electrophysiology with a Transgenic Cre-Dependent Optopatch Mouse, J. Neurosci. 36:11059-73 1 53 Results from previous support from CRCNS The P.I. and co-P.I. have not received previous CRCNS support Coordination Plan Division of labor The P.I. (A. Reyes) and his group at NYU will perform the experiments and analyses detailed in Aims 1 and 2 of this proposal. The co-P.I (S. Fusi) and his group at Columbia University will be responsible for the development of the theory and computational models that will incorporate the experimental data. Data Sharing between NYU and Columbia groups: Analyzed data, model codes, and manuscripts in progress will be disseminated between the groups via Google drive. All members of the Fusi and Reyes groups will have access upon email request. Larger, raw data will be stored on local servers and accessible via FTP. Coordination between NYU and Columbia University Because NYU and Columbia are easily acces- sible by subway, we will have monthly joint lab meetings, alternating between NYU and Columbia. Graduate students and postdoctoral fellows will be encouraged to submit joint poster presentations or talks to annual meetings such as Cosyne and Society for Neuroscience Meetings. 16 54