A wide range of mnemonic and cognitive behaviors are dependent on persistent neural activity, sustained discharge in a neuronal population that far outlasts the duration of the stimulus. Persistent neural activity generally exhibits low-dimensional dynamics, meaning only a few patterns of firing across the population can be sustained, even when the population size numbers in the thousands. These low-dimensional dynamics are suggested by theoretical studies to be generated by attractor network mechanisms. However, in most settings there is an extremely large gap between abstract attractor network models and a quantitative understanding of the circuit mechanisms underlying persistent neural firing. Here we aim to close this gap by bridging between neuronal dynamics and circuit connectivity through a cross- scale approach using a combination of whole-network two-photon imaging, targeted optogenetic perturbations, whole-cell patch electrophysiology, serial-section electron microscopy, and next- generation network modeling. These studies will take advantage of the highly quantifiable saccadic and fixation behavior of the larval zebrafish, where persistent firing in the well-localized oculomotor neural integrator encodes a memory of desired eye position.
In Aim 1, we will combine imaging of neuronal dynamics with pseudo-random perturbations that explore the state-space of the integrator network to gain insight into the network's attractor landscape and the functional relationships between constitutive neurons.
In Aim 2, we will perform high- resolution serial-section electron microscopy on the same brain to gain insight into the circuit's anatomical connectivity.
In Aim 3, we will directly fit an attractor network model to these dynamical and structural data to find the expected physiological interactions in the network, thereby obtaining concrete predictions about the direct influence of one neuron upon another.
In Aim 4, we will test these predictions by obtaining whole-cell patch recordings of integrator neurons to measure synaptic responses following single-cell optogenetic stimulation. These tools and results will thus allow us to begin defining in a quantitative manner the patterns of physiological interactions in a circuit that give rise to persistent firing and temporal integration.
Persistent activity is a prevalent brain dynamics important for a range of behaviors from sensorimotor integration to decision making. In this proposal we will combine whole-network imaging, optogenetic perturbations, serial-section electron microscopy, next-generation network modeling, and whole-cell patch electrophysiology to determine the physiological interactions in a network generating persistent activity. This will allow us to identify the circuit mechanisms governing an essential brain dynamic.
|Vishwanathan, Ashwin; Daie, Kayvon; Ramirez, Alexandro D et al. (2017) Electron Microscopic Reconstruction of Functionally Identified Cells in a Neural Integrator. Curr Biol 27:2137-2147.e3|