Representing and Transforming Sensory Stimuli using Spike Trains

Abstract

This final report summarizes the results from the four (3+1 NCE) years of support under the grant. Technical advances were made in three interrelated research areas, with novel optimization frameworks conferring the common theme.(1) An optimization framework that tunes the dynamics of a network of non-spiking neurons to display an observed behavior, ws developed. The framework was then used to identify synaptic profiles in a parsimonious network model of the elementary motion detector. To our knowledge, this is the first conductance based compartmental model neuronal network implementation replicating the behavior of the T4 neuron in the fly optic lobe. (2) A framework was developed for the coding and decoding of continuous time signals using an ensemble of spike trains. The technique distinguishes itself in the quality of reconstruction achieved under low spike rate regimes. (3) A framework named Spike-triggered descent was developed to complement the widely used technique, Spike-triggered average, to characterize the response of a neuron to sensory stimuli. The framework improves upon the model assumption from spikes generated using an inhomogeneous Poisson process to spikes generated using a cumulative spike response model. Superior performance was demonstrated on a Locusta migratoria tympanal nerve dataset. The technique has wide applicability to all neural systems that display low levels of noise.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 19, 2020
Accession Number
AD1104310

Entities

People

  • Arunava Banerjee

Organizations

  • University of Florida

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Coding
  • Computational Neuroscience
  • Computational Science
  • Data Sets
  • Detection
  • Detectors
  • Equations
  • Information Processing
  • Information Science
  • Neural Networks
  • Probabilistic Models
  • Scientific Research
  • Signal Processing
  • Students
  • Universities

Readers

  • Computational Modeling and Simulation
  • Neuroscience