notebooks/01-introduction.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# A Computational Approach to Vestibular Decision Making"
      ],
      "metadata": {
        "collapsed": false,
        "outputHidden": false,
        "inputHidden": false
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "We investigate the vestibular (perceptual) decision making. \n",
        "\n",
        "Two sub projects\n",
        "\n",
        "## Decision variable for perceptual decisions about self-motion\n",
        "Computational model decision making process. In particular, focus on the role of sensory \"evidence\": this is the output of a state inference model.\n",
        "\n",
        "We can conceptualize decision making either in the traditional DDM framework, or as a partially observable Markov decision process (POMDP). The latter approach is particularly interesting, as this framework is commonly used in robotics.\n",
        "\n",
        "A POMPD consists of a decision making agent, which acts on an environment, and in turn receives both observations of the state of the environment and a reward signal.\n",
        "\n",
        "State inference is typically performed by a filtering algorithm, e.g. particle filtering.\n",
        "\n",
        "Response time arising in a typical decision making can be viewed as the result of a POMDP given the task of categorising stimuli into either one of two possible categories. At each time step, the agent (decision maker) has to either decide or defer its decision and collect more information. This means that the agent has three permissible actions: decide __left__, decide __right__ and __wait__. Waiting means obtaining a further observation. In order to make a decision, the agent has to evaulate the __value of information__, that is the increase in expected reward with the new observation. There may also be a cost in obtaining new observations, e.g. if each time step incurs a cost.\n",
        "\n",
        "## Effect of prior beliefs on vestibular decision making\n",
        "\nReplication of Wertheim experiment."
      ],
      "metadata": {
        "collapsed": false,
        "outputHidden": false,
        "inputHidden": false
      }
    },
    {
      "cell_type": "markdown",
      "source": [],
      "metadata": {}
    }
  ],
  "metadata": {
    "kernel_info": {
      "name": "julia-1.0"
    },
    "language_info": {
      "file_extension": ".jl",
      "name": "julia",
      "mimetype": "application/julia",
      "version": "1.0.2"
    },
    "kernelspec": {
      "name": "julia-1.0",
      "language": "julia",
      "display_name": "Julia 1.0.2"
    },
    "nteract": {
      "version": "0.12.3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 4
}