Sensory Systems/Computer Models

From Wikibooks, open books for an open world
Jump to navigation Jump to search
Technological Aspects
In Animals

TOC[edit | edit source]

Short Description of Chapter Contents[edit | edit source]

Simulations of Neurons[edit | edit source]

The chapter on neural simulations presents simulations on on different levels:

  • The behavior of the neural membrane can be described exceedingly well with the Hodgkins-Huxley equations, which can be simulated with dynamical systems tools.
  • To simplify the dynamics of the membrane, the Hodgkins-Huxley equations can be simplified to a 2-parameter system, described by the Fitzhugh-Nagumo model.
  • Idealized effects of feedback on the membrane potiential, and simple neural networks, can be simulated with simple feedback systems.
  • Neuromorphic systems, which implement neural circuits with analogVLSI, allow the systematic investigation of realistic larger neural networks, and the application of such systems for ultra-low energy and/or very fast responding applications.

Simulations of the Visual System[edit | edit source]

The chapter on simulations of the visual system primarily describes how linear 2d-convolutions can capture many aspects of information processing in the visual system:

  • The effect of the cornea
  • The contrast detection in the retina
  • The edge detection in V1, the primary visual cortex.
  • A short summary is also given about digital image processing.

The chapter on simulations of the retinal function provides a software simulation of human retina based on a biological inspired model. It shows how retina, receives, transmits and processes visual information for higher visual cortical structures.

If you want to know more about visual perception, the chapter on the Physiology of the Visual System has special chapters on:

  • Motion perception
  • Color perception

Simulations of the Auditory System[edit | edit source]

The chapter on simulations of the Auditory System covers a fairly large area:

  • Since auditory signals are often described in the time domain, this chapter first presents a brief overview over the Fourier Transform.
  • This section also describes the effect of narrowing the time window (to improve temporal accuracy) on the observed frequency spectrum.
  • The next section shows how control system models can reproduce the effects of the pinna and the outer ear on the auditory signal.
  • The deflection of the basilar membrane by auditory inputs is very elegantly and efficiently described with Gammatone filters.
  • The section on human speech shows computational approaches to speech perception.

Simulation of the Vestibular System[edit | edit source]

The Simulation of the vestibular system has two components, which have to modeled separately:

  • The semicircular canals transduce angular velocity, and can be approximated well with 3 sensitivity vectors. As a result, they can be modeled nicely with control system models.
  • The otoliths transduce linear acceleration (including gravity). Their response has to be simulated with more involved finite element simulations.
  • The downstream processing of the signal by the brainstem can again be approximated well with control system tools.

Simulations of the Somatosensory System[edit | edit source]

The chapter on simulations of the somatosensory system describes how muscle spindles can be simulated. Complete simulations of the somatosensory system are difficult, since this system includes not only the limb dynamics (which are complex on their own), but also additional complexities caused by the redundancy of muscles for limb movements.

Efficient coding[edit | edit source]

The chapter on efficient coding describes how natural images and natural sounds are encoded by the brain and how the efficient coding model replicates this process. It was found that the process of both input signals could be modeled with very similar methods. The goal of efficient coding theory is to conceil a maximal amount of information about a stimulus by using a set of statistically independent characteristics.

Computational Models of Olfaction[edit | edit source]

Our understanding of the neural processes underlying hearing, seeing, and orientation sensing are remarkably advanced. On the gadget side, we have corresponding vision, sound, and movement sensors available for robotics applications. The processes of smell and taste, however, seem to be remarkably more complex. This complexity is hinted at by the sheer number of physiological sensors available to our respective senses: we have one to two types of transducing cells for hearing and for movement sensing (regular and irregular hair-cells), and about four types of light transducing cells in the retina (the three types of cones for color vision, and the rods for light/dark sensation). Instead, we have hundreds of taste- and smell- receptors in our nose and on our tongue. It makes sense, therefore, that despite being functional, current artificial noses and tongues are not yet very advanced. The chapter on simulations of the olfactory system describes our first attempts at computationally describing the olfactory system and its properties - that is to say, how we come to smell.

Neurosensory Implants · Arthropods