What exactly is the neural receptive field?

Question

Neural receptive fields map the spatial or temporal distribution of the data to individual neuron excitation, if I understand correctly, but I do not understand if receptive fields (especially in the higher cognitive areas such as the various layers of the visual cortex, etc.) are:

1. Actual structural distributions of neurons that help map specific spatial or temporal properties of stimuli (such as concentric arrangement for contrast detection, rectangular arrangement for orientation detection, or movement detection, etc.), or,

2. Merely a mapping of the structure of the stimuli that can excite specific neurons.

If it's the latter, isn't it basically as simple as a pattern/feature/vector that individual neurons are sensitive to?

Answer 1

I think is closer to 2. It is the subset of sensory space that elicits a reaction in a given neural population.

So you should:

• Pick your neural population. Usually it is a single neuron (but can be easily extrapolated to bigger populations). It may be a sensory neuron, or a neuron in a higher processing region.

• Determine your sensory space. That is, you should decide the kind of stimuli you'll be dealing with. For example, not only "visual stimuli", but you should identify the parameters, or independent variables, you will vary in your experiment.

• Define a criterion to discriminate between reaction and non-reaction. This implies to choose what is the dependent variable in your study (for example neuron firing rate). Then you could define what features in the relation $\frac{dependent\:variables}{independent\:variables}$ you will search for. That is the tuning curve (which well may be a high dimensional curve).

• Perform your experiment and analyze your curve. The region of parameters (independent variables) whose reaction goes above threshold according to your criterion is the receptive field. Sometimes is not trivial to define such a criterion (see: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.0040092)

A good general reference is the article in Scholarpedia: http://www.scholarpedia.org/article/Receptive_field

This type of analysis is very useful for position coding, that is, when you expect that the activation of a specific region in the brain correlates to stimuli in a region of sensory space. But sometimes things are more complicated (e.g Olfactory coding http://www.jneurosci.org/content/30/27/9017.full).

Answer 2

No, it is not 1. the structural distribution of the neurons themselves. If by mapping you mean a kind of function that maps real space to cognitive space (or such) then it is not that either (at least, not explicitly/directly).

The receptive field is the region of 3D space in which sensory neurons can make their relevant detection. It's not a pattern/feature/vector, just the coordinates defining some spatial region itself.

For example, take the neurons making up your audio sensory system. If an object is making noise and you can hear it, then that object is within the receptive field of the hair cells responsible for the audible frequencies. That is, the object's physical location is within the spatial coordinates that define your audio system's perceptive field.

Fun fact: for two sounds of the same amplitude, lower frequency detecting hair cells tend to have a larger receptive field since lower frequency signals travel farther. In other words, the receptive field is somewhat dependent on the quality of the source of perception, not just the sensory neurons.

It's strange terminology coming from physics, because we generally define "fields" by sources, not by receivers. An audio source has an unambiguous "field" of acoustic radiation. A receiver does not generate a field but is sensitive to a certain threshold of field quantity. So to define the field as something coming from the receivers (the sensory neurons) is an unintuitive idea.

Answer 3

Stick a probe in your neuron-of-interest. Now stimulate the organisms in some way.

All the stimuli that make your neuron fire are the neuron's receptive field.

Example 1:

You probe a neuron in the primary sensory cortex of a human subject. Next, you stimulate the patient with light touch to their skin. Let's say you find that the neuron only fires when you stimulate the right thumb. Thus, light touch to the right thumb is that neuron's receptive field. The neuron doesn't fire to light touch to any other part of the body.

Example 2:

You probe a neuron in the primary auditory ganglion. Next you stimulate the patient with different sound frequencies. The neuron only fires at only a narrow band of sound frequencies. The receptive field of that neuron is that specific band of sound frequencies.