In previous blogs we have discussed the proliferation of low cost sensors and their potential positive impact on our environment. The cost of the sensors and the design of the sensor system will determine the density of sensors. The density of smart sensors in our environment the lower the cost of data transport, the lower the energy demands, the greater the redundancy, the better the reliability, the less the need for replacement.

For this article we refer to a sensor’s output as a channel. We do this so that we can create virtual channels which can carry user created data or data created from combining multiple channels. Both channels and virtual channels can emulate the human nervous system using a publish subscribe low latency messaging network combined in the eight models of inhibitory and excitatory feedback mechanisms used by the human nervous system. This blog focuses on how to emulate what is called the lateral inhibitory circuit. A lateral inhibitory circuit is excellent at what is called edge detection. Edge detection can be used to differentiate between the uniformity of two fields, any fields, multiple auditory fields, multiple visual fields, multiple pressure fields.


It is important to note that nerves transmit all the sensory data collected by our bodies in the exact same way, by sending a signal, an action potential, along the axon of the nerve to other nerves. The frequency of the signal determines the intensity of the stimulus. This mechanism can be considered to be normalized with every nerve impulse being determined by:

  • type of nerve
  • location of stimulus or nerve endpoint
  • connectivity (which network the nerve belongs to)
  • frequency of pulse

Here is a diagram of a lateral inhibitory circuit.

Diagram from Introduction to Neurons and Neuronal Networks by John H. Byrne, Ph.D.

Our eyes use this type of circuit to enhance the edges within our visually recognized environment. “Edge enhancement is mediated, at least in part by lateral inhibition in the retina.” quoted by John H. Byrne. Here are two diagrams from Doctor Byrne one showing a circuit without lateral inhibition and one using lateral inhibition.

Optic Nerve Representation Without Lateral Inhibition

Optic Nerve Representation With Lateral Inhibition

Diagrams from Introduction to Neurons and Neuronal Networks by John H. Byrne, Ph.D.

In fact, here is how your brain recognizes the two shaded areas.

Without Lateral Inhibition

With Lateral Inhibition

Publish/Subscribe IoT Neural Network Mimicry

How can we mimic edge detection using a low cost sensory network? We can accomplish this with the creative use of virtual channels and a publish subscribe mechanism built into a sensor and smart edge node devices. Here is an image of an array of sensors connected to a smart node that has virtual channels.

HealthInRealTime Inc. Lateral Inhibitory Neural Network Publish/Subscribe Design

In the above diagram sensors are feeding Virtual HealthInRealTime Messaging Devices each having the capability of publishing and subscribing to other virtual devices. Publishing and subscribing emulates neurons perfectly as shown in the diagram Emulating Neurons. Each virtual device can publish multiple channels. In the example above each sensor input is broken into three published channels, one excitatory channel, and two inhibitory channels. The inhibitory channels are a percentage of the excitatory channel. A series of virtual channels subscribe to the published inhibitory channels from neighboring published channels and from the central excitatory channel. Analytics sums the subscription channels and publishes a single accumulated channel with edge detection. Simple, elegant, beautiful, just like nature.