How to build a [water] brain

By Fabiana Kubke 18/06/2010

Whenever I try to teach some aspects of neuronal integration in class, I run into trouble, since most of the neuronal properties are defined by mathematical formulae that describe the electrical properties of neurons that are sometimes difficult for the students to grasp. Without a basic knowledge of electricity, it is hard to build a conceptual image of what neurons are doing.

Or is it?

I was invited to talk about the brain to a group of 9-11 year old pupils in a primary school in the North Shore yesterday, when I thought it might be fun to try to build neurons and discover how they worked. So, here is my water neuron:

The water neuron

It turns out, this little water neuron (which can be built with pretty much household items) has a lot to show about the passive properties of neurons.

Synaptic currents:

The pipette dropper was used to inject [current] water into the different dendrites. Because of the properties of the dropper, there is a limit to the amount of current that can be injected at a given time, and the injection of current is not instantaneous but has a time course that is analogout to the time course of the synaptic potential.

Spatial and temporal integration:

Current can be injected in one or more dendrites with different time patterns. Injecting into all dendrites at the same time, or into one or more dendrites at different time intervals provides a good idea of how the output of the neuron is shaped by spatial and temporal integration.


By tilting the ‘soma’ to different degrees the amount of current needed to be injected into the dendrites to allow for an output of the axon will increase. Therefore, one can build neurons with different thresholds and see how that affects the output of the neuron.

Leak currents:

One can poke tiny holes into the soma so that some of the current injected into the neuron leaks out. Combining this with changing threshold and the temporal patterns of injection into the dendrites is a good way of showing how temporal and spatial integration work in different ways to produce an output through the axon.  One can also put some leaks into the axon, and ‘myelinate’ it with saran wrap to show the insulating  properties of the myelin sheath.

Although this ‘water neuron model’ cannot illustrate the active properties of the neurons, it does contribute to an intuitive construct of how currents may be acting in individual neurons. The different neurons can be connected to form a circuit, and then one could examine how the output of the circuit is affected by changing things likethreshold, leak and number of inputs into individual neurons.

Well, it was fun. I may give this a go in my next neuro class at Uni.