Fully analog implementation of adaptive spiking level-tuned synapse with adjustable parameters. Intended for researchers, students and teachers.

The circuit performs three synaptic functions in parallel:

• Gaussian Weight Function (pinched absolute deviation of input from weight)
• Spike generation (pulse rate coded)
• Weight adaptation (unsupervised pure Hebbian -- saturation easily prevented be tuning weight so that weight = average input when neuron is firing)

Learning rates, max-min spike frequency and pulse length can all be controlled by the potentiometers and variable capacitors. An LED is included for visualization of the spike train.

A header is provided for input, output, post-synaptic feedback and weight monitoring.

With some minor modifications to the circuit design, to reduce power consumption, the 2 capacitors can be replaced by memristors (now available to the public on knowm.org).