A sigmoid neuron is similar to a Perceptron but they are modified so that small changes in their weights and biases only cause a small change in their output. This allows a network of sigmoid neurons to learn.
Just like a perceptron, the sigmoid neuron has inputs and outputs as well. However, these inputs and outputs will not be binary. Inputs and outputs can take on any values between 0 and 1. Now, instead of the output being 0 or 1, it is: where is the Sigmoid Activation Function (logistic function).