how can we reduce the number of hidden layer and in the same time how to increase the number of neuron in that reduced hidden layer???

X = abalone_dataset;
 hiddenSize = 100;
 autoenc = trainAutoencoder(X,hiddenSize);

1)what 'hiddensize' means???

does it mean ---- auto encoder has 100 hidden layer or auto encoder has only one hidden layer and 100 neurons in one hidden layer?????(FIG. GIVEN BELLOW)

2)Another question is that ::---- how can we reduce the number of hiddenlayer and in the same time how to increase the number of neuron in that reduced hiddenlayer???