Stacked Autoencoder for regression output layer

I am trying to use stacked auto encoder (SAE) by following this link:

https://uk.mathworks.com/help/deeplearning/ug/train-stacked-autoencoders-for-image-classification.html

I get that the first two autoencoders can be stacked normally, however my problem is regression output and the last layer which is stacked in the link is trainSoftmaxLayer.

Below is the code i would like to replace:

softnet = trainSoftmaxLayer(feat2,tTrain,'MaxEpochs',400);

Which should look like:

regNet = trainRegressionLayer(feat2,tTrain,'MaxEpochs',400);

Is there any way to train regression layer with SAE?