It introduces a Hebbian-LMS algorithm, an integration of unsupervised Hebbian learning and supervised LMS learning in neural networks, as a mathematical representation of a general theory for synaptic learning in the brain, and adaptation and functional control of homeostasis in living systems.