Why choose this book? by Read Montague
Computational Theory of the Mind + Reinforcement Learning. As seems normal for recent psychology-related books, the notes prove better reading than the main text.
Principle 1: Compute as slow and as softly as possible.
Principle 2: Be as imprecise as possible and compress everything
Principle 3: Stay off the lines, don't repeat yourself, and be as noisy as possible.
Principle 4: Expect goals, desires (error signals), and values.
The nervouse system's parts and their interactions learn (adjust), guided by the new goal. As the internal neural structures adjust to achieve the goal, cognition literally changes, and often in hard-to-reverse ways. A familiar example is drug addiction.
Anyone who has ever dealt closely with a drug addict knows that she no longer views the world the same way. One reason is that drugs of abuse derail exactly the guidance signals that wwe have been considering. Under the perturbed guidance of drugs, an addict's nervous system readjusts complext goal-seeking mechanisms and chages her mental world; it's very much like a pathological software update...
Once a goal is selected by your nervous system, a kind of natural information cycle is set up to guide behavioral choices. There are four basic steps in this cycle. The nervouse system must (1) hold the current goal in mind, (2) produce a critic signal for this goal, (3) use the critic signal to guide chioces and improve the brain's model of the goal, and (4) select the next goal (or keep the current one active).
The sad story of Heaven's Gate cult shows that ideas, however strange, can act with the potency of primary reinforcers -- actually something even more important than any natural reinforcers. Dopamine systems located in your midbrain, in collabroation with your prefrontal cortex, striatum, and other areas, participate in the computation and storage of values attached to ideas. The central concept was that an idea, formed in your prefrontal cortex, comes to play the role of a special control signal, a reward-dependent guidance signal.