Wanting, Liking, Learning

Three jobs that come apart under the knife — and the one that turned out to be an algorithm somebody had already written down.

A course on the thought of The Reward System.

Psychology · Neuroscience

What you will learn

• Explain why a behavioural measure of pleasure was needed at all, and what taste reactivity actually scores.
• Describe the salt-appetite reversal and say what it establishes about the measure.
• State the limit of the method in one sentence, without apologising for it.
• State the size of the accumbens hedonic hotspot and what opioid stimulation inside it does.
• Describe the coldspot result and explain why it rules out a simple dose account.
• Recognise that liking is produced by a small distributed circuit rather than a pleasure organ.
• State the self-stimulation result and the common-sense reading of it.
• Explain incentive salience as an alternative account, and what evidence motivated it.

Course outline

1. How Do You Measure a Pleasure? — You cannot ask a rat whether it enjoyed that. The measure the field settled on is the face — and one experiment with salt shows the face is not reporting the taste.
2. A Cubic Millimetre — The place where liking is made turns out to be about the size of a grain of rice. The same drug, a short way behind it, does the opposite.
3. Wanting Without Liking — A rat and a man both pressed a button thousands of times. The obvious explanation is that it felt wonderful. There is another one, and the people who propose it are careful to say they are only…
4. The Neuron That Reports Surprise — One cell, one monkey, one drop of juice. The cell fires for the reward — until the reward becomes predictable, and then it stops, and moves.
5. The Silence That Arrives On Time — The predicted reward does not come, and the neuron goes quiet at exactly the moment it should have arrived. Nothing occurred at that moment. The timing is the finding.
6. The Algorithm Was Already There — The rule the midbrain turned out to be running had been written down fifteen years earlier by people trying to make machines predict. It explains learning. It does not explain the cubic millimetre.

Works read in this course

• Pleasures of the Brain
• A Neural Substrate of Prediction and Reward

Terms introduced

hedonic hotspot, taste reactivity, incentive salience, reward prediction error, temporal difference learning, nucleus accumbens, self-stimulation

Part of the Simposeum course library.