Motivational and representational theories of pain
| Ben Seymour |
| Institute of Neurology , UCL , UK |
The Marrian approach to systems neuroscience places primacy on understanding and defining the function of the systems under question. For pain, the brain must solve two broad problems: firstly, it has to decide whether or not, and to what degree, an ascending nociceptive input is painful, and secondly, what to do about it. Respectively, these problems are referred to as representational, and motivational.
Representational theories of pain deal with the problem of inference – how best to integrate information regarding the intensity or unpleasantness of pain provided by peripheral input, with that provided by prior expectation. Statistical theories of perception, more commonly applied to modalities such as vision and audition, provide an account of many perceptual characteristics of pain (such as the placebo effect), and promote a somewhat different account than advanced in many traditional theories of pain modulation, which often place precedence on attenuation of input, rather than perceptual synthesis.
Motivational theories of pain deal with the problem of control – how best to make decisions in the face of pain, and learn from ones experiences to optimise behaviour in the future. This draws strongly on an extensive literature on aversive (and appetitive) learning, usually studied in the context of classical and instrumental learning. Computational theories of optimal control, such as reinforcement learning, provide a solid framework for understanding the motivational basis of pain, and are starting to make direct contact with diverse bodies of physiological, pharmacological, behavioural and imaging data.
Workshop HOME
