Magazine R9 Primer Decision making Jeffrey D Schall Daily we encounter numerous alternative courses of action among which we must choose These alternatives vary in complexity and consequence Some alternatives are considered lightly such as the way to turn on the journey home Others require fervent deliberation such as Hamlet s agonizing decision whether to end his life The neural basis of decision making has been reviewed extensively 1 4 so the goal of this primer is to orient the interested reader to this growing literature with emphasis on neurophysiological work rather than the expansive literature on functional brain imaging and neuropsychology Another goal is to articulate the key concepts of choosing deciding intending and acting 5 6 Dwelling on terminology may seem an unnecessary tangent but science travels on its vocabulary inconsistent and vague terms can only yield confusion This is all the more important when the object of this scientific investigation ultimately is nothing less than human agency Choice A choice is required when an organism is confronted with alternatives for which an action is required to acquire or avoid one of the alternatives because of a desire goal or preference In its most fundamental sense a choice is an overt action performed in the context of alternatives for which explanations in terms of purposes can be given Consider the options of selecting between two envelopes A and B containing different amounts of money Table 1 Choice 1 is easy because it offers the guarantee of obtaining 1 000 by selecting envelope A instead of 1 by selecting envelope B Choice 2 requires more thought because the amount of money is unpredictable but a suitable analysis leads to a wise choice Choice 3 requires deliberation to contend with a paradox of choice prediction 7 Regardless of the complexity of the alternatives though you must execute your choice through some action such as grasping one of the envelopes Choices have particular characteristics First choices are evaluated as good or bad according to whether goals are achieved and consequences are as expected Second choices take time a choice process evolves from a state of more or less equipotentiality immediately after the alternatives are presented to a state of commitment before the overt action is performed Third with prior knowledge of the alternatives and preferences some choices can be predicted According to this definition neurophysiological correlates of choice behavior are studied whenever an experimental subject has alternative responses This has been studied for example with psychophysical discrimination 8 and visual search 9 With easily distinguished alternatives and predictable reinforcements like choice 1 the neural representation of the stimuli is resolved unambiguously leading to an earlier and more accurate motor response When the stimuli are more difficult to distinguish 10 11 or the reinforcement becomes less predictable 12 13 like choice 2 the neural representation of the alternative choices takes more time to resolve and that representation can be modulated by the expected value of the choice Finally although no neurophysiological studies have been conducted in the context of Newcomb s paradox new studies are exploring situations in which the experimental subject chooses alternatives supposedly based on the actions of another agent 14 Choices made in the context of the choices of another agent are also guided by reinforcement history but also with some sensitivity to the strategies employed by the other agent This progression of experimental conditions should make clear that some choices are easy and others are hard When alternatives are vague payoffs are unclear and stakes are high then it seems natural to regard the subjects as making a decision before they act Decision The term decision is used casually and technically in several not entirely consistent senses On the one hand decision theory describes the characteristics of effective behavior in relation to expected consequences 15 For example the best solution to the second choice in Table 1 is to select the envelope for which the product of the probability and the monetary amount is greatest This is the domain of decision theory and related frameworks such as economics and game theory A very active area of inquiry of late has employed concepts of decision theory and related disciplines in the investigation of the neural preliminaries of choice behavior 16 Table 1 Choices based on predicted monetary reward Envelope A Condition Payoff Envelope B Condition Payoff 1 100 1 000 100 1 2 9 10 000 90 1 000 3 100 1 000 If only B 1 000 000 If A and B 0 Two envelopes labeled A and B are offered for you to obtain Choice 1 is between two transparent envelopes envelope A contains 1 000 and envelope B 1 Choice 2 is between two opaque envelopes Envelope A has a 9 chance of containing 10 000 and envelope B has a 90 chance of containing 1 000 Choice 3 is Newcomb s paradox Envelope A is transparent and contains 1 000 Envelope B is opaque and will contain either 1 million or nothing You will be given the choice of taking either B alone or A with B Before you are presented with the envelopes an infallible brain scan is done to predict your choice If the prediction is that you will choose only envelope B then 1 million is placed in it If it is predicted that you will choose both envelope A and B then nothing is put in envelope B Current Biology Vol 15 No 1 R10 On the other hand we can refer to a decision as a process that results in the overt act of choosing Such a process must have a particular architecture but measures of outcome derived from decision theory do not uniquely specify mechanism Decision as a process has two logically and practically distinct meanings We must distinguish between decide to which is a selection between alternative actions and decide that which is a selection between alternative categories of an image or concept The logical distinction is easy to see You can decide that falsely but you cannot decide to falsely it is not intelligible The decision that envelope B always contained 1 million is false but in what sense can you decide falsely to take envelope B You just take it or you do not Decide to like choosing is judged as good or bad but not as true or false The fact that choices can be predicted makes it possible to choose in advance it is intelligible to say I will choose envelope B when given the opportunity This sense of