^A The inability to differentiate imagination from reality is, of course, a key component of several psychological disorders (such as schizophrenia).³

^B Working memory keeps ideas, concepts, and results easily accessible for immediate processing. The complexity and the mechanisms that allow working memory are a very large topic and are the subject of multiple books.⁶ It is related to the concept of rehearsal, where one repeats the thing-to-be-remembered over and over again,⁷ and prospective memory, where one has to remember to be ready for a future event⁸ (“Don’t forget the milk”). The capacity of working memory is strongly tied to IQ, mental ability, success in life, and self-control.⁹ Recent experiments have suggested that it is possible to train working memory, and that this training has effects on decision-making.¹⁰

^C For my metric-using readers, that’s –17° Celsius! The University of Minnesota has most of its buildings connected by skyways, bridges between buildings set above the street, usually on the second or higher stories. These skyways are often beautiful, with large glass windows, but they form a complex maze. Some interbuilding connections are above the street, while others are tunnels below the street. If you know the secret passages, you can often get quite a ways across campus without ever going outside.

^D Interestingly, although search algorithms seem to be able to reliably beat human players at chess, other games, such as the game of Go, have been much more difficult.

^E We weren’t looking for this. We were actually trying to examine hippocampal representations on a specific maze to compare them with our dorsal striatal recordings from rats running that same maze.³¹ But one day, Adam came into my office and said, “Dave, my animals are showing mental time travel!” To put it politely, I didn’t believe him; I was deeply incredulous. As we checked and rechecked our findings for possible alternative hypotheses, we were eventually left with the conclusion that our rats were indeed representing possible paths ahead of themselves. This is the fun of science—discovering something completely unexpected. This is captured elegantly in the famous quote usually attributed to Isaac Asimov, “The most important phrase in science is not actually ‘Eureka,’ but rather ‘Hey, that’s funny.’”

^F One exception is the work on race-to-threshold models, which accumulate evidence for individual choices until one of the choices reaches a threshold.³⁵ Although there is neural evidence for these race-to-threshold models, the evidence tends to come from experiments that include perceptual ambiguity, but which have not incorporated the multiple decision-making systems known to exist. One experiment has looked at the effect of value ambiguity and found evidence for race-to-threshold integration in humans making decisions between equally valued but different foods.³⁶ We will examine race-to-threshold models in our discussion of the role of perception in decision-making (Chapter 11). An interesting possibility is that evaluation uses the same mechanisms as perception, applied to hypotheses about future internal states.

^G It is worth taking a digression to talk about introspection and psychology. Introspection is the conscious internal examination of our own mental processes. It has a very long tradition in psychology, going back to the earliest philosophy—the famous dictum was “Know thyself.”³⁸ The problem is that, as we all know (and as much of what this book is about), we are often wrong about our own decision-making processes.³⁹

Normal psychology (the study of normal mental processing, as compared to “abnormal psychology,” the study of mental disorders) arose in the 1800s with the study of perception.⁴⁰ These experimental psychologists developed a technique in which they trained themselves to respond to signals quickly, and quantitatively measured how quickly they recognized something (such as a picture of a face).

The psychologists using introspection at the end of the 19th century called themselves Gestalt psychologists, from the German word Gestalt, meaning form or shape, because they believed that they were learning to recognize the overall form or shape of the world. The problem was that they were measuring introspective responses. (How quickly do I recognize this signal?) With the discovery by Sigmund Freud (1856–1939) that some psychology was unconscious, the conscious-based experiments of the Gestaltists became problematic. (It is important to recognize that the Gestalt experiments were quantitative [measuring how quickly they responded to a signal], even if they were introspective.) In the early part of the 20th century, these introspective experiments were abandoned for direct measures of behavior, under the guise of behaviorism, which led to the belief that since it was impossible to measure internal cognition, introspection was not good scientific practice.⁴¹ The argument was that we should treat the brain as a black box, since the only thing we can measure is behavior. Of course, as we have seen throughout this book, we can now open the black box and measure things other than behavior (such as neural activity; see Appendix B). We can now measure those internal cognitive states.⁴² Nevertheless, the problems with introspection remain.

The fundamental problem with introspection, of course, is that, as we saw in Chapter 1, we contradict ourselves, we each contain multitudes, and not all of our internal processes are accessible to our consciousness.⁴³ This doesn’t mean that we can’t use introspection to guide our scientific hypotheses, but we are going to need to check those hypotheses with careful experiments based on direct observations. In psychology, this is called operationalizing our hypothesis—the idea is that we can’t measure introspective concepts directly, but we can measure behavior (or neural signals). If we can find something that correlates with or differentiates our introspective hypotheses, then we can scientifically study that.

^H Ventral means “toward the chest” in Latin and can be contrasted to dorsal, which means “toward the back” in Latin. In animals that run on all fours (quadrupeds, like dogs and cats), the chin of the animal is toward the chest, while the top of the head is toward the back. In humans, because we’ve stood up (become bipedal), the bottom of the brain is toward the ground and the dorsal part of the brain is toward the sky. Scientists stick to these terms so that they can use the same terminology for the same parts of the brain, such as the dorsal and ventral striatum, which are well conserved between other animals and humans.⁵² Think of ventral as toward the bottom of the brain and dorsal as toward the top.

^I Thus orbito, from the “orbit” of the eyes, referring to the circular bone that surrounds the eye, and frontal. We encountered the orbitofrontal cortex in our discussion of emotion and Pavlovian action-selection systems (Chapter 8).

^J Remember the television example from Chapter 3. When given a choice between $100, $200, and $300 televisions, people are much more likely to pick the $200 television over the $300 one. But when given a choice between $200, $300, and $400 televisions, people are much more likely to pick the $300 television over the $200 one. This is called framing, because the set of choices frames your decisions.⁵⁷

^K As we saw in our discussion of multiple decision-making systems (Chapter 6) and will return to in our discussion of Procedural learning (Chapter 10), there are some tasks that individuals do learn gradually and incrementally.⁷⁷ Suddenly getting the answer right (an “Aha!” moment) is a sign that a task is being solved by the Deliberative system rather than the Procedural system.