Insights from Cognitive Neuroscience
Total Page:16
File Type:pdf, Size:1020Kb
OUP UNCORRECTED PROOF – FIRSTPROOFS, Sat Jun 22 2013, NEWGEN C H A P T E R Expectancies and Beliefs: Insights 23 from Cognitive Neuroscience Lauren Y. Atlas and Tor D. Wager Abstract AQ: Per OUP guidelines for Expectations influence clinical outcomes and ongoing experience across nearly all psychological abstracts, no domains. They color our perceptions, drive learning and memory, and shape the generation of first perosn is used. Abstract emotional responses. Despite their profound influence, researchers have only recently begun to focus is edited on the mechanisms by which expectancies actually modulate subjective experience. This chapter accordingly; please revise as describes a cognitive neuroscience approach to the study of expectations, focusing on expectancy needed. effects on affective experience. First a brief history is provided of the development of expectation as a construct with explanatory power in psychology, and several distinct types of expectancy are discussed. Next, the chapter describes the role of expectations in affective processes, both during anticipation and during the experience of hedonic outcomes. The chapter ends with a discussion on the brain mechanisms currently thought to underlie expectations and their effects, first focusing on expectancies across domains, and then specifically on pain, an area that has proven to be a particularly tractable and informative model system. Key Words: expectancy, beliefs, placebo, anticipation, emotion, cognitive neuroscience, conditioning, pain, learning High expectations are the key to everything . —Sam Walton, founder of Wal-Mart I f nd my life is a lot easier the lower I keep my expectations. —Calvin, from “Calvin and Hobbes,” by Bill Watterson Expect nothing, live frugally on surprise . —Alice Walker Expectations shape the world we perceive, for who believe the rats were specifi cally bred to per- better and for worse. Students tend to score lower form poorly in mazes, in contrast to experiment- on IQ tests when teachers expect them to perform ers who believe the rats come from a brighter breed poorly (Raudenbush, 1984; Rosenthal, 1994). (Rosenthal & Fode, 1963). Expectations can also be Experimenters’ expectations infl uence experimental profoundly benefi cial. Expectations that a medical outcomes (Rosenthal & Rubin, 1978), even in sim- treatment will be benefi cial can elicit placebo eff ects ple observational studies of animal behavior: Rats that infl uence pain (Price, Craggs, Verne, Perlstein, are slower when they are tested by experimenters & Robinson, 2007), depression (Kirsch & Sapirstein, 359 223_Ochsner-V2_Ch23.indd3_Ochsner-V2_Ch23.indd 335959 66/22/2013/22/2013 99:20:32:20:32 PPMM OUP UNCORRECTED PROOF – FIRSTPROOFS, Sat Jun 22 2013, NEWGEN 1998), symptoms of Parkinson’s disease (de la expectations involve a belief that something will Fuente-Fernandez et al., 2001), and other physiolog- happen in the future (Kirsch, Lynn, Vigorito, & ical outcomes (Meissner, Distel, & Mitzdorf, 2007). Miller, 2004). Expectancy theory dates back to the Although outcomes are shaped toward expectations middle of the twentieth century, when behavior- in all of these examples, expectations can also bias ism dominated the fi eld. Woodworth (1947) and perception in the opposite direction. Individuals Tolman (1949) argued that when an animal learns often complain when peers rave about a popular new that a tone predicts a shock, the animal is essentially fi lm, for fear of experiencing disappointment if the developing an expectation about the timing of the fi lm fails to live up to expectations. In sum, expec- shock and the relationship between the tone and tancies color all areas of aff ective experience. T ey the stimulus. T ese ideas were further developed by fi ll us with dread or excitement. T ey aff ect how we Bolles (1972), who argued that conditioned stimuli experience events themselves. T ey bias our memo- do not directly elicit responses; instead, contingent ries of signifi cant occasions, sometimes outweighing reinforcements cause animals to develop expectan- the infl uence of our feelings during the event. T us, cies about outcomes, which in turn elicit responses it is not surprising to see disagreement over whether insofar as the animal is motivated to achieve or it is better to expect the best, expect the worst, or avoid that outcome. More formally, animals learn attempt to live life without expectations. either stimulus-outcome contingencies (S-S*) or T e study of expectations is a very broad one response-outcome contingencies (R-S*), and behav- indeed, as expectations play a critical role in vir- iors are exhibited as a function of the value of the tually every area of psychology and neuroscience. expected outcome (S*); thus a hungry animal will T eir eff ects operate across multiple levels of analy- be more likely to exhibit responses than a satiated sis, from social behavior to low-level neurobiological one because of a diff erence in S*. Finally, Rescorla responses. T ey are critical in perception, learning and Wagner (1972) formalized a model of classical (from simple conditioning to complex problem conditioning to explain phenomena such as block- solving), memory, attention, judgment and deci- ing and conditioned inhibition, which suggest sion-making, social behavior, and disorders of the that learning does not depend on simple contigu- mind and brain. T eir study is embedded within ity between conditioned and unconditioned stim- each fi eld of psychology and is encapsulated by uli. Instead, conditioned responses are elicited on none. Despite the pervasive infl uence of expectancy, the basis of the information value of conditioned researchers have only recently begun to examine the cues, not simply as refl exive responses to the cues precise brain mechanisms by which expectancies themselves (Rescorla, 1988). T us, according to modulate perception, emotion, and judgment. this perspective, expectancies underlie most forms Here, we describe a cognitive neurosci ence of learning (Reiss, 1980). Notably, these accounts approach to the study of expectations, focusing on the also suggest that expectancies can be studied in basic eff ects of expectations on brain processes and behav- animal models. ior, and on the cognitive and brain mechanisms that T e defi nition of expectancy evolved with the underlie them. Rather than comprehensively review- cognitive revolution. Researchers focused on a more ing expectancy eff ects across all areas of psychology, we cognitive interpretation of the notion of expectancy, will focus on expectancy eff ects in aff ective processes, requiring that expectancies involve explicit, verbal- including the experience of pain and other responses izable awareness of contingencies. T is gave rise to to events with high relevance for physical and social important distinctions between conditioning and well-being. Among all the domains of aff ective and expectancy. Conditioning in humans can produce clinical outcomes infl uenced by expectations, pain is explicit awareness of stimulus contingencies and the most well studied, with strong evidence for causal thus lead to conscious expectancies (Brewer, 1974; eff ects of expectations on the experience of both clini- Kirsch, 1985). Expectancies of this type can be dis- cal and experimental pain (Benedetti, Carlino, & tinguished from other types of conditioned learn- Pollo, 2011; Vase, Petersen, Riley, & Price, 2009; Vase, ing that are unconscious, in several ways (Benedetti Riley, & Price, 2002). et al., 2003; Kirsch et al., 2004). First, insofar as conditioning can occur without conscious aware- A Brief History: Conditioning ness (Clark, Manns, & Squire, 2002; Lovibond vs. Expectancy & Shanks, 2002), it is distinct from expectancy. Expectancies have been defi ned in many ways, Second, conscious expectations can be elicited by but a basic commonality across defi nitions is that verbal information or social observation, without 360 EXPECTANCIES AND BELIEFS: INSIGHTS FROM COGNITIVE NEUROSCIENCE 223_Ochsner-V2_Ch23.indd3_Ochsner-V2_Ch23.indd 336060 66/22/2013/22/2013 99:20:34:20:34 PPMM OUP UNCORRECTED PROOF – FIRSTPROOFS, Sat Jun 22 2013, NEWGEN any previous experience with a given stimulus or Montgomery & Kirsch, 1997; Voudouris et al., situation. T ird, if conditioned eff ects are impervi- 1985, 1989, 1990; for a thorough review, see ous to changes in expectations, they are likely dis- Stewart-Williams & Podd, 2004). One infl uential tinct from expectancy. experiment (Benedetti et al., 2003) tested the basis Interest in the relationship between conditioning of placebo eff ects on consciously accessible outcomes and expectancy grew when the medical community (pain in healthy controls and motor performance in began to acknowledge the power of expectancy. At patients with Parkinson’s disease) and physiological the same time that Tolman and others were arguing outcomes that are not accessible to direct conscious for a new interpretation of classical conditioning, experience (cortisol and growth hormone secretion). Henry Beecher published an infl uential article enti- T e critical groups went through conditioning tled “T e Powerful Placebo” (Beecher, 1955), which phases (pretreatment with the analgesic ketorolac for included an early meta-analysis of 15 studies that pain conditioning, subthalamic nucleus stimulation administered placebos for conditions as diverse as for Parkinson’s patients, and treatment with suma- wound pain, seasickness, anxiety, and the common triptan for cortisol and growth hormone secretion) cold. Beecher