Manuscripts submitted to Social Cognitive and Affective Neuroscience

“Willpower” Over the Life Span: Decomposing Impulse Control

For Peer Review

Journal: Social Cognitive and Affective Neuroscience

Manuscript ID: Draft

Manuscript Type: Original Manuscript Date Submitted by the Author:

Complete List of Authors: Mischel, Walter; Columbia University, Psychology Ayduk, Ozlem; University of California - Berkeley, Psychology Berman, Marc; University of Michigan, Psychology Casey, B. J.; Weill Cornell Medical College, Cornell University, Sackler Institute for Developmental Psychobiology Gotlib, Ian; Stanford, Psychology Jonides, John; University of Michigan, Psychology Kross, Ethan; University of Michigan, Psychology Teslovich, Theresa; Weill Cornell Medical College, Cornell University, Sackler Institute for Developmental Psychobiology Wilson, Nicole; University of Washington - Seattle, Psychology Zayas, Vivian; Cornell University, Psychology Shoda, Yuichi; University of Washington - Seattle, Psychology

Keywords: Impulse Control, Delay of Gratification

http://mc.manuscriptcentral.com/scan Page 1 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 FOR SPECIAL ISSUE on AGING for SCAN 4 5 6 7 8 “Willpower” Over the Life Span: Decomposing Impulse Control 9 10 11 Walter Mischel 12 13 Columbia University 14 15 Ozlem Ayduk 16 17 18 ForUniversity Peer of California, Review Berkeley 19 20 Marc G. Berman 21 22 University of Michigan 23 24 25 B. J. Casey 26 27 Sackler Institute of Weill Cornell Medical College 28 29 Ian Gotlib 30 31 32 Stanford University 33 34 John Jonides and Ethan Kross 35 36 37 University of Michigan 38 39 Theresa Teslovich 40 41 Sackler Institute of Weill Cornell Medical College 42 43 44 Nicole Wilson 45 46 University of Washington 47 48 Vivian Zayas 49 50 51 Cornell University 52 53 Yuichi Shoda 54 55 University of Washington 56 57 58 59 60 1 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 2 of 31

1 2 3 Abstract 4 5 6 7 8 Preschool children at Stanford University’s Bing nursery school 40 years ago 9 10 11 were tested for their ability to delay gratification on the “marshmallow test”. 12 13 Longitudinal follow up studies traced the long-term correlates of the observed early life 14 15 individual differences, focusing on consequential social, cognitive, and mental health 16 17 18 outcomes in each decadeFor over Peerthe developmental Review course. As they reach mid-life, 19 20 individuals from this study consistently high or low in delay ability and impulse control 21 22 are being assessed by our inter-disciplinary team with fMRI and a variety of cognitive 23 24 25 information processing measures to try to further identify and decompose the processes 26 27 that underlie the phenomena of “willpower” and impulse control. This article reviews 28 29 key findings from the longitudinal work, and from earlier experiments on the cognitive 30 31 32 appraisal and attention control strategies that influence the ability to delay gratification. 33 34 We also report preliminary findings from the mid-life follow up in progress, and discuss 35 36 37 implications of the research for the prediction and potential enhancement of important 38 39 life outcomes in the aging process. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2 http://mc.manuscriptcentral.com/scan Page 3 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 More than four decades ago, the Bing Longitudinal Study began at Stanford 4 5 6 University’s preschool to try to identify and demystify the processes that underlie the 7 8 phenomena of “willpower.” These phenomena include diverse aspects of control over 9 10 11 one’s cognitions, emotions, and actions, particularly the ability to delay immediate 12 13 gratification and resist impulsive responding. With that goal, experimental studies and 14 15 long-term longitudinal investigations have been examining children’s abilities to delay 16 17 18 immediate gratification.For This abilityPeer is measured Review by assessing how long the child could 19 20 resist settling for a small, immediately available reward (e.g., one mini marshmallow) in 21 22 order to get a larger reward later (e.g., two mini marshmallows), dubbed the 23 24 25 “marshmallow test” by the media (e.g., Mischel, Ebbesen & Zeiss, 1972; Mischel, 26 27 Shoda, & Rodriguez, 1989; Mischel & Ayduk, 2004). 28 29 30 31 32 Attentional Control and Cognitive Reappraisals Enabling Delay: Early Experiments 33 34 This research began in the early 1970s with a series of experiments designed to 35 36 37 elucidate the attentional and cognitive mechanisms that enable young children to delay 38 39 gratification. These experiments helped to identify the role of attentional control, for 40 41 example, by self-distraction to reduce the frustration of continuing to wait (e.g., looking 42 43 44 away from vs. toward the temptations). The studies also documented the importance of 45 46 cognitive transformations or reappraisals in which the “hot” (e.g., appetitive, 47 48 consummatory) features of the tempting stimuli are “cooled” by changing how they are 49 50 51 mentally represented (e.g., envisioning the marshmallow as clouds). Such reappraisal 52 53 processes to “cool” hot stimuli have been shown to be highly effective in enhancing delay 54 55 of gratification. The same preschool child who yielded immediately to the temptation by 56 57 58 59 60 3 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 4 of 31

1 2 3 focusing on the consummatory features (e.g., its yummy, sweet, chewy taste) could wait 4 5 6 for otherwise unbearable delay periods for the same tempting stimulus by focusing on its 7 8 non-consummatory qualities (e.g., its shape). These results have been fully described 9 10 11 elsewhere (see e.g., Mischel, Ebbesen, & Zeiss, 1972; Mischel, 1974; Mischel, Shoda, & 12 13 Rodriguez, 1989; Mischel & Ayduk, 2004)). Although these experimental demonstrations 14 15 have been short-term and confined to brief laboratory situations, they suggest that these 16 17 18 strategies required Forto successfully Peer self-regulate Review can be taught. 19 20 The findings from these experiments and related research also led Metcalfe and 21 22 Mischel (1999) to propose a “Hot/Cool System Analysis of Delay of Gratification” to 23 24 25 explain the dynamics of “willpower.” In that conceptualization, there are two interacting 26 27 neurocognitive systems that enable self-control and willpower in the execution of one’s 28 29 intentions. A cool system, likely involving the top-down attentional network that includes 30 31 32 the dorsal frontoparietal neural circuit (Corbetta and Shulman, 2002) which underlies 33 34 processing that is affectively-neutral, flexible, slow, and strategic. It is the seat of self- 35 36 37 regulation and self-control (Mischel & Ayduk, 2004; Metcalfe & Mischel, 1999). The 38 39 other (hot) system, likely mediated by the amygdala and other limbic structures, is the 40 41 basis of emotionality, fears, and passion; it is impulsive and reflexive, and sometimes it is 42 43 44 initially controlled by innate releasing stimuli (and is thus literally under stimulus 45 46 control). The hot system is fundamental for emotional (classical) conditioning, and it 47 48 undermines efforts at self-control. The balance between the hot and cool systems is 49 50 51 determined by stress, development, and the individual's self-regulatory dynamics. Recent 52 53 work has successfully extended this model to help individuals cope adaptively with 54 55 intense negative emotions that require “cooling” in order to facilitate adaptive 56 57 58 59 60 4 http://mc.manuscriptcentral.com/scan Page 5 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 psychological adjustment (e.g., Kross, Ayduk, & Mischel, 2005; Kross & Ayduk, 2008). 4 5 6 Stumbling into Longitudinal Research: The Bing Life-Span Study 7 8 To the surprise of the initiators of this research program, what began as a set of 9 10 11 experiments with preschoolers turned into a life-span developmental study. Four decades 12 13 later this research is continuing to reveal remarkable patterns of coherence in 14 15 consequential psychological, behavioral, health, and economic outcomes from early 16 17 18 childhood to mid-life—theFor current Peer age of the Reviewparticipants. Given these provocative 19 20 findings and the methodological advances now available for probing self-control and 21 22 executive functions with increasing depth, this longitudinal sample provides a unique 23 24 25 opportunity for deepening our understanding of the basic neuro-cognitive mechanisms 26 27 underlying “willpower”. The present paper reports some of our current efforts to take up 28 29 this challenge. 30 31 32 Overview of Past Findings. Participants in the Bing Longitudinal Study come 33 34 from a sample of more than 300 participants who were enrolled in Stanford University’s 35 36 37 Bing preschool between 1968 and 1974. Since the original testing, the ability of these 38 39 participants to pursue long-term goals in the face of immediate temptation has been 40 41 assessed once every decade. Participants have now reached their late 30s and early to mid 42 43 44 40s, and information about their life outcomes, such as their occupational, marital, 45 46 physical health, and mental health status are continuing to become available. The findings 47 48 have surprised us from the start, and they continue to do so. For example, preschoolers 49 50 51 who delayed longer relative to other participants in the original “marshmallow test” 52 53 earned higher SAT scores and exhibited better social-cognitive and emotional coping in 54 55 adolescence (Mischel, Shoda, & Peake, 1988; Shoda, Mischel, & Peake, 1990). As 56 57 58 59 60 5 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 6 of 31

1 2 3 adults, these participants continued to have better cognitive-social functioning, and better 4 5 6 educational and economic life outcomes than did their low-delaying peers (Ayduk et al., 7 8 2000). 9 10 11 The high-delay individuals also were buffered against the development of diverse 12 13 mental health problems: they used cocaine less frequently, were less likely to suffer from 14 15 16 low self-esteem and self-worth (Ayduk et al., 2000), and had fewer features of borderline 17 18 personality disorderFor than did matchedPeer controls Review with similar dispositional vulnerabilities 19 20 (Ayduk et al., 2008). Parallel findings come from diverse demographic populations, 21 22 23 including middle-school children in the South Bronx, N.Y. (e.g., Ayduk et al., 2000; 24 25 Mischel & Ayduk, 2004) and children in a summer residential treatment program for 26 27 28 youths at high risk for problems of aggression/externalization and depression/withdrawal 29 30 (e.g., Mischel & Shoda, 1995; Rodriguez et al., 1989). For example, spontaneous use of 31 32 self-control strategies in the delay task (e.g., looking away from the rewards and using 33 34 35 self-distraction) predicted lower verbal and physical aggression as directly observed over 36 37 six weeks in the summer camp study (e.g., Rodriguez et al., 1989; Wright & Mischel, 38 39 1987, 1988). 40 41 42 In search of the mechanisms underlying long-term stability in individual differences 43 44 45 in self-control abilities 46 47 The early experiments examining delay of gratification showed that mental 48 49 representations that are “hot” or appetitive (consummatory) hinder delay and are 50 51 52 ineffective because they make it too difficult to resist the prepotent response of settling 53 54 for the immediately available treat. In contrast, representations based on attention to the 55 56 “cool,” cognitive, abstract aspects of the situation have the opposite effect. Thus, delay 57 58 59 60 6 http://mc.manuscriptcentral.com/scan Page 7 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 ability in this paradigm depends critically on the aspects of the situation to which the 4 5 6 child attends, the mental representations that are activated in working memory during the 7 8 waiting period, and the degree of control the child has over his or her response behavior. 9 10 11 It is precisely these three aspects of cognitive control that we have been pursuing in 12 13 ongoing work with the Bing sample. 14 15 16 Cognitive control mechanisms . How is cognitive control achieved? The weight of 17 18 evidence, presentedFor in a meta-analysis Peer of 41 neuroimagingReview studies of a variety of tasks 19 20 measuring cognitive control, such as the Flanker, Stimulus-response-incompatibility 21 22 23 (SRIC), Go-no/go, Simon, Stop-signal, and Stroop tasks (Nee, Jonides, & Berman, 2007), 24 25 does not favor the view that there is a unitary mechanism. The activation peaks in these 26 27 28 tasks are widely dispersed in the brain, suggesting that there are multiple mechanisms of 29 30 control. This conclusion is supported by the oft-reported low behavioral correlations 31 32 among tasks that allegedly recruit inhibitory mechanisms (Earles et al., 1997; Grant & 33 34 35 Dagenbach, 2000; Kramer et al., 1994; Shilling et al., 2002; Tipper & Baylis, 1987). 36 37 Thus, both imaging and behavioral evidence suggest that processes involved in resolving 38 39 interference come from a “family of functions” rather than from a “single unitary 40 41 42 construct” (Friedman and Miyake, 2004; Dempster, 1993; Harnishfeger, 1995; Nigg, 43 44 2000). 45 46 47 Guided by these analyses, for the first time in the four-decade long Bing 48 49 Longitudinal Study we are currently collecting data on both behavioral and cognitive, and 50 51 52 neural functioning (via functional and structural brain imaging) in the participants. 53 54 55 Current data collection. Before examining the brain functioning underlying self- 56 57 regulation, we sought to obtain information about the participants’ cognitive control 58 59 60 7 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 8 of 31

1 2 3 abilities using behavioral tasks designed to measure cognitive control. Because the 4 5 6 participants are now widely dispersed across the globe, we mailed laptop computers 7 8 preprogrammed with software for tasks designed to assess cognitive control. We focused 9 10 11 on two subsamples of the Bing study: participants who were consistently above average 12 13 in their self-regulatory abilities, beginning with their performance on the delay of 14 15 gratification task in preschool in the early 1970s, through the three follow-up assessments 16 17 18 conducted in 1984,For 1993, and 2003;Peer and those Review who were consistently below average 19 20 thorough the first four decades of their life. Our rationale for focusing on these 21 22 participants is that if individual differences in self-control reflect differences in the basic 23 24 25 operation of the brain, these may be most clearly seen by comparing individuals with life- 26 27 long trajectories of stably high, or stably low, self-control. Indeed, these individuals have 28 29 a four-decade long “track record” of high vs. low self-control. 30 31 32 One hundred and seventeen individuals met the criterion of being either 33 34 35 consistently above average or consistently below average in self-control in their original 36 37 delay of gratification performance as well as in their self-control ability in subsequent 38 39 follow-ups. In these follow-up assessments, self-regulation was assessed using items 40 41 42 from the California Child Q-set (CCQ) in 1984 (when the participants were in their late 43 44 teens), in 1993 (when they were in their late twenties), and in 2003 (when they were in 45 46 47 their late thirties). These 117 individuals were invited to participate in the current follow- 48 49 up (we did not include their 1984 score in this requirement because of the relatively small 50 51 sample size in 1984). Of those who were invited, 57 agreed to participate, were sent 52 53 54 laptop computers, completed the tasks, and returned them. Of the 57, 31 were from the 55 56 “consistently high self-regulation” group, and 26 were from the “consistently low self- 57 58 59 60 8 http://mc.manuscriptcentral.com/scan Page 9 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 regulation” group. Figure 1 shows the self-control trajectories across these three waves of 4 5 6 follow-up assessments for these 57 consistently high and consistently low participants. 7 8 The vertical axis represents the self-regulation indices, standardized within each 9 10 11 assessment period. The horizontal axis represents the year in which the self-regulation 12 13 data were collected, including the initial preschool assessment that took place between 14 15 1968 and 1973. The two lines represent the average developmental trajectories for those 16 17 18 participants whoseFor preschool delayPeer time was consistentlyReview above mean (shown in blue), 19 20 and those who were consistently below mean (shown in red). 21 22 23 ------24 25 Insert Figure 1 about here 26 ------27 28 The figure illustrates that the two subgroups we targeted in our current data 29 30 collection clearly differ from each other at every assessment during the past four decades. 31 32 33 While this is not surprising because they were chosen on the basis of being above or 34 35 below across the follow-up waves, the fact that many of the participants fell in either of 36 37 these groups, and that their means at each data point were many standard errors apart, 38 39 40 reflects the overall stability of individual differences in self-regulation. 41 42 43 Preliminary Findings: Behavioral Indices of Cognitive Control 44 45 46 Are life-long patterns of high vs. low self-regulation related to behavioral indices 47 48 of cognitive control? Specifically, we examined the following three aspects of cognitive 49 50 control: (1) blocking the entry of unwanted information (e.g., shutting out information by 51 52 53 paying attention to something else); (2) suppressing unwanted thoughts (e.g., by thinking 54 55 about something else); and (3) stopping themselves from acting upon salient cues (e.g., 56 57 58 59 60 9 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 10 of 31

1 2 3 inhibiting a response or impulse). In order to assess them, we used a set of laboratory 4 5 6 procedures designed to measure each process, as described below. 7 8 9 Evaluating the ability to block the entry of unwanted information with the Ignore 10 11 Task. We used a version of Sternberg’s (1966) item-recognition task to assess the first 12 13 two of the three aspects of cognitive control described above (Jonides, Smith, et al., 1998, 14 15 16 Jonides, Marshuetz, et al., 2000; D’Esposito et al., 1999; Thompson-Schill et al., 2002; 17 18 Nelson et al., 2003;For Nee et al., Peer 2007). Specifically, Review we modified this task to allow us to 19 20 study interference-resolution in perception and working memory (Nee & Jonides, 2008, 21 22 23 2009). 24 25 26 Participants were presented with a set of 4 words for .5 seconds, half of which 27 28 were printed in teal and half in blue. Prior to each trial, participants were give a cue (the 29 30 word “teal” or the word “blue”) telling them which 2 of the words in the upcoming 31 32 33 display were relevant on that trial. Immediately after seeing the display, participants were 34 35 given a probe word and had to decide whether it matched one of words in the relevant 36 37 38 color. If the probe did not match one of the words in the relevant color, it might have 39 40 been a word originally presented in the other color. As a control, the probe might have 41 42 been a word that had not been presented on the current trial or on any of the three 43 44 45 previous trials. A comparison between these two types of non-matching probes reveals 46 47 how successfully participants were able to “ignore” (hence the name of the task) the 48 49 irrelevant information. 50 51 52 This paradigm serves as the basis for our examination of whether adults who 53 54 55 could or could not delay gratification as children differ in their effectiveness in resolving 56 57 interference at the time of original perception of information (encoding). 58 59 60 10 http://mc.manuscriptcentral.com/scan Page 11 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Evaluating the ability to control unwanted information with the Suppress Task. 4 5 6 We also designed a task that assessed how effectively participants could “suppress” 7 8 information that already entered working memory. The design was similar to that of the 9 10 11 Ignore task. For the Suppress task, participants were presented with 4 words, two in teal 12 13 and two in blue. After the display was encoded, a cue appeared telling subjects which 14 15 two of the words (either both teal words or both blue words) to forget. After that, a probe 16 17 18 appeared. Again, ifFor the probe Peerdid not match theReview relevant words, it could be one of the 19 20 words that was to be forgotten or some control word that had not appeared recently in the 21 22 experiment. By comparing these non-matching trials, this task provides an opportunity to 23 24 25 assess how well participants can suppress irrelevant information that has already entered 26 27 working memory but is no longer relevant. Figure 2 shows the train of events on a 28 29 typical trial. 30 31 32 ------33 34 Insert Figure 2 about here 35 36 ------37 38 39 40 The data collection is still on-going. So far, however, in a sample of 57 Bing 41 42 43 participants, we found no clear association between preschool delay time and 44 45 participants’ performance on the Ignore task. As for their ability to control contents of 46 47 working memory, we found a trend of low-delay participants making a greater number of 48 49 50 errors on the Suppress task (figure 3). Recall that the participants were selected on the 51 52 basis of their either consistently high, or consistently low, self-regulation trajectory over 53 54 the last four decades. Thus, those with above average preschool delay time in this sample 55 56 57 are a subsample of such participants who also maintained above average self-regulation 58 59 60 11 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 12 of 31

1 2 3 indices during the subsequent decades; similarly, those with below average preschool 4 5 6 delay time in this sample are a subsample who also remained below average self- 7 8 regulation so far in their lives. 9 10 11 ------12 13 Insert Figure 3 about here 14 ------15 16 17 18 For Peer Review 19 Cognitive control of prepotent reaction to “hot” stimuli. The go/no-go task 20 21 (Luria, 1961) is a measure of cognitive control that requires over-riding a prepotent 22 23 24 response. This task is computer-administered and requires participants to press a button 25 26 whenever a target (go) stimulus is present, but not to respond to an infrequently presented 27 28 nontarget (no-go) stimulus. We included this task because of our interest in determining 29 30 31 the extent to which the ability to delay gratification involves the ability to stop oneself 32 33 from acting (Eigsti et al, 2006). That is, performing well on this task requires not only 34 35 36 pressing a button in response to the go stimulus, but also withholding a response to the 37 38 no-go stimulus. Therefore, performance on the no-go trials was of particular interest. 39 40 We administered an adapted version of the traditional go/no-go task (Hare et al 41 42 43 2005) in which emotional facial expressions were used as stimuli (i.e., emotional go/nogo 44 45 task). Specifically, stimuli were digital photographs of happy and fearful expressions 46 47 from eight individuals (four female and four male) obtained from the NimStim set 48 49 50 (available at www.macbrain.org; identities of the faces used were: 8, 10, 14, 16, 27, 36, 51 52 39, and 45). Hare et al. (2005) suggested that participants have a more difficult time 53 54 withholding a response when the no-go (nontarget) stimulus is a happy expression 55 56 57 because happy faces are positive approach-related stimuli, while fearful faces are 58 59 60 12 http://mc.manuscriptcentral.com/scan Page 13 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 negative and alarming stimuli. Thus, while we were interested in the no-go trials in 4 5 6 general, we were particularly interested in performance on the happy no-go trials as these 7 8 trials were hypothesized to require greater self-control in order to successfully inhibit an 9 10 11 approach response. 12 13 Figure 4 is a schematic representation of a block of trials in which fearful 14 15 expressions were the go stimuli and happy expressions were the no-go stimuli. 16 17 18 For Peer------Review 19 20 Insert Figure 4 about here 21 22 ------23 24 25 Participants were instructed to respond to the target facial expression (fearful or 26 27 happy) by pressing the space bar on a keyboard. Stimuli were presented for 500ms with 28 29 30 an interstimulus interval (ISI) of 1000ms, during which a fixation cross was presented on 31 32 the screen. Therefore, trial length was 1500ms. Two 160-trial runs comprised the task. In 33 34 35 one run, the go stimulus (target) was happy expressions. The go stimulus (target) was 36 37 fearful expressions in the other run. 38 39 40 Thus far, we have collected data on this task from a sample of 57 Bing 41 42 participants, and data collection is still on-going. Preliminary results suggest that the 43 44 45 accuracy on trials in which the happy expressions were the no-go stimuli is significantly 46 47 related to self-regulatory abilities over the first four decades of their lives, starting in 48 49 preschool, as shown in Figure 5. Among the 56 participants with valid go/no-go data (one 50 51 52 participant’s data were not used because of extremely long latencies), the correlation was 53 54 r =.27 ( p < .05). Delay of gratification at age four did not appear to be related to accuracy 55 56 57 58 59 60 13 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 14 of 31

1 2 3 on trials in which fearful expressions were the no-go stimulus or to reaction times 4 5 6 associated with any trial types. 7 8 9 ------10 Insert Figure 5 about here 11 12 ------13 14 In addition, we also administered the Barratt Impulsiveness Scale (BIS-11), a 15 16 widely used measure of impulsivity consisting of 30 items assessing attentional, motor, 17 18 and nonplanning impulsiveness.For Peer As shown in ReviewFigure 6, participants, accuracy in happy 19 20 21 no-go trials was negatively and significantly correlated with this measure of impulsivity 22 23 (r = -.37, p < .01). The correlation between standardized delay and BIS-11 scores was - 24 25 26 .58, p < .001 ( N = 56). 27 28 ------29 30 Insert Figure 6 about here 31 32 ------33 34 Exploring brain functions and structure underlying life-long patterns of self- 35 36 regulation 37 38 39 Ultimately our goal is to specify the neural correlates of these processes using 40 41 magnetic resonance imaging studies that are underway, although we are still in an early 42 43 44 stage of data collection. Prior research has elucidated areas of the brain that become 45 46 differentially activated when people engage in the three processes described above 47 48 (blocking unwanted information, suppressing unwanted thoughts, and inhibiting 49 50 51 responses), as well as the connections among areas of the brain that seem to play a key 52 53 role in these tasks. We are currently assessing these functional and anatomical features to 54 55 examine whether they are related to these three aspects of effective self-control. 56 57 58 Specifically, resolution of interference between goal irrelevant (but alluring) and relevant 59 60 14 http://mc.manuscriptcentral.com/scan Page 15 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 (and long-term goal-specific) information. There is by now a well-specified set of brain 4 5 6 regions that have been discovered in previous research that can be associated with the 7 8 ability to resolve interference between relevant and irrelevant information at the time of 9 10 11 encoding, at the time information is in working memory, and at the time when one is 12 13 producing a response. For example, the right inferior frontal gyrus has been well- 14 15 documented as a critical region in resolving interference in the go/no-go task which 16 17 18 involves response interferenceFor Peer (e.g., Casey et Reviewal., 2002; Aron & Poldrack, 2006). 19 20 Likewise, the left inferior frontal gyrus region has been implicated when interference 21 22 occurs after information has entered working memory (e.g., Jonides & Nee, 2006). And 23 24 25 the dorsolateral prefrontal and posterior parietal regions appear critical to resolving 26 27 interference at the time of encoding (e.g., Casey et al., 2000; Nee & Jonides, 2008). 28 29 30 These tasks share a key feature with the delay-of-gratification task used in our 31 32 longitudinal studies: in all these cases, prepotent information interferes with other goal- 33 34 35 specific information, thus requiring processes of cognitive control to resolve the 36 37 interference. Specifically, we predict that participants with low levels of self-control, 38 39 compared with their high-control counterparts, will be characterized by less activity and 40 41 42 less refined connectivity (e.g., less myelination and orientation regularity) in 43 44 frontostriatal (Liston et al., 2006; Casey et al., 2007) and frontoparietal circuitry (Nagy, 45 46 47 Westerberg, & Klingberg, 2004) during performance of the cognitive control tasks. We 48 49 are testing this prediction by performing diffusion tensor imaging (DTI) and functional 50 51 MRI. Participants in this study have been invited to the Lucas Center for Imaging at 52 53 54 Stanford University and only a few have been scanned to date. 55 56 57 58 59 60 15 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 16 of 31

1 2 3 Finally, the participants are reaching middle adulthood and the most productive 4 5 6 years of their lives. Therefore, we are continuing to assess consequential outcomes, 7 8 including occupational and marital status, social, cognitive, and emotional functioning, as 9 10 11 well as mental and physical health and behavior patterns relevant to mental and physical 12 13 health and economic and social well-being. 14 15 Implications and next challenges: Advancing NIA goals 16 17 18 Work by our teamFor and other Peer investigators Reviewshows that willpower (adaptive self- 19 20 regulation) in young children and adolescents is facilitated by a set of specific cognitive 21 22 and affective strategies that allow them to control their attention strategically and to 23 24 25 transform their thoughts, all in the service of effectively resisting impulsive responding 26 27 and delaying immediate gratification for the sake of later, more valued, future goals. 28 29 These cognitive-affective strategies serve to “cool” and thus reduce the emotion-eliciting 30 31 32 power of the impulse-triggering temptations and frustrations that otherwise undermine 33 34 even people’s best intentions to control their own impulsive behavior in light of delay 35 36 37 consequences. 38 39 Overall, decades of research on delay of gratification suggest an emerging view of 40 41 “willpower” and impulse control (e.g., Mischel et al.,1989; Mischel & Ayduk, 2004; 42 43 44 Kross, Mischel & Shoda, in press). Considered collectively, the results of the laboratory 45 46 experiments, conducted when the Bing study participants were in preschool, suggest that 47 48 the concept of willpower can be decomposed into at least two general classes of 49 50 51 neurocognitive processes: (1) attentional and inhibitory processes, often referred to as 52 53 cognitive-control processes ; and (2) reappraisal processes , that control the mental 54 55 representation in working memory, transforming a tempting representation into another 56 57 58 59 60 16 http://mc.manuscriptcentral.com/scan Page 17 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 (e.g., a 5-year-old might succeed in delaying gratification by transforming his/her 4 5 6 perceptual representation of a single marshmallow into a representation of a pictured 7 8 marshmallow). 9 10 11 We suggest that the pre-schoolers who were high-delayers made more use of certain 12 13 inhibitory processes than did those who were low-delayers, and it is this difference in 14 15 inhibitory ability that persisted into adulthood and led to the sequelae of being able to 16 17 18 delay gratification Forand the advantageous Peer health-protective Review outcomes and adaptive social- 19 20 cognitive development that followed. The relation we propose between cognitive control 21 22 and willpower, as manifested in the delay-of-gratification studies and their resulting 23 24 25 sequelae, is this: Willpower requires skill in overcoming tempting immediate rewards, 26 27 distractions, and frustrations in favor of greater but delayed rewards. This skill, in turn, 28 29 requires encoding only information from the environment that is relevant, keeping 30 31 32 wanted information active in working memory and suppressing unwanted information, 33 34 and selecting desired responses while withholding responses that are not optimal. 35 36 37 Findings so far are encouraging but still tentative: with further follow-up assessments we 38 39 hope to test these hypotheses with greater precision. 40 41 42 From preschool impulse control to social and economic behaviors in aging. With 43 44 NIA goals in mind, we are also including genetics in this project, as well as gathering 45 46 47 other types of aging-relevant data from our participants now, ranging from brain imaging 48 49 data to information about such economic behaviors as savings, retirement planning, 50 51 health care planning/protection, and mental and physical health. By obtaining such 52 53 54 information on our unique longitudinal sample, we are laying the groundwork for 55 56 extending the study into the coming decades to identify factors that underlie lifespan 57 58 59 60 17 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 18 of 31

1 2 3 longitudinal trajectories of self-regulation and consequent behavioral outcomes. Of 4 5 6 particular interest for understanding economic behavior over the life span, we are 7 8 including standard measures of temporal discounting, thus enabling systematic 9 10 11 comparison of such measures with the predictive and explanatory value of our behavioral 12 13 measures of delay of gratification and self-control. We are now well-positioned to 14 15 conduct such assessments over the life span, obtaining comprehensive data that are 16 17 18 relevant for our originalFor Bing participantsPeer now Review at mid-life, their young children, and their 19 20 parents (in many cases still surviving) now in their 70s, thus spanning three generations. 21 22 23 Concluding remarks . Taken collectively, findings over many decades from the Bing 24 25 study converge with those from many other studies of “willpower” and executive 26 27 28 functions at the social-cognitive (e.g., Mischel & Ayduk, 2004), and brain (Casey et al., 29 30 1997; Hare et al., 2005)levels point to a clear, albeit still tentative, set of conclusions. The 31 32 skills and motivations that enable the phenomenon of “willpower”, and particularly the 33 34 35 ability to inhibit prepotent “hot” responses and impulses in the service of future 36 37 consequences, appear to be important early life markers for long-term adaptive mental 38 39 and physical development. They predict a wide variety of consequential life outcomes, 40 41 42 including social and economic well-being and higher educational and achievement levels. 43 44 Further, these skills seem to have long-term protective effects that may importantly 45 46 47 influence how a variety of hugely problematic dispositional vulnerabilities develop and 48 49 play out (e.g., high delay ability early in life predicts fewer features of borderline 50 51 personality, less substance use). Ultimately it may be possible to target and harness the 52 53 54 underlying mechanisms into readily teachable interventions for sustained and 55 56 consequential behavior change. Our research currently seeks to identify, develop and test 57 58 59 60 18 http://mc.manuscriptcentral.com/scan Page 19 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 the tool kit of basic cognitive skills that enable willpower. Most encouraging, the 4 5 6 cognitive and affective strategies that underlie this aptitude may be fairly easy to teach 7 8 and learn (e.g., Bandura & Mischel, 1965; Mischel, et al., 1989; Mischel & Ayduk, 9 10 11 2004). 12 13 Relevant to NIA goals, the Bing cohort and the data it has already yielded 14 15 provides a unique basis for future studies of the ageing process as these participants move 16 17 18 into the next phaseFor of the life cycle.Peer The possibility Review that the self-regulatory competencies 19 20 reflected in the ability to delay gratification exert benign and protective influences on the 21 22 ageing process seems a particularly exciting prospect for further investigation. Such work 23 24 25 is certain to benefit from inter-disciplinary teams working at multiple levels of analysis 26 27 from the social cognitive and behavioral, to the neural, to the genetic. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 19 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 20 of 31

1 2 3 4 5 6 7 8 Author Notes 9 10 11 Correspondence regarding this article should be addressed to Walter Mischel 12 13 ([email protected]) or Yuichi Shoda ([email protected]). The Bing 14 15 Longitudinal Project was supported by a number of grants from NIMH and NSF to 16 17 18 Walter Mischel, ofFor which the mostPeer recent and Review active is MH39349. Ayduk, Berman, 19 20 Gotlib, Casey, Jonides, Kross, Teslovich, Wilson and Zayas are listed alphabetically. 21 22 Shoda served as the overall PI, funded by NSF, for the most recent wave of data 23 24 25 collection from the Bing longitudinal study that yielded the new data reported here. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 20 http://mc.manuscriptcentral.com/scan Page 21 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 References 4 5 6 Ayduk, O., Mendoza-Denton, R., Mischel, W., Downey, G., Peake, P., & Rodriguez, M. 7 L. (2000). Regulating the interpersonal self: Strategic self-regulation for coping with 8 rejection sensitivity. Journal of Personality and Social Psychology, 79, 776-792. 9 10 11 Ayduk, O., Zayas, V., Downey, G., Cole, A. B., Shoda, Y., & Mischel, W. (2008). 12 Rejection Sensitivity and Executive Control: Joint predictors of Borderline Personality 13 features. Journal of Research in Personality, 42, 151-168. 14 15 Aron, A. R., & Poldrack, R. A. (2006). Cortical and subcortical contributions to stop 16 signal response inhibition: Role of the subthalamic nucleus. The Journal of Neuroscience, 17 18 26(9), 2424-2433. For Peer Review 19 20 Bandura, A., & Mischel, W. (1965). Modification of self-imposed delay of reward 21 through exposure to live and symbolic models. Journal of Personality and Social 22 Psychology, 2, 698-705. 23 24 25 Casey, B., Epstein, J., Buhle, J., Liston, C., Davidson, M., Tonev, S., et al. (2007). 26 Frontostriatal connectivity and its role in cognitive control in parent-child dyads with 27 ADHD. The American Journal of Psychiatry, 164, 1729-36. 28 29 Casey, B. J., Thomas, K. M., Davidson, M. C., Kunz, K., & Franzen, P. L. (2002). 30 31 Dissociating striatal and hippocampal function developmentally with a stimulus-response 32 compatibility task. Journal of Neuroscience, 22, 8647-8652. 33 34 Casey, B. J., Thomas, K. M., Welsh, T. F., Badgaiyan, R., Eccard, C. H., Jennings, J. R., 35 & Crone, E. A. (2000). Dissociation of response conflict, attentional control, and 36 37 expectancy with functional magnetic resonance imaging (fMRI). Proceedings of the 38 National Academy of Sciences, 97, 8728-8733. 39 40 Casey, B. J., Trainor, R. J., Orendi, J. L., Schubert, A. B., Nystrom, L. E., Cohen, J. D., 41 Noll, D. C., Giedd, J., Castellanos, X., Haxby, J., Forman, S. D., Dahl, R. E., & Rapoport, 42 J. L. (1997). A pediatric functional MRI study of prefrontal activation during 43 44 performance of a Go-No-Go task. Journal of Cognitive Neuroscience, 9 , 835-847. 45 46 Corbetta, M., Shulman, G. L. (2002). Control of goal-directed and stimulus-driven 47 attention in the brain. Nature Reviews Neuroscience, 3, 201-215. 48 49 50 D’Esposito, M., Postle, B. R., Ballard, D., & Lease, J. (1999). Maintenance versus 51 manipulation of information held in working memory: An event-related fMRI study. 52 Brain and Cognition, 41, 66-86. 53 54 Dempster, F. N. (1993). Resistence to interference: Developmental changes in a basic 55 processing mechanism. In M. L. Howe & R. Pasnak (Eds.), Emerging themes in 56 57 cognitive development: Vol. 1. Foundations (pp. 3–27). New York: Springer-Verlag. 58 59 60 21 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 22 of 31

1 2 3 4 5 Earles, J. L., Connor, L. T., Frieske, D., Park, D. C., Smith, A. D., & Zwahr, M. (1997). 6 Age differences in inhibition: Possible causes and consequences. Aging, 7 Neuropsychology, and Cognition, 4(1), 45-57. 8 9 Eigsti, I., Zayas, V., Mischel, W., Shoda, Y., Ayduk, O., Dadlani, M. B., et al. (2006). 10 11 Attentional control in preschool predicts cognitive control at age eighteen. Psychological 12 Science, 17, 478-84. 13 14 Friedman, N. P., & Miyake, A. (2004). The relations among inhibition and interference 15 control functions: A latent-variable analysis. Journal of Experimental Psychology: 16 General. 133(1), 101-135. 17 18 For Peer Review 19 Grant, J. D., & Dagenbach, D. (2000). Further considerations regarding inhibitory 20 processes, working memory, and cognitive aging. American Journal of Psychology, 21 113(1), 69-94. 22 23 24 Hare, T. A., Tottenham, N., Davidson, M. C., Glover, G. H., & Casey, B. J. (2005). 25 Contributions of amygdala and striatal activity in emotion regulation. Biological 26 Psychiatry, 57, 624–632 27 28 Harnishfeger, K. K. (1995). The development of cognitive inhibition: Theories, 29 definitions, and research evidence. In F. N. Dempster & C. J. Brainerd (Eds.), 30 31 Interference and inhibition in cognition (pp. 175-204). San Diego, CA: Academic Press. 32 33 Jonides, J., Marshuetz, C., Smith, E. E., Reuter-Lorenz, P. A., Koeppe, R. A., & Hartley, 34 A. (2000). Age differences in behavior and PET activation reveal differences in 35 interference resolution in verbal working memory. Journal of Cognitive Neuroscience, 36 37 12(1), 188-196. 38 39 Jonides, J., & Nee, D.E. (2006). Brain mechanisms of proactive interference in working 40 memory. Neuroscience, 139, 181-193. 41 42 Jonides, J., Smith, E. E., Marshuetz, Koeppe, R. A., & Reuter-Lorenz, P. A. (1998). 43 44 Inhibition in verbal working memory revealed by brain activation. Proceedings of the 45 National Academy of Sciences, 95(14), 8410-8413. 46 47 Kramer, A. F., Humphrey, D. G., Larish, J. F., Logan, G. D., & Strayer, D. L. (1994). 48 Aging and inhibition: Beyond a unitary view of inhibitory processing in attention. 49 50 Psychology of Aging, 9(4), 491-512. 51 52 Kross, E. & Ayduk, O. (2008). Facilitating adaptive emotional analysis: Distinguishing 53 distanced-analysis of depressive experiences from immersed-analysis and distraction. 54 Personality and Social Psychology Bulletin, 34, 924-938. 55 56 57 58 59 60 22 http://mc.manuscriptcentral.com/scan Page 23 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Kross, E., Ayduk, O., & Mischel, W. (2005). When asking “why” doesn’t hurt: 4 5 Distinguishing rumination from reflective processing of negative emotions. 6 Psychological Science, 16, 709-715. 7 8 Kross, E., Mischel, W., & Shoda, Y. (in press). Enabling self-control: A cognitive 9 affective processing system (CAPS) approach to problematic behavior. In J. Maddux & J. 10 11 Tangney (Eds.), Social Psychological Foundations of Clinical Psychology. New York: 12 Guilford. 13 14 Liston, C., Watts, R., Tottenham, N., Davidson, M., Niogi, M., Ulug, A., et al. (2006). 15 Frontostriatal microstructure predicts individual differences in cognitive control. Cerebral 16 Cortex, 16, 553-560. 17 18 For Peer Review 19 Luria, A. R. (1961). The role of speech in the regulation of normal and abnormal 20 behavior. New York: Liveright. 21 22 Metcalfe, J., & Mischel, W. (1999). A hot/cool system analysis of delay of gratification: 23 24 Dynamics of willpower. Psychological Review, 106, 3-19. 25 26 Mischel, W. (1974). Processes in delay of gratification. In L. Berkowitz (Ed.), 27 Advances in experimental social psychology (Vol. 7, pp. 249-292). New York: 28 Academic Press. 29 30 31 Mischel, W., & Ayduk, O. (2004). Willpower in a cognitive-affective processing system: 32 The dynamics of delay of gratification. In R. F. Baumeister & K. D. Vohs (Eds.), 33 Handbook of self-regulation: Research, Theory, and Applications(pp. 99-129). New 34 York: Guilford. 35 36 37 Mischel, W., Ebbesen, E. B., & Zeiss, A. R. (1972). Cognitive and attentional 38 mechanisms in delay of gratification. Journal of Personality and Social Psychology, 21, 39 204-218. 40 41 Mischel, W. & Shoda, Y. (1995). A cognitive-affective system theory of personality: 42 Reconceptualizing situations, dispositions, dynamics, and invariance in personality 43 44 structure. Psychological Review, 102, 246-268. 45 46 Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent competencies 47 predicted by preschool delay of gratification. Journal of Personality and Social 48 Psychology, 54, 687-699. 49 50 51 Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. 52 Science, 244, 933-938. 53 54 Nagy, Z., Westerberg, H., Klingberg, T. (2004). Maturation of white matter is associated 55 with the development of cognitive functions during childhood. Journal of Cognitive 56 57 Neuroscience, 16, 1227-1233. 58 59 60 23 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 24 of 31

1 2 3 4 5 Nee, D.E., & Jonides, J. (2008). Dissociable interference-control processes in perception 6 and memory. Psychological Science, 19, 490-500. 7 8 Nee, D.E., & Jonides, J. (2009). Common and distinct neural correlates of perceptual and 9 memorial selection. NeuroImage, 45, 963-75. 10 11 12 Nee, D. E., Jonides, J., & Berman, M. G. (2007). Neural mechanisms of proactive 13 interference-resolution. NeuroImage, 38(4), 740-751. 14 15 Nee, D.E., Wager, T.D., & Jonides, J. (2007). Interference resolution: Insights from a 16 meta-analysis of neuroimaging tasks, Cognitive, Affective, and Behavioral Neuroscience, 17 18 7 (1), 1-17. For Peer Review 19 20 Nelson, J.K., Reuter-Lorenz, P.A., Sylvester, C-Y, Jonides, J., & Smith, E.E. (2003). 21 Dissociable neural mechanisms underlying response-based and familiarity-based conflict 22 in working memory. Proceedings of the National Academy of Sciences, 100, 1171-1175. 23 24 25 Nigg, J. T. (2000). On inhibition/disinhibition in developmental psychopathology: Views 26 from cognitive and personality psychology and working inhibition taxonomy. 27 Psychological Bulletin, 126(2), 220-246. 28 29 Rodriguez, M. L., Mischel, W., & Shoda, Y. (1989). Cognitive person variables in the 30 31 delay of gratification of older children at risk. Journal of Personality and Social 32 Psychology, 57, 358-367. 33 34 Shilling, V. M., Chetwynd, A., & Rabbitt, P. M. A. (2002). Individual inconsistency 35 across measures of inhibition: An investigation of the construct validity of inhibition in 36 37 older adults. Neuropsychologia, 40(6), 605-619. 38 39 Shoda, Y., Mischel, W., & Peake, P. K. (1990). Predicting adolescent cognitive and self- 40 regulatory competencies from preschool delay of gratification: Identifying diagnostic 41 conditions. Developmental Psychology, 26, 978-986. 42 43 44 Sternberg, S. (1966). High-speed scanning in human memory. Science, 153(3736), 652- 45 654. 46 47 Thompson-Schill, S. L., Jonides, J., Marshuetz, C., Smith, E.E., D’Esposito, M., Kan, I., 48 Knight, R. T., & Swick, D. (2002). Effects of frontal lobe damage on interference effects 49 50 in working memory. Cognitive, Affective, & Behavioral Neuroscience, 2, 109-120. 51 52 Tipper, S. P., & Baylis, G. C. (1987). Individual differences in selective attention: The 53 relation of priming and interference to cognitive failure. Personality and Individual 54 Differences, 8(5), 667-675. 55 56 57 58 59 60 24 http://mc.manuscriptcentral.com/scan Page 25 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Wright, J. C., & Mischel, W. (1987). A conditional approach to dispositional constructs: 4 5 The local predictability of social behavior. Journal of Personality and Social Psychology, 6 53, 1159-1177. 7 8 Wright, J. C., & Mischel, W. (1988). Conditional hedges and the intuitive psychology of 9 traits. 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 25 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 26 of 31

1 2 3 Figure 1 4 5 6 7 8 9 10 Life-long trajectories of high vs. low self-regulation groups 11 12 1.5 13

14 1 15 16 0.5 17 18 For Peer Review 19 0 20 -0.5

21 -regulation index (Z) 22 23 Self -1 24 25 -1.5 26 1968-1974 1984 1993 2003 27 28 Year 29 30 High self-regulators Low self-regulators 31 32 Note: Bars around the mean are standard errors 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan Page 27 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Figure 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 A schematic of the "Suppress" task showing the timeline of events on three typical trials. 44 45 46 A memory set in two colors is presented, followed by a cue to forget half the items in the 47 48 indicated color. Then one of three sorts of probes ensues: a probe that matches an item 49 50 51 in memory, a probe that never appeared on that trial, or a probe that was one of the items 52 53 to be forgotten. 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 28 of 31

1 2 3 Figure 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan Page 29 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Figure 4 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Figure 4. Temporal layout of the emotional go/nogo task for the condition in which happy expressions 43 were the nontargets. 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan Manuscripts submitted to Social Cognitive and Affective Neuroscience Page 30 of 31

1 2 3 Figure 5 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Plot of correlation between nogo performance for happy expressions in adulthood and 50 51 same participants standardized delay score at age 4 years. 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan Page 31 of 31 Manuscripts submitted to Social Cognitive and Affective Neuroscience

1 2 3 Figure 6 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/scan