When Is an Adolescent an Adult? Assessing Cognitive Control in Emotional and Nonemotional Contexts PDF

627625 research-article2016 PSSXXX10.1177/0956797615627625Cohen et al.When Is an Adolescent an Adult? Psychological Science OnlineFirst, published on February 24, 2016 as doi:10...

627625 research-article2016 PSSXXX10.1177/0956797615627625Cohen et al.When Is an Adolescent an Adult? Psychological Science OnlineFirst, published on February 24, 2016 as doi:10.1177/0956797615627625 Research Article Psychological Science When Is an Adolescent an Adult? 1 ­–14 © The Author(s) 2016 Reprints and permissions: Assessing Cognitive Control in Emotional sagepub.com/journalsPermissions.nav DOI: 10.1177/0956797615627625 and Nonemotional Contexts pss.sagepub.com Alexandra O. Cohen1, Kaitlyn Breiner2, Laurence Steinberg3, Richard J. Bonnie4, Elizabeth S. Scott5, Kim A. Taylor-Thompson6, Marc D. Rudolph7, Jason Chein3, Jennifer A. Richeson8,9, Aaron S. Heller10, Melanie R. Silverman1, Danielle V. Dellarco1, Damien A. Fair7, Adriana Galván2, and B. J. Casey1 1 Department of Psychiatry, Sackler Institute for Developmental Psychobiology, Weill Cornell Medical College; 2 Department of Psychology, University of California, Los Angeles; 3Department of Psychology, Temple University; 4 University of Virginia School of Law, University of Virginia; 5Columbia Law School, Columbia University; 6 New York University School of Law, New York University; 7Department of Behavioral Neuroscience and Psychiatry, Oregon Health & Science University; 8Department of Psychology, Northwestern University; 9 Institute for Policy Research, Northwestern University; and 10Department of Psychology, University of Miami Abstract An individual is typically considered an adult at age 18, although the age of adulthood varies for different legal and social policies. A key question is how cognitive capacities relevant to these policies change with development. The current study used an emotional go/no-go paradigm and functional neuroimaging to assess cognitive control under sustained states of negative and positive arousal in a community sample of one hundred ten 13- to 25-year-olds from New York City and Los Angeles. The results showed diminished cognitive performance under brief and prolonged negative emotional arousal in 18- to 21-year-olds relative to adults over 21. This reduction in performance was paralleled by decreased activity in fronto-parietal circuitry, implicated in cognitive control, and increased sustained activity in the ventromedial prefrontal cortex, involved in emotional processes. The findings suggest a developmental shift in cognitive capacity in emotional situations that coincides with dynamic changes in prefrontal circuitry. These findings may inform age-related social policies. Keywords adolescence, cognitive control, development, emotion, fMRI, legal policy, young adult Received 8/6/15; Revision accepted 12/28/15 Definitions of adulthood in the United States differ accord- to have reached maturity. Extant studies suggest that this ing to state law and policy. Although most states set the may vary depending on the context in which adolescents age of majority at 18, the legal age for purchasing alcohol are assessed. In the current study, we compared the devel- is 21 (Institute of Medicine & National Research Council, opment of cognitive control in neutral and emotionally 2014), and the minimum age for criminal prosecution is arousing situations because the latter seem highly relevant 14 or younger in most states (Taylor-Thompson, 2014). In to many policies relating to definitions of adulthood. scientific studies, 18 is often used as the cutoff for adult- hood even though government research policies, until recently, considered individuals under 21 to be minors. Corresponding Author: B. J. Casey, Weill Cornell Medical College, Sackler Institute for Thus, the legal definition of adulthood is fluid and impre- Developmental Psychobiology, 1300 York Ave., Box 140, New York, cise. One consideration in defining adulthood is when NY 10065 behavior, and the underlying neural circuitry, can be said E-mail: [email protected] Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 2 Cohen et al. Although a large developmental literature shows that such as those related to criminal responsibility and adolescents’ speed and accuracy on simple cognitive accountability. Prior research examining motivational and tasks can resemble adults’ (Luna, Marek, Larsen, Tervo- social influences on cognitive capacities in young adults Clemmens, & Chahal, 2015), mounting evidence suggests has used varying age ranges and experimental manipula- that contextual factors influence performance differen- tions that have produced mixed results (Chein et al., tially as a function of age. Studies show that adolescence, 2011; Cohen-Gilbert et al., 2014; Silva, Shulman, Chein, & typically defined as ages 13 through 17, is a time of Steinberg, 2015; Steinberg et al., 2009). We attempted to heightened sensitivity to motivational, social, and emo- control for several of these variables by testing the impact tional information (Casey, 2015; Steinberg, 2010). Specifi- of both brief and sustained positive and negative emo- cally, during adolescence, cognitive-control capacities tional states on cognitive control, using predefined age and decision making appear to be especially influenced groups as well as age as a continuous variable. We by incentives (Galvan et al., 2006; Geier, Terwilliger, Tes- hypothesized that there would be a developmental shift lovich, Velanova, & Luna, 2010; Somerville, Hare, & in cognitive control in emotional situations that would Casey, 2011; Van Leijenhorst et al., 2010), threats (Cohen- correspond to dynamic changes in prefrontal circuitry. Gilbert & Thomas, 2013; Dreyfuss et al., 2014; Grose- Specifically, we predicted that young adults 18 to 21 years Fifer, Rodrigues, Hoover, & Zottoli, 2013; Hare et al., old would differ from adults over age 21 in cognitive 2008), and peers (Chein, Albert, O’Brien, Uckert, & Stein- control in emotionally arousing conditions (as teens do) berg, 2011; Gardner & Steinberg, 2005). Behavioral regu- but not in neutral conditions. lation in response to these inputs has been shown to rely on prefrontal circuitry (Dreyfuss et al., 2014; Hare et al., 2008; Somerville et al., 2011), which shows marked Method change into the early 20s (Gogtay et al., 2004; Sowell Participants et al., 2004). Prominent neurobiological theories of adolescence Participants were 110 individuals from a larger sample of suggest that dynamic and asymmetric trajectories in 147 healthy, right-handed 13- to 25-year-olds who under- structural and functional development of limbic and pre- went functional MRI (fMRI) while performing an adapted frontal circuitry are implicated in motivated behavior and emotional go/no-go task (Hare et al., 2008) under sus- its control, respectively, and may lead to a propensity tained emotional states of threat and excitement and toward risky and impulsive actions (Casey, 2015; Casey, under nonemotional states (Cohen et al., 2016). Data Getz, & Galvan, 2008; Ernst, Pine, & Hardin, 2006; Mills, from 5 participants were excluded because of their poor Goddings, Clasen, Giedd, & Blakemore, 2014; Steinberg, overall performance (> 2 SD below the group’s average 2010). Phylogenetically older brain regions, such as sub- performance as measured by d′). Data from 14 partici- cortical limbic regions, show nonlinear developmental pants were excluded because of excessive head motion changes and appear to be functionally sensitized during (more than 10% of time points within a run censored adolescence (Galvan et al., 2006; Hare et al., 2008; Raz- because of translational motion > 1.56 mm, or half a nahan et al., 2014), whereas development of prefrontal voxel, or rotational motion > 1°), and data from 18 par- cortex (PFC) exhibits a roughly linear trajectory (Galvan ticipants were excluded because of technical problems et al., 2006; Gogtay et al., 2004; Sowell et al., 2004). Rest- that led to errors in coding and recording of behavioral ing-state functional-connectivity data show prolonged data in the scanner. A total of 110 usable scans were development of long-range cortical connectivity that included in the final analyses reported here (41 teens— does not stabilize until the 20s (Dosenbach et al., 2011; 23 females and 18 males, ages 13–17 years, M = 16.19, Fair et al., 2009). Together, these results suggest contin- SD = 1.20; 35 young adults—17 females and 18 males, ued refinement of brain circuitry, particularly prefrontal ages 18–21 years, M = 19.88, SD = 1.09; 34 adults—17 cortical circuitry, into young adulthood, but the behav- females and 17 males, ages 22–25 years, M = 24.08, SD = ioral implications of this protracted brain development 1.04). Portions of the data from 38 adults in this sample remain unclear. are included in a separate report (Cohen et al., 2016) The current study compared the development of cog- focusing on different experimental questions. nitive control under brief and prolonged states of emo- Participants were a diverse community sample tional arousal and nonemotional states. We focused on recruited from New York City and Los Angeles as part of the 18-to-21 age range given the protracted development an ongoing multisite project. They self-identified as Cau- of prefrontal circuitry and the particular legal and social casian (32.7%), African American (27.3%), Hispanic relevance of this age group. Our key premise was that (24.6%), Asian (12.7%) and “other” (2.7%). The recruit- responses in emotional situations would provide insight ment target for this portion of the study was 125 partici- on cognitive capacities relevant to social and legal policy, pants, in anticipation of 20% attrition due to excessive Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 When Is an Adolescent an Adult? 3 head motion, poor performance, or technical issues. reports: Cohen et al., 2016; Dreyfuss et al., 2014; Hare Because of exclusions due to poor task performance and et al., 2008; Somerville et al., 2011). Participants practiced technical issues in the scanner environment, 22 addi- the task prior to entering the scanner, so that they under- tional participants were run. Participants reported no use stood the instructions and conditions. of psychotropic medications or past diagnoses of or treat- We included both blocks with a sustained state of ment for psychiatric or neurological disorders. Adults and threat and blocks with a sustained state of excitement in parents provided informed written consent, and minors order to dissociate effects of arousal and effects of provided assent. The institutional review board at each valence. Threat was induced by telling participants that site approved the study. they might experience an unpredictable aversive auditory stimulus. Excitement was induced by telling participants that they had a chance of winning up to $100. Partici- Experimental task pants were instructed that the probability of an event Participants completed a modified emotional go/no-go occurring, the volume of the noise, and the amount of paradigm (Hare et al., 2008) called the Cognitive Control money won would not be tied to their performance, but Under Emotion (CCUE) task (Cohen et al., 2016). In this rather would be determined by the computer. They were task, happy, fearful, and calm emotional expressions (Fig. also told that events of a given type would occur ran- 1a) are presented as targets, which participants are domly, only when the background screen was a particu- instructed to respond to (go trials), and nontargets, which lar color (blue for one event and purple for the other). In participants are instructed not to respond to (no-go tri- reality, each participant heard the noise once and won als). The task is performed in blocks of sustained antici- $20 once over the course of the task, and these events pation of a negative event (aversive sound), a positive occurred in a pseudorandomized order. Each event event (winning up to $100), and no event; each type of always occurred near the end of an experimental run, so block is denoted with a different background color on that these time points could be eliminated from the anal- the screen (Fig. 1b). (Further descriptions of this task and yses. During blocks of a sustained neutral state (depicted task-related neural activations are available in other with a yellow background), participants were told there a b Example Block: “Only press to calm faces” Neutral State Negative State Positive State $$ Time Fig. 1. The Cognitive Control Under Emotion (CCUE) paradigm (from Cohen et al., 2016): (a) examples of the fearful, happy, and calm faces used as cues and (b) schematic of one run of the task. In a given run, participants were instructed to respond to one type of cue (go trials) and not to respond to another (no-go trials). These cues were presented within blocks of sustained negative emotion (anticipation of an unpredict- able aversive noise), positive emotion (anticipation of an unpredictable monetary reward), and neutral emotion (no event anticipated); the block type was indicated by the background color of the screen (yellow, blue, or purple). Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 4 Cohen et al. was no chance of either event occurring as they per- Collection of debriefing data from 2 of the 110 subjects formed the task. Each state (75-s duration) was induced was accidently omitted. twice during each run. Data were acquired in six 8-min 2-s runs (total of 48 min fMRI data acquisition 12 s). Each run consisted of a unique combination of the emotional expressions that served as go and no-go cues Whole-brain fMRI data were acquired using Siemens (calm–go/fearful–no go, calm–go/happy–no go, happy– Magnetom Trio 3.0-T scanners located at the Citigroup go/fearful–no go, happy–go/calm–no go, fearful–go/calm– Biomedical Imaging Center at Weill Cornell Medical Col- no go, fearful–go/happy–no go), in a mixed-block event- lege or at the Staglin Center for Cognitive Neuroscience related design. Run orders were pseudocounterbalanced, at the University of California, Los Angeles. Scanning and pairing of the background color and emotional state parameters were identical at the two data-collection sites. was counterbalanced. Before each run, participants were A high-resolution, T1-weighted magnetization-prepared told which type of emotional expression was the target and rapid-acquisition gradient-echo (MPRAGE) sequence reminded of the meaning of each colored background. We scan was acquired using Biomedical Informatics Research then asked participants a series of four questions to be sure Network ( Jovicich et al., 2006) optimized sequences with they were aware of each of these contingencies. On each the following parameters: repetition time (TR) = 2,170 trial, a face appeared for 500 ms; the intertrial interval was ms, echo time (TE) = 4.33 ms, 256-mm field of view jittered (2–7 s). A total of 114 trials were presented in each (FOV), 160 sagittal slices with a thickness of 1.2 mm. run, in a pseudorandomized order (84 go trials and 30 Functional images were acquired using T2*-sensitive no-go trials). For each emotional state, we acquired data on echo planar pulse sequences covering the full brain. a total of 168 go trials and 60 no-go trials. Thirty-eight 4-mm-thick axial slices were acquired per 2,500-ms TR (TE = 30 ms, FOV = 200 mm, flip angle = Behavioral and psychophysiological 90°, 3.1- × 3.1- × 4.0-mm voxels). data acquisition Behavioral data analysis Participants completed a final screening for MRI safety before being positioned in the scanner, with a five-but- Behavioral data were analyzed for accuracy using the ton (New York) or four-button (Los Angeles) MR-com- sensitivity index d′, which incorporates the rates of both patible button box. The experimental task was presented hits and false alarms (Macmillan & Creelman, 2004). We using E-Prime 1.0 (New York) or 2.0 (Los Angeles; Psy- calculated d′ by subtracting the normalized false alarm chology Software Tools, Inc., http://www.pstnet.com) rate from normalized accuracy on go trials. Behavioral and was projected onto a flat screen mounted in the data, stimulus timing, and emotional-state timing were scanner bore. Participants viewed the screen via a mirror extracted and calculated using MATLAB and Statistics mounted on a 12-channel head coil. Skin conductance Toolbox Release 2013b (The MathWorks, Natick, MA). All response (SCR) was acquired using disposable, isotonic statistical analyses of the behavioral data were conducted gel electrodes, which were attached to the first and sec- using R (Release 3.1.0; R Core Team, 2014). We tested for ond fingers of the left hand between the first and second age-related differences in performance (d′) using analysis phalanges. The electrode cables were grounded through of variance (ANOVA) models that included sex and scan- a radio-frequency filter panel. During fMRI scanning, the ning site as between-subjects variables. To investigate skin conductance signal was recorded (200-Hz sam- performance responding to the emotional cues, without pling) and amplified using a Biopac recording system effects of emotional state, we tested for main effects of and AcqKnowledge 4.0 software. E-Prime software was age group on performance with each cue type in the used to indicate the onset and offset of the emotional neutral state. To investigate performance during the emo- states during the task. SCR data were acquired from all tional states, controlling for effects of the emotional cues, the participants. we tested for main effects of age group on performance After exiting the scanner, participants were asked responding to the calm face cues in each emotional state. debriefing questions about the believability of task con- A Bonferroni-adjusted alpha of less than.01 was used to ditions. Specifically, they were asked how much they correct for multiple comparisons in determining the sta- expected to win money or hear the noise during the tistical significance of these ANOVA results. Bonferroni- blocks in which the background color signaled the pos- corrected post hoc t tests were used to determine the sibility of those events (e.g., “Did you expect to win statistical significance of differences between age groups. money more during the purple blocks than the blue or Linear and quadratic models were also fitted to each yellow blocks?”). Each question was answered using a dependent variable, with age modeled continuously. As 7-point Likert scale (1 = not at all, 7 = very much). in the age-group analyses, we used a Bonferroni-adjusted Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 When Is an Adolescent an Adult? 5 alpha of less than.01 to determine statistical significance. emotional states (i.e., the threat, excitement, and neutral All analyses were performed on the data from the 110 sustained states), an additional regressor corresponding subjects with usable imaging and behavioral data. to trials with incorrect responses (both go and no-go tri- We examined responses to the debriefing questions als), and 6 motion estimation parameters. Baseline trends and the SCR data to assess the efficacy of our emotional- were estimated to capture shifts in signal change. Activa- state manipulation. A 1-Hz filter was applied to the raw tion in response to the face cues was modeled with a SCR data. Data were smoothed for each subject. Six sub- three-parameter gamma hemodynamic-response func- jects had no SCR data because of technical difficulties in tion (HRF); activation during the sustained states was the collection of these data, and 29 of the remaining 104 modeled using a single-parameter block HRF. Time participants had no discernible variation in SCR across points with motion greater than half a voxel (1.56 mm) the experiment or individual runs and so were removed were censored, along with the preceding and following from the SCR analyses. SCR slope was extracted for each time points. emotional-state block within each run and was z-scored Individual-level regression coefficients for the 110 par- within subjects to account for individual differences in ticipants were submitted to group linear mixed-effects SCR. Each individual’s average slope was calculated for (LME) analyses using the AFNI 3dLME function (Chen, each emotional state (excitement, threat, and neutral). Saad, Britton, Pine, & Cox, 2013), which is robust to small Change in skin conductance was computed as the differ- amounts of missing data. All group-level analyses included ence between average SCR slope in an aroused state a random intercept for each participant and included sex (excitement or threat) and average SCR slope in the neu- and scanning site as between-subjects variables. Separate tral state. Given the directionality of our hypotheses with models were used to assess effects of transient cues (mod- respect to these validation measures, we performed one- eled as brief events) and sustained states (modeled as tailed one-sample Student’s t tests to test whether prolonged blocks) on brain activity. The first group-level responses to debriefing questions were significantly dif- LME model assessed effects of the transient cues (fearful, ferent from 1 (the lowest value on the 7-point scale) and happy, and calm faces) on go and no-go trials. The sec- whether SCR differences were significantly different from ond group-level LME model assessed effects of the sus- zero. tained states (threat, excitement, and neutral). Age-group contrasts (general linear tests) were specified within each model to directly probe the neural correlates of behav- fMRI data analysis ioral findings. Two additional models assessed effects of Image processing. Functional imaging data were pre- the emotional cues and emotional states as a function of processed and analyzed using Analysis of Functional exact age as a continuous variable (i.e., interactions of NeuroImages (AFNI) software (Cox, 1996). Preprocessing emotional cues or states with exact age). of functional scans included correction for slice-time In group whole-brain analyses, individual voxels were acquisition using sinc interpolation, volume registration thresholded at a p value of.005; the cluster-size threshold using a 6-parameter rigid-body transformation to account was a p value of.05 after correction for multiple compari- for head motion, and normalization to the Montreal Neu- sons (performed using Monte Carlo simulation via the rological Institute (MNI) 152 1-mm T1 template using a 3dClustSim program in AFNI). For the threat condition, 12-parameter affine transformation and nonlinear trans- given our a priori hypotheses regarding differences in formations (AFNI 3dQWarp function). Data were resam- prefrontal activation, we used an anatomical region of pled to 3-mm isotropic voxels and were smoothed using interest (ROI) for the PFC (obtained from the Harvard- a full-width/half-maximum Gaussian kernel of 6 mm. Sig- Oxford probabilistic atlas in FSL; http://fsl.fmrib.ox.ac nal intensity of each voxel time series was normalized to.uk/fsl/fslwiki/; Smith et al., 2004). Similar to the PFC ROI percentage signal change. in previous studies (e.g., Foerde, Steinglass, Shohamy, & Walsh, 2015), this ROI combined the frontal pole, supe- Image analysis. A general linear model (GLM) was rior frontal gyrus, middle frontal gyrus, inferior frontal created for each participant to estimate activation in gyrus (triangularis and opercularis), frontal medial cor- response to the emotional cues and sustained-emotional- tex, subcallosal cortex, paracingulate gyrus, cingulate state blocks. To disentangle the neural responses to the gyrus anterior division, and frontal orbital cortex bilater- cues and to the sustained states, which were presented ally; a threshold of 50% probability was used for all sub- simultaneously, we included 16 regressors in each par- regions within the PFC. A p value of.005 was used as the ticipant’s GLM: 6 task regressors for correct responses to threshold for individual voxels (p <.05 after PFC volume the emotional cues (fearful, happy, or calm faces on go correction for multiple comparisons was performed using trials and fearful, happy, or calm faces on no-go trials), 3 Monte Carlo simulation via the 3dClustSim program in task regressors modeling the longer (30-TR) sustained AFNI). Regression coefficients for individual participants Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 6 Cohen et al. were extracted from regions with significant effects and Results were tested for brain-behavior correlations in R (Release 3.1.0; R Core Team, 2014). Behavioral results Validation of the paradigm. Responses to the debrief- Psychophysiological interaction analysis. General- ing questions and SCR slope differences were tested inde- ized psychophysiological interaction (gPPI) analyses pendently, so we used a Bonferroni-adjusted alpha of less (McLaren, Ries, Xu, & Johnson, 2012) were conducted in than.025 in our validation tests. These validation mea- AFNI to examine task-dependent connectivity across the sures were collapsed across age. Participants expected whole brain. Seed regions were the two PFC regions both the money, t(107) = 24.49, p <.001, d = 3.35, and identified as having age-group effects. The gPPI analyses loud noise, t(107) = 31.87, p <.001, d = 4.36, to occur were carried out by removing sources of noise and arti- during the blocks in which they were led to anticipate fact, deconvolving the neural signal, extracting the func- these possibilities (Fig. 2a). tional time course within the seed regions (5-mm spheres Participants’ mean SCR difference scores (arousal state around peak activation), and convolving the time-course minus neutral state) were positive for both the excite- data with task timings and the canonical HRF (McLaren ment condition, t(74) = 1.92, p =.029, d = 0.32, and the et al., 2012). The 16-regressor GLM used for the individ- threat condition, t(74) = 1.65, p =.051, d = 0.27 (Fig. 2b). ual-level image analyses was implemented, but for the SCR difference scores for the excitement and threat con- gPPI analyses, these models also included regressors for ditions were not significantly different from each other, the seed time course and each Time Course × Task Con- t(74) = 0.26, p >.250, d = 0.04. These validation results dition interaction, for a total of 27 regressors. The group- replicate previous results for adults performing this same level LME model (controlling for sex and scanning site) task (Cohen et al., 2016). was used to test the specific age-group contrasts. Specifi- cally, group-level LME models tested the effects of tran- Main effects of age for each type of emotional cue. sient cues (fearful, happy, and calm faces) and sustained In the neutral-state blocks, there were significant main states (threat, excitement, and neutral) separately. Age- effects of age group on performance in response to fear- group contrasts (general linear tests) were specified ful cues, F(2, 98) = 11.11, p <.001, ηp2 =.16; happy cues, within each model. The models used a p threshold of.05, F(2, 98) = 10.90, p <.001, ηp2 =.15; and calm cues, F(2, corrected for multiple comparisons at the whole-brain 98) = 7.81, p <.001, ηp2 =.10 (see Fig. 3a and Behavioral level using 3dClustSim, as described previously. Results and Figs. S1a and S2a in the Supplemental a 7 b 0.3 6 SCR Slope Difference (From Neutral) 5 Anticipation of Event 0.2 4 3 0.1 2 1 0 0.0 Excitement Threat Excitement Threat Emotional State Emotional State Fig. 2. Validation of the Cognitive Control Under Emotion (CCUE) paradigm: (a) participants’ mean ratings of how likely they thought they were to win money and hear a loud sound in the excitement and threat blocks, respectively, and (b) participants’ mean skin conductance response (SCR) during those blocks relative to the neutral blocks. Error bars represent ±1 SE. Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 When Is an Adolescent an Adult? 7 Material available online). Post hoc t tests revealed that condition. We found that both linear and quadratic func- teens and young adults showed diminished performance tions significantly fit the data in both the threat condi- relative to adults in response to fearful cues—teens ver- tion—linear function: adjusted R2 =.23, p <.001, F(1, sus adults: t(62.39) = 4.08, p <.001, d = 0.95; young 108) = 34.08; quadratic function: adjusted R2 =.24, p < adults versus adults: t(64.82) = 3.33, p =.0019, d = 0.80;.001, F(2, 107) = 17.78—and the excitement condition— teens versus young adults: t(70.09) = 0.61, p >.250, d = linear function: adjusted R2 =.13, p <.001, F(1, 108) = 0.14. However, young adults and adults showed enhanced 17.09; quadratic function: adjusted R2 =.13, p <.001, F(2, performance relative to teens in response to happy 107) = 9.21 (see Fig. 4). cues—teens versus adults: t(71.15) = 4.14, p <.001, d = 0.96; young adults versus adults: t(65.77) = 1.79, p >.250, Imaging results d = 0.43; teens versus young adults: t(73.96) = 2.55, p =.042, d = 0.59—and only teens and adults differed signifi- To probe the neural correlates of the observed behav- cantly in their performance with calm cues—teens versus ioral effects, we examined blood-oxygen-level-depen- adults: t(64.05) = 3.54, p =.001, d = 0.82; young adults dent (BOLD) activity in the age-group contrasts specified versus adults: t(64.60) = 1.56, p >.250, d = 0.38; teens in the group LME models for each emotional cue and versus young adults: t(71.54) = 2.14, p =.140, d = 0.49. state. Specifically, general linear tests comparing brain We also examined effects of age as a continuous vari- activity (relative to implicit baseline, i.e., overall baseline able, fitting both linear and quadratic functions to perfor- brain activity) of teens and young adults with that of mance with each cue type in the neutral-state blocks. adults were specified for the relevant conditions. We also Linear and quadratic functions significantly fit the data examined BOLD activity using group LME models in for all three cue types—fearful cues, linear: adjusted R2 = which age was modeled continuously. For these models,.12, p <.001, F(1, 108) = 15.68; fearful cues, quadratic: in the absence of any specific general linear tests, we adjusted R2 =.13, p <.001, F(2, 107) = 9.23; happy cues, examined activation maps showing the interactions of linear: adjusted R2 =.14, p <.001, F(1, 108) = 18.33; age with type of emotional cue and emotional-state happy cues, quadratic: adjusted R2 =.13, p <.001, F(2, condition. 107) = 9.25; calm cues, linear: adjusted R2 =.10, p <.001, F(1, 108) = 13.6; calm cues, quadratic: adjusted R2 =.095, Effects of emotional cues as a function of age. Two p =.002, F(2, 107) = 6.75. However, the fit of the qua- clusters survived whole-brain correction in the age-group dratic function completely overlapped with the fit of the analyses of response to fearful cues, showing less activity linear function for the calm cues (see Fig. 4 for perfor- in teens and young adults than in adults: right dlPFC (x = mance in response to calm cues in all three sustained −41.5, y = −9.5, z = 36.5, 47 voxels; Z = −4.66, p <.02, emotional states and in response to fearful and happy corrected; Figs. 3b and 3c) and right thalamus (x = −20.5, cues in the neutral-state blocks). y = 23.5, z = 6.5, 57 voxels; Z = −3.88, p <.02, corrected). MR signal change in dlPFC was positively correlated with Main effects of age for each emotional state. There behavioral performance (in the neutral condition) were significant main effects of age group on performance responding to fearful cues across age, r(108) =.203, p = in response to calm cues when participants were in emo-.033 (Fig. 3d), but this correlation did not remain signifi- tionally arousing states of threat, F(2, 98) = 17.57, p <.001, cant when we controlled for age, r(107) =.087, p =.365. ηp2 =.24 (Fig. 5a), and of excitement, F(2, 98) = 8.65, p < A general linear test corresponding to the behavioral.001, ηp2 =.13 (Fig. S1b). Post hoc t tests revealed that, result was performed for happy cues in the group-level although young adults performed better than teens, teens emotional-cue model to compare brain activity (relative and young adults both showed diminished performance to implicit baseline) of teens with that of adults and relative to adults under the state of threat—teens versus young adults in response to happy cues. A single cluster adults: t(60.47) = 5.40, p <.001, d = 1.24; young adults in the right inferior frontal gyrus (x = –32.5, y = –24.5, z = versus adults: t(59.51) = 2.75, p =.014, d = 0.66; teens ver- –11.5, 30 voxels; Z = –4.18, p <.02, corrected) survived sus young adults: t(73.25) = 3.25, p =.014, d = 0.64. In whole-brain correction, showing more activity in teens contrast, only teens and adults’ performance differed sig- than in both young adults and adults (see Imaging Results nificantly under the state of excitement—teens versus and Figs. S2b and S2c in the Supplemental Material). No adults: t(58.52) = 4.28, p <.001, d = 0.98; young adults clusters survived whole-brain correction in the analyses versus adults: t(66.95) = 2.03, p =.087, d = 0.49; teens ver- of activation in response to calm cues, and no interac- sus young adults: t(61.39) = 1.83, p =.213, d = 0.42. tions of emotional cue with age group were observed. We also examined effects of age as a continuous vari- Four clusters in the dorsal anterior cingulate cortex able, fitting both linear and quadratic functions to perfor- (dACC), parietal cortex, and right and left cerebellum sur- mance in response to the calm cues in each emotional-state vived whole-brain correction when we examined the Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016  8 a *** b c 1.5 d 3 ** 3 2 1.0 1 2 d′ x = –42 0 dlPFC 0.5 −1 1 dlPFC Signal Change dlPFC Signal Change −2 Teens Young Adults −3 Adults 0 0.0 Teens Young Adults Adults Teens Young Adults Adults 0 1 2 3 4 5 6 Age Group Age Group d′ e f g 10 10 8 8 x = –3 6 6 4 4 2 2 0 0 Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 −2 −2 dACC Signal Change −4 Teens −4 Teens Parietal Cortex Signal Change −6 Young Adults −6 Young Adults Adults dACC Parietal Adults Cortex 0 1 2 3 4 5 6 0 1 2 3 4 5 6 d′ d′ Fig. 3. Results for the fearful cues. The graph in (a) shows mean performance in response to the brief fearful cues in the neutral-state condition, as indexed by d′, for each of the age groups. The brain image in (b) shows the location of the region in the dorsolateral prefrontal cortex (dlPFC) for which signal-change results are presented. The graphs in (c) and (d) show mean signal change in the dlPFC in response to the fearful cues for each age group and as a function of d′ in the neutral-state condition (separately for each age group), respectively. The brain image in (f) shows the location of the regions in the dorsal anterior cingulate cortex (dACC) and parietal cortex for which mean signal change in response to the fearful cues is graphed as a function of d′ in the neutral-state condition, separately for each age group, in (e) and (g). Error bars represent ±1 SE. Asterisks indicate significant differences (**p <.01, ***p <.001). When Is an Adolescent an Adult? 9 Fearful Cues, Neutral State Happy Cues, Neutral State 6 Female 6 Female 5 Male 5 Male 4 4 3 3 d′ d′ 2 2 Calm Cues, Neutral State 1 1 0 6 Female 0 5 Male 4 14 16 18 20 22 24 14 16 18 20 22 24 3 Age Age d′ 2 Calm Cues, Threat State Calm Cues, Excitement State 1 6 Female 6 Female 0 Male Male 5 5 4 4 14 16 18 20 22 24 3 3 Age d′ d′ 2 2 1 1 0 0 14 16 18 20 22 24 14 16 18 20 22 24 Age Age Fig. 4. Scatterplots showing male and female participants’ performance, as indexed by d′, as a function of age, along with linear and quadratic functions fitted to the data. Results are shown for each kind of emotional cue in the neutral-state condition, as well as for calm cues in the threat and excitement conditions. interaction of age as a continuous variable with type of regions are important for behavioral performance of the emotional cue (see Table S1 in the Supplemental Mate- task. rial). In the two largest regions, the dACC and the parietal cortex (Fig. 3f), activity in response to fearful cues was Effects of emotional states as a function of age. positively correlated with age, r(108) =.196, p =.040, and Although no activations survived whole-brain correc- r(108) =.32, p <.001, respectively. MR signal change in tion for the contrast of age groups in emotional states, response to happy cues was negatively correlated with a single cluster in the ventromedial PFC (vmPFC; x = age in the dACC, r(108) = −.189, p =.048, but not in the 3.5, y = −33.5, z = −17.5, 13 voxels; Z = 3.58, p <.05, parietal cortex, r(108) = −.164, p =.087. Activity in PFC corrected; Fig. 5b) survived PFC volume correction response to calm cues was not significantly correlated for responses in the state of threat. Teens’ and young with age in either of these regions, r(108) = −.088, p = adults’ BOLD activity in the vmPFC during the threat.363, and r(108) = −.079, p =.412, respectively. condition showed a sustained increase relative to We examined whether changes in dACC and parietal adults’ (Fig. 5c). MR signal change in this region in the activity in response to fearful cues were correlated with threat condition was negatively correlated with behav- behavioral performance. In both of these regions, MR ioral performance (in response to the calm cues) in the signal change in response to fearful cues was positively threat condition, r(108) = −.308, p =.001 (Fig. 5d), and correlated with d′ in the neutral-state condition, r(108) = this correlation remained significant even when we.222, p =.020, and r(108) =.359, p <.001, respectively controlled for age, r(107) = −.215, p =.023, and when (Figs. 3e and 3g). Similar patterns were observed even we removed the one extreme outlier, r(107) = −.253, when we controlled for age, r(107) =.166, p =.081, and p =.008. No interactions of emotional state with age r(107) =.277, p =.002. These results suggest that these group were observed. In analyses for the excitement Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016  10 a b c d *** * * x=3 3 0.4 2 1 0.2 0 2 d′ −1 vmPFC 0.0 −2 1 vmPFC Signal Change Teens vmPFC Signal Change −3 −0.2 Young Adults −4 Adults 0 Teens Young Adults Adults Teens Young Adults Adults 0 1 2 3 4 5 6 Age Group Age Group d′ e f g 10 10 x = –6 x = –32 5 5 0 0 Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 −5 −5 Parameter Estimate Parameter Estimate dACC PPC Functional Connectivity Functional Connectivity −10 −10 Teens Young Adults Adults Teens Young Adults Adults Age Group Age Group Fig. 5. Results for the sustained-threat condition. The graph in (a) shows mean performance on calm-cue trials, as indexed by d′, for each of the age groups under the sustained state of threat. The brain image in (b) shows the location of the region in ventromedial prefrontal cortex (vmPFC) for which signal-change results are presented. The graphs in (c) and (d) show mean signal change in the vmPFC in the threat condition for each age group and as a function of d′ on calm-cue trials (separately for each age group), respectively. The brain images in (f) show the location of the regions in the dorsal anterior cingulate cortex (dACC) and posterior parietal cortex (PPC) for which functional coupling with the vmPFC is graphed for the three age groups in (e) and (g), respectively. Error bars represent ±1 SE. The asterisk indicates a significant difference (*p <.05, ***p <.001). When Is an Adolescent an Adult? 11 and neutral-state conditions, no clusters survived et al., 2015; Vincent et al., 2008), and in the dACC and whole-brain or PFC volume correction. parietal cortex. The dlPFC and parietal cortex have recip- A single cluster in the parietal cortex (x = −2.5, y = rocal projections with the dACC, and all three regions 68.5, z = 54.5, 29 voxels; F = 11.90, p <.05, corrected) have been implicated in cognitive control and are coacti- survived whole-brain correction when we examined the vated during cognitive-control tasks (Botvinick, Nystrom, interaction of emotional state and age as a continuous Fissell, Carter, & Cohen, 1999; Platt & Glimcher, 1999; variable (see Fig. S3 in the Supplemental Material). MR Roy, Shohamy, & Wager, 2012). Further, activity in these signal change in this region showed similar positive asso- regions not only was consistently lower in younger par- ciations with age in the threat and excitement conditions, ticipants, but also was positively correlated with task per- r(108) =.181, p =.058, and r(108) =.305, p =.001, respec- formance. Together, these findings are consistent with tively, but not in the neutral-state condition, r(108) =.151, the hypothesis that lower levels of activity within this p =.116. Because BOLD activity in the threat and excite- circuitry in younger individuals reflects diminished cog- ment conditions showed positive correlations with age, nitive control in the face of negative emotional cues that we collapsed the MR signal across these conditions and signal potential threat in the environment. tested for associations between activation in this region Although under sustained states of negative emotional and behavioral performance (in response to the calm arousal (threat), young adults performed better than cues) in these conditions. MR signal change was posi- teens, they performed worse than adults. Teens’ and tively correlated with behavioral performance, r(108) = young adults’ diminished performance relative to adults.209, p =.028 (see Fig. S3), but this correlation did not in the threat condition was paralleled by increased activ- hold when we controlled for age, r(107) =.11, p =.251. ity in the vmPFC. This region has been implicated in vari- ous processes, including self-referential thought and Seed-based functional connectivity with prefrontal integration of affective information, and is a proposed regions in the three age groups. Whole-brain gPPI hub for affective computations and regulation (Roy et al., analyses were performed using the dlPFC and vmPFC 2012). Increased sustained recruitment of the vmPFC regions as seeds. Nine clusters of voxels showing signifi- under threat in teens and young adults may suggest cantly less functional coupling with the vmPFC in teens heightened sensitivity to potential threat, leading to emo- and young adults than in adults across the threat condi- tional interference and diminished cognitive control. This tion were observed (see Table S2 in the Supplemental interpretation is supported in part by our finding of Material). Areas showing this pattern included the dACC decreased functional coupling of the vmPFC with cogni- (Figs. 5e and 5f) and posterior parietal cortex (Figs. 5f tive-control circuitry of the dACC and posterior parietal and 5g). No significant clusters were observed in the age- cortex in the threat condition among teens and young group contrast for fearful cues using the dlPFC seed. adults relative to adults. The negative functional connec- tivity between cognitive and emotional brain regions dur- ing this emotional state may underlie the poorer Discussion performance of the younger age groups. Our findings suggest a developmental shift in cognitive Taken together, these findings suggest that young control in negative emotional situations during young adulthood is a time when cognitive control is still vulner- adulthood that is paralleled by dynamic developmental able to negative emotional influences, in part as a result changes in prefrontal circuitry. Specifically, young adults of continued development of lateral and medial prefron- showed diminished cognitive control under both brief tal circuitry. This temporal developmental shift in cogni- and prolonged negative emotional arousal relative to tive-control capacity in negatively arousing situations slightly older adults, a pattern not observed in neutral or relative to neutral (or positive) situations is consistent positive situations. This behavioral pattern was paralleled with the classic notion of developmental cascades in by altered recruitment of lateral and medial prefrontal brain and behavior (Casey, Galván, & Somerville, 2015; circuitry in young adults and adolescents, a finding con- Masten & Cicchetti, 2010). Accordingly, dynamic brain sistent with structural imaging studies showing protracted changes during late adolescence may enhance receptivity development of prefrontal circuitry (Gogtay et al., 2004; to or processing of emotional inputs in order to facilitate Sowell et al., 2004). meeting changing socioemotional pressures that accom- Teens’ and young adults’ diminished cognitive control pany adulthood (Casey et al., 2015). in response to negative cues was paralleled by their Our findings have potential implications for informing decreased activity in cognitive-control circuitry. When age-related legal and social policies. Developmental find- presented with fearful cues, teens and young adults ings based largely on teens have been referenced in sev- showed less activity than older adults in dlPFC, a region eral U.S. Supreme Court decisions regarding treatment of implicated in affective and cognitive regulation (Silvers juvenile offenders over the past decade, with the Court Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 12 Cohen et al. acknowledging immature cognitive functioning in juve- prolonged, affects this capacity in individuals ages 18 to niles as a mitigating factor in judgments of criminal cul- 21 more than in older individuals. Few studies have pability (Cohen & Casey, 2014; Scott, 2013; Steinberg, examined cognitive capacities under emotional influ- 2013). Scientific research has demonstrated that adoles- ences, and fewer still have taken this approach to study cents show heightened sensitivity to motivational and developmental differences in capacities of potential rele- socioemotional information, which potentially renders vance to legal and social policies. Our findings provide them more vulnerable to poor decision making in these support for consideration of contextual influences on situations, compared with younger and older individuals behavior and brain function, such as the influence of (Chein et al., 2011; Cohen-Gilbert & Thomas, 2013; Drey- emotional arousal, when evaluating appropriate age cut- fuss et al., 2014; Galvan et al., 2006; Grose-Fifer et al., offs for such policies. Although the data in this study do 2013; Hare et al., 2008; Somerville et al., 2011; Steinberg not speak directly to these policy issues, they may inform et al., 2009). The extension of this work to young adults, dialogues about the age of adulthood in a variety of who show diminished cognitive control relative to slightly social and policy contexts. older adults in negative emotional situations, may have implications for legal policy. This is not to suggest that Author Contributions teens and young adults should not be held accountable B. J. Casey, A. Galván, and L. Steinberg developed the study for their actions, but rather, the boundaries of juvenile- concept. A. O. Cohen, K. Breiner, A. S. Heller, M. R. Silverman, court jurisdiction, criminal-court sentencing, and punish- and D. V. Dellarco collected the data and performed data analy- ment may be informed by developmental considerations sis under the supervision of B. J. Casey and A. Galván. All (Bonnie & Scott, 2013). authors contributed to interpretation of the data. A. O. Cohen The implications of our findings must be considered and B. J. Casey drafted the manuscript, and all authors provided within the limitations of the study. First, behaviors were critical revisions and approved the final version of the manu- measured within a controlled research setting. Although script for submission. the emotionally arousing conditions may be relevant to emotional arousal in the real world, they were limited to Acknowledgments experimentally manipulated emotional conditions that did We gratefully acknowledge the assistance of Doug Ballon, not capture the complex real-world situations in which Kristine Caudle, Jonathan Dyke, Hillary Raab, Ahrareh Rahdar, individuals typically make decisions. Second, the sample, and the Citigroup Biomedical Imaging Center at Weill Cornell although community based and representative of the racial Medical College. We thank the anonymous reviewers for their and ethnic distribution in Los Angeles and New York City, constructive feedback. was relatively small, with 110 participants 13 to 25 years of age; replication of these findings is warranted. Declaration of Conflicting Interests Prior research examining motivational and social influ- B. J. Casey and L. Steinberg serve as paid consultants to the ences on cognitive capacities in young adults as a unique John D. and Catherine T. MacArthur Foundation. The authors age group has produced mixed results (Chein et al., 2011; declared that they have no other conflicts of interest with Cohen-Gilbert et al., 2014; Silva et al., 2015; Steinberg respect to their authorship or the publication of this article. et al., 2009). The present and previous findings suggest that teens’ and young adults’ cognitive capacities may be Funding affected differently by various situations. For instance, This work was supported by a National Science Foundation although negative emotional arousal may diminish cogni- Graduate Research Fellowship (to A. O. Cohen). Preparation of tive control in both teens and young adults, positive this article was supported by a grant from the John D. and emotional arousal and the presence of peers may not Catherine T. MacArthur Foundation to Vanderbilt University. Its influence young adults as strongly as teens (Chein et al., contents reflect the views of the authors, and do not necessarily 2011). Identifying specific situations in which the behav- represent the official views of either the John D. and Catherine ior of young adults may differ from that of slightly older T. MacArthur Foundation or the MacArthur Foundation Research adults will be important in informing potential changes Network on Law and Neuroscience (www.lawneuro.org). to existing policies and laws. Moreover, further examina- tion of changes in brain structure, activity, and connectiv- Supplemental Material ity during this developmental period may provide clearer Additional supporting information can be found at http://pss insights into why and when researchers may or may not.sagepub.com/content/by/supplemental-data observe group-level behavioral changes in young adults. We examined the influence of emotional arousal on Open Practices cognitive control from early adolescence through the mid The data reported here are part of an ongoing multisite project. 20s and found that negative emotional arousal, brief or An optimized version of the Cognitive Control Under Emotion Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 When Is an Adolescent an Adult? 13 (CCUE) task (both behavioral and jittered for use in the scan- Fair, D. A., Cohen, A. L., Power, J. D., Dosenbach, N. U., ner) will soon be made available at https://www.sacklerinstitute Church, J. A., Miezin, F. M.,... Petersen, S. E. (2009)..org/cornell/assays_and_tools/. The complete Open Practices Functional brain networks develop from a “local to dis- Disclosure for this article can be found at http://pss.sagepub tributed” organization. PLoS Computational Biology, 5(5),.com/content/by/supplemental-data. Article 381. doi:10.1371/journal.pcbi.1000381 Foerde, K., Steinglass, J. E., Shohamy, D., & Walsh, B. T. (2015). Neural mechanisms supporting maladaptive food choices References in anorexia nervosa. Nature Neuroscience, 18, 1571–1573. Bonnie, R. J., & Scott, E. S. (2013). The teenage brain: doi:10.1038/nn.4136 Adolescent brain research and the law. Current Directions Galvan, A., Hare, T. A., Parra, C. E., Penn, J., Voss, H., Glover, in Psychological Science, 22, 158–161. G., & Casey, B. J. (2006). Earlier development of the accum- Botvinick, M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen, bens relative to orbitofrontal cortex might underlie risk- J. D. (1999). Conflict monitoring versus selection-for-action taking behavior in adolescents. The Journal of Neuroscience, in anterior cingulate cortex. Nature, 402, 179–181. 26, 6885–6892. doi:10.1523/JNEUROSCI.1062-06.2006 Casey, B. J. (2015). Beyond simple models of self-control to Gardner, M., & Steinberg, L. (2005). Peer influence on risk taking, circuit-based accounts of adolescent behavior. Annual risk preference, and risky decision making in adolescence Review of Psychology, 66, 295–319. and adulthood: An experimental study. Developmental Casey, B J, Galván, A., & Somerville, L. (2015). Beyond simple Psychology, 41, 625–635. models of adolescence to an integrated circuit-based account: Geier, C. F., Terwilliger, R., Teslovich, T., Velanova, K., & Luna, A commentary. Developmental Cognitive Neuroscience. B. (2010). Immaturities in reward processing and its influ- Advance online publication. doi:10.1016/j.dcn.2015.12.006 ence on inhibitory control in adolescence. Cerebral Cortex, Casey, B. J., Getz, S., & Galvan, A. (2008). The adolescent brain. 20, 1613–1629. Developmental Review, 28, 62–77. Gogtay, N., Giedd, J. N., Lusk, L., Hayashi, K. M., Greenstein, Chein, J., Albert, D., O’Brien, L., Uckert, K., & Steinberg, L. D., Vaituzis, A. C.,... Thompson, P. M. (2004). Dynamic (2011). Peers increase adolescent risk taking by enhanc- mapping of human cortical development during child- ing activity in the brain’s reward circuitry. Developmental hood through early adulthood. Proceedings of the National Science, 14, F1–F10. Academy of Sciences, USA, 101, 8174–8179. Chen, G., Saad, Z. S., Britton, J. C., Pine, D. S., & Cox, R. W. Grose-Fifer, J., Rodrigues, A., Hoover, S., & Zottoli, T. (2013). (2013). Linear mixed-effects modeling approach to FMRI Attentional capture by emotional faces in adolescence. group analysis. NeuroImage, 73, 176–190. Advances in Cognitive Psychology, 9, 81–91. Cohen, A. O., & Casey, B. J. (2014). Rewiring juvenile justice: Hare, T. A., Tottenham, N., Galvan, A., Voss, H. U., Glover, The intersection of developmental neuroscience and legal G. H., & Casey, B. J. (2008). Biological substrates of policy. Trends in Cognitive Sciences, 18, 63–65. emotional reactivity and regulation in adolescence dur- Cohen, A. O., Dellarco, D. V., Breiner, K., Helion, C., Rahdar, ing an emotional go-nogo task. Biological Psychiatry, A., Pedersen, G.,... Casey, B. J. (2016). The impact of 63, 927–934. emotional states on cognitive control circuitry and function. Institute of Medicine & National Research Council. (2014). Journal of Cognitive Neuroscience 28, 446–459. Investing in the health and well-being of young adults (R. J. Cohen-Gilbert, J. E., Killgore, W. D. S., White, C. N., Schwab, Bonnie, C. Stroud, & H. Breiner, Eds.). Washington, DC: Z. J., Crowley, D. J., Covell, M. J.,... Silveri, M. M. (2014). The National Academies Press. Differential influence of safe versus threatening facial Jovicich, J., Czanner, S., Greve, D., Haley, E., van der Kouwe, expressions on decision-making during an inhibitory con- A., Gollub, R.,... Dale, A. (2006). Reliability in multi- trol task in adolescence and adulthood. Developmental site structural MRI studies: Effects of gradient non-linearity Science, 17, 212–223. correction on phantom and human data. NeuroImage, 30, Cohen-Gilbert, J. E., & Thomas, K. M. (2013). Inhibitory control 436–443. during emotional distraction across adolescence and early Luna, B., Marek, S., Larsen, B., Tervo-Clemmens, B., & adulthood. Child Development, 84, 1954–1966. Chahal, R. (2015). An integrative model of the maturation Cox, R. W. (1996). AFNI: Software for analysis and visualization of cognitive control. Annual Review of Neuroscience, 38, of functional magnetic resonance neuroimages. Computers 151–170. and Biomedical Research, 29, 162–173. Macmillan, N. A., & Creelman, C. D. (2004). Detection theory: A Dosenbach, N. U. F., Nardos, B., Cohen, A. L., Fair, D. A., user’s guide (2nd ed.). Mahwah, NJ: Erlbaum. Power, D., Church, J. A.,... Schlaggar, B. L. (2011). Masten, A. S., & Cicchetti, D. (2010). Developmental cascades. Prediction of individual brain maturity using fMRI. Science, Development and Psychopathology, 22, 491–495. 329, 1358–1361. McLaren, D. G., Ries, M. L., Xu, G., & Johnson, S. C. (2012). Dreyfuss, M., Caudle, K., Drysdale, A. T., Johnston, N. E., A generalized form of context-dependent psychophysi- Cohen, A. O., Somerville, L. H.,... Casey, B. J. (2014). ological interactions (gPPI): A comparison to standard Teens impulsively react rather than retreat from threat. approaches. NeuroImage, 61, 1277–1286. Developmental Neuroscience, 36, 220–227. Mills, K. L., Goddings, A.-L., Clasen, L. S., Giedd, J. N., & Ernst, M., Pine, D. S., & Hardin, M. (2006). Triadic model of Blakemore, S.-J. (2014). The developmental mismatch in the neurobiology of motivated behavior in adolescence. structural brain maturation during adolescence. Develop- Psychological Medicine, 36, 299–312. mental Neuroscience, 36, 147–160. Downloaded from pss.sagepub.com at UNIV CALIFORNIA IRVINE on May 2, 2016 14 Cohen et al. Platt, M. L., & Glimcher, P. W. (1999). Neural correlates of deci- Somerville, L. H., Hare, T., & Casey, B. J. (2011). Frontostriatal sion variables in parietal cortex. Nature, 400, 233–238. maturation predicts cognitive control failure to appetitive R Core Team. (2014). R: A language and environment for statis- cues in adolescents. Journal of Cognitive Neuroscience, 23, tical computing. Retrieved from https://cran.r-project.org/ 2123–2134. doc/manuals/r-release/fullrefman.pdf Sowell, E. R., Thompson, P. M., Leonard, C. M., Welcome, S. E., Raznahan, A., Shaw, P. W., Lerch, J. P., Clasen, L. S., Greenstein, Kan, E., & Toga, A. W. (2004). Longitudinal mapping of D., Berman, R.,... Giedd, J. N. (2014). Longitudinal four- cortical thickness and brain growth in normal children. The dimensional mapping of subcortical anatomy in human Journal of Neuroscience, 24, 8223–8231. development. Proceedings of the National Academy of Steinberg, L. (2010). A dual systems model of adolescent risk- Sciences, USA, 111, 1592–1597. taking. Developmental Psychobiology, 52, 216–224. Roy, M., Shohamy, D., & Wager, T. D. (2012