TY - JOUR
AU - PhD, Angela D Bryan,
AB - Abstract Background Identifying cognitive and neural mechanisms of decision making in adolescence can enhance understanding of, and interventions to reduce, risky health behaviors in adolescence. Delay discounting, or the propensity to discount the magnitude of temporally distal rewards, has been associated with diverse health risk behaviors, including risky sex. This cognitive process involves recruitment of reward and cognitive control brain regions, which develop on different trajectories in adolescence and are also implicated in real-world risky decision making. However, no extant research has examined how neural activation during delay discounting is associated with adolescents’ risky sexual behavior. Purpose To determine whether a relationship exists between adolescents’ risky sexual behavior and neural activation during delay discounting. Methods Adolescent participants completed a delay discounting paradigm during functional magnetic resonance imaging (fMRI) scanning, and they reported risky sexual behavior at baseline, 3-, 6-, 9-, and 12-month follow-up time points. Latent growth curve models were employed to determine relationships between brain activation during delay discounting and change in risky sexual behavior over time. Results Greater activation in brain regions associated with reward and cognitive control (caudate, putamen, nucleus accumbens, anterior cingulate, insula, orbitofrontal cortex, inferior frontal gyrus, dorsolateral prefrontal cortex) during delay discounting was associated with lower mean levels of risky sexual behavior but greater growth over the period from baseline to 6 months. Conclusions Neural activation during delay discounting is cross-sectionally and prospectively associated with risky sexual behavior in adolescence, highlighting a neural basis of risky decision-making as well as opportunities for early identification and intervention. Introduction Nearly half of American high school students have had sexual intercourse, but more than 40% of sexually active adolescents report not using a condom at their last intercourse [1]. Unprotected sex can lead to numerous adverse outcomes, including the transmission of sexually transmitted infections as well as unplanned pregnancy. Half of new sexually transmitted infections occur in individuals aged 15–24, and one quarter of new HIV cases occur in individuals aged 13–24 [2]. Clearly, reducing adolescents’ risky sexual behavior is of great importance to public health. In order to do so, it is necessary to enhance our understanding of cognitive, social, and biological influences on decisions to engage in risky sexual behavior, especially those that are specific to adolescence. Adolescence is a unique neurodevelopmental period that is frequently associated with increased engagement in behaviors that are both high-risk and hold the potential for high-reward (e.g., substance use, sexual risk behaviors, etc.) [3]. Importantly, during adolescence, the neural machinery supporting the critical evaluation of the possible risks versus rewards associated with a given action is still in development [4]. Though the specific developmental mechanisms that contribute to adolescent risk-taking are not fully understood, a number of theories have converged to highlight the imbalance between neural reward and control systems that occurs during adolescence. The Dual Systems Model is a highly influential approach, which proposes that the affective/approach system that responds to rewards matures earlier than the cognitive control system [5]. Since it is this control system that contributes to the inhibition of behavior, this imbalance is posited to account for the increased reward-seeking and engagement in risky behaviors observed during adolescence [6]. A parallel model proposed by Ernst and colleagues [7] is the triadic model, which indicates an imbalance across three neurobiologically mediated systems, the approach/reward-driven system (subserved by ventral striatum circuitry), the harm-avoidance system (subserved by limbic circuitry), and the regulatory system (subserved by the medial/ventral prefrontal cortex). This model states that during adolescence, hormonal changes shift the balance of these systems such that the reward-system is strengthened and the avoidance and regulatory systems are weakened. Some researchers have utilized this approach to frame a neurodevelopmental account of adolescents’ and young adults’ risky sexual decisions, suggesting that the combination of decreased threat reactivity in the amygdala and increased reward reactivity in the ventral striatum leads to increased risky sexual behavior [8]. Finally, the Imbalance Model [6, 9, 10] takes a more integrated, circuit-based perspective, and proposes that differential developmental trajectories of subcortical and limbic brain circuits compared to cortical control circuits results in an imbalance (i.e., in terms of functional connectivity, neurochemical availability, etc.) that drives adolescent behavior. Specifically, subcortical circuitry that drives emotional and reward salience matures earlier than do cortical control circuits, creating an “imbalance” that can lead to risky behavior engagement (e.g., substance use, risk-taking in peer social contexts) when control resources are insufficient to regulate emotional and reward drives. This model is informed not only by human imaging work, but also by nonhuman animal studies demonstrating regional differences across development in terms of synaptogenesis, synaptic pruning, availability of neurochemicals and receptor densities, such that subcortical and reward circuitry matures faster than control regions such as the prefrontal cortex [11–14]. To date, human structural and functional imaging studies have provided some of the most compelling evidence to advance our developmental understanding of adolescent risk-taking. In particular, several functional neuroimaging tasks have been leveraged in recent years to examine the relationship between adolescent risk behavior and neural reward and control processing. For example, in a monetary reward paradigm, activation of the nucleus accumbens during anticipation of receiving rewards of various magnitudes and avoiding losses of various magnitudes [15]. Additionally, in the Go-No Go task, which requires inhibitory control, activation of the right inferior frontal gyrus (IFG) and middle occipital gyrus during inhibition trials are associated with adolescent risky sexual behavior [16]. Furthermore, numerous studies have related real-world risk-taking behavior to performance on the Balloon Analogue Risk Task, which assesses risk-taking propensity in the context of rewards. Notably, riskier performance on this task in adolescents has been correlated with greater substance use, risky sexual behavior and delinquency [17–19], and adolescents show greater activation of regions including the striatum, thalamus, dorsal anterior cingulate, insula, and prefrontal cortex as they seek larger and larger (and riskier) rewards [20]. Further, longitudinal analyses have demonstrated that striatal reward responding and risk-taking both peak during adolescence [21], while ventrolateral prefrontal cortex activation during risk-taking declines over time, with associated declines in self-reported risk-taking behavior [22]. Taken together, results from these functional imaging paradigms have illuminated critical neurobiological correlates of risk-taking in adolescence, including in health-related domains such as risky sexual behavior. However, to date, research on the cognitive neuroscience of adolescents’ risky decision making and sexual behavior has not addressed relationships between delay discounting and risky sex. Delay discounting, or temporal discounting, is the propensity to discount the value of rewards that are offered in the future as compared to those that are available immediately. This propensity is frequently assessed in the context of monetary decisions, in which individuals are asked if they would prefer to receive a smaller reward sooner (e.g., $20 today) or a larger reward later ($200 one month from now). Individuals who discount the magnitude of future outcomes to a greater degree (i.e., those with greater delay discounting propensities) are more likely to prefer the more immediate rewards, despite their lesser face value. This process is not purely irrational in a broad sense, given that the future is less certain than the present. Nevertheless, this propensity can underlie decision-making that leads to adverse consequences if more immediately salient but smaller rewards are consistently chosen at the expense of temporally distal but more valuable rewards or outcomes. Importantly, this propensity tends to be relatively stable over time within individuals (e.g., [23, 24]), and individual differences in delay discounting propensity are associated with different behavioral and health outcomes. Although the propensity to discount the value of future rewards is frequently measured in the context of monetary choices [25], it is exhibited in many health contexts, including substance use, unhealthy diet, and risky sexual behavior, all of which can involve electing to pursue a small immediate reward while sacrificing future good health or life expectancy. Importantly, responses to intertemporal monetary choices are indeed associated with these types of risky health behaviors and ill health status. Increased delay discounting is associated with engagement in addictive behaviors, including use of various addictive substances (e.g., methamphetamine, alcohol, tobacco) and gambling [26, 27]. This relationship is especially robust among individuals who meet criteria for abuse and dependence [26], and delay discounting has been proposed as an endophenotype of diverse addictive disorders [28]. Evidence also suggests that delay discounting propensity is associated with specific health risk behaviors, including risky sexual behavior. For example, research has demonstrated that increased delay discounting is associated with higher rates of unprotected sex under the influence of alcohol in adults [29] and greater engagement in several risky sexual behaviors among adolescents [30]. Beyond these behavioral measures, neuroscientific investigations have highlighted patterns of neural activation during delay discounting that are associated with health behaviors as well as health status. For example, alcohol use disorder severity is positively associated with activation in brain regions including the insula and IFG when choosing delayed rewards [31]. Additionally, during delay discounting, activation in regions including the anterior cingulate cortex, caudate, and dorsolateral prefrontal cortex (dlPFC) differs between methamphetamine dependent individuals and controls [32, 33]. Moreover, research has demonstrated that behavioral and neural responses to intertemporal monetary choices can prospectively predict health behaviors and outcomes. For example, greater delay discounting propensity in adolescence and young adulthood is predictive of cigarette smoking onset [24], suggesting that this cognitive and affective process may contribute to the etiology of risky health decision-making and is not merely a consequence of behaviors that may affect neural and behavioral reward processing. Additionally, among obese adult women, neural activation in frontal regions (e.g., IFG) during delay discounting is associated with future successful or unsuccessful weight loss [34]. Risky sexual decisions can be construed as intertemporal choices. That is, the decision to have unprotected intercourse is more immediately rewarding than waiting until a later time when a condom is available or abstaining from intercourse completely. However, this immediate reward is associated with increased risk of adverse health outcomes, such as sexually transmitted infections or unwanted pregnancy. Deciding whether to engage in risky sexual behavior, which involves weighing immediate physiological and social rewards against risk of probabilistic but highly negative future health outcomes, would involve the recruitment of both reward and cognitive control brain regions. However, to our knowledge, no study to date has examined relationships between neural activation during delay discounting and adolescent risky sexual behavior. The present study addresses this gap in the extant literature by linking brain activation during intertemporal monetary choices to real-world risky sexual behavior. Study participants were high-risk adolescents recruited through the juvenile justice system, and they completed a delay discounting paradigm during functional magnetic resonance imaging (fMRI) scanning before receiving one of two interventions designed to reduce risky sex. They reported risky sexual behaviors, defined as frequency of unprotected intercourse, in 3-month intervals from baseline to 12 months postintervention, permitting data analyses to highlight the neural substrates of decision-making that are prospectively associated with changes in risky sexual behavior. Given the novelty of this approach and lack of extant research in this domain, our hypotheses were exploratory. We posited that greater activation during intertemporal monetary choices in neural regions associated with reward and cognitive processing, specifically during Hard as compared to Easy choices, would be associated with decreased engagement in risky sexual behavior at baseline and over time during the follow-up period. Method Participants Participants were recruited from a juvenile justice programming center in the southwestern United States. To be eligible for the study, participants had to (i) be between the ages of 14–18 years old, (ii) have verbal consent via tape-recorded phone calls from a parent or legal guardian, (iii) not currently be taking psychotropic medications, and (iv) be free of magnetic resonance imaging (MRI) contraindications, including a history of injury to the brain or brain-related medical conditions, nonremovable metallic implants, recent tattoo, and pregnancy (if female). Juvenile justice staff were not involved in recruitment, and participation decisions had no impact on the adolescent’s juvenile justice status or treatment. Both adolescent assent and parental consent via digital audio recording over the telephone were obtained prior to involvement in the study. The Institutional Review Board at the University of New Mexico and the federal Office of Human Research Protection approved all aspects of the study, and a certificate of confidentiality was obtained to protect participants and encourage full disclosure. All participants were paid for their time at each research study appointment, and they received a total of $185 if they completed all of the sessions (i.e., baseline and fMRI session, 3-, 6-, 9-, and 12-month follow-ups). Intervention Trial The study was a between-subjects randomized controlled intervention with two conditions, each designed to reduce risky sexual behavior. All participants completed a baseline assessment that included behavioral and self-report measures as well as an MRI scan and an individual motivational interviewing (MI) session focused on substance abuse. Participants were then randomly assigned to receive a group sexual risk reduction plus alcohol risk reduction intervention or an information-only comparison sexual risk reduction intervention. Group interventions were conducted at the justice center, and adolescents participated in the single-session group intervention prior to their discharge from the justice center. Participants completed immediate posttest measures after the intervention session, and the same behavioral and self-report measures as at the baseline assessment 3 months, 6 months, 9 months, and 1 year after the intervention at locations convenient for them. The present analysis examines data from the baseline behavioral and neuroimaging assessments as well as the 3-, 6-, 9-, and 12-month follow-up behavioral assessments. Study Procedures Baseline assessment The baseline assessment consisted of a neuroimaging scan, questionnaires, and an individual MI intervention focused on substance abuse. Questionnaire data were collected using Audio Computer-Assisted Self-Interviewing (ACASI) technology on individual laptop computers. Previous experience with this high-risk population indicates that poor literacy can be a limitation, and the ACASI technology helps eliminate many issues with understanding the content of questions and navigating complicated skip patterns in questionnaires [35]. Self-Report measures Participants were asked a series of questions regarding their sexual history, including whether they had ever had sex, the age at which they first had intercourse, and their number of unique sexual partners. Additionally, they were asked whether they had ever been or gotten someone else pregnant, and whether they ever had a diagnosis of an STI. At each time point, we assessed frequency of intercourse (“On average, in the past 3 months, how often have you had sexual intercourse?”) on a scale ranging from 1 (never) to 6 (almost every day); and condom use (“In the past 3 months, how much of the time did you use condoms when you had sexual intercourse?”) on a scale ranging from 1 (never) to 5 (always). Risky sexual behavior was examined at all time points using a composite variable calculated as how frequently a participant had sex multiplied by how often they used a condom when having sex (reverse coded), such that higher scores indicate riskier sexual behavior. Additionally, participants completed a set of demographic items as well as the Pubertal Development Scale [36]. Image Acquisition and Processing MRI was performed on a 3T Siemens Trio (Erlangen, Germany) whole body scanner with a 12-channel radio frequency coil. Participants were placed in the scanner and a piece of tape across each participant’s forehead reduced movement. A high-resolution T1-weighted structural image was acquired with a 5-echo multiecho MPRAGE sequence with TE = 1.64, 3.50, 5.36, 7.22, and 9.08 ms, TR = 2.53 s, TI = 1.20 s, flip angle = 7°, NEX = 1, slice thickness = 1 mm, 33 slices, FOV = 256 × 256 mm, resolution = 256 × 256 × 176, voxel size = 1 × 1 × 1 mm, and pixel bandwidth = 650 Hz. Whole brain resting state T2-weighted functional images were acquired using a gradient-echo echo planar imaging sequence with TE = 29 ms, TR = 2 s, flip angle = 75°, slice thickness = 3.5 mm, 33 slices, slice gap = 1.05 mm, FOV = 240 × 240 mm, 64 × 64 matrix, voxel size = 3.75 × 3.75 × 3.5 mm. Functional images were preprocessed using an automated pipeline based in SPM 5 (http://www.fil.ion.ucl.ac.uk/spm/software/spm5), including realignment, slice-timing correction, unified segmentation spatial normalization, reslicing, and smoothing with a 10 mm Gaussian kernel [37]. Time-series of cerebrospinal fluid and white-matter fluctuations, motion parameters, and first derivative of motion parameters were progressively regressed from the time-series data and then the data were band-pass filtered (0.01–0.1 Hz) using in-house scripts [38]. Further, individuals with excessive motion (>3 mm translational movement) were excluded from analyses. Regions of interest Neuroanatomical regions of interest were selected based on prior delay discounting and reward processing literature suggesting involvement of nucleus accumbens and orbitofrontal cortex (OFC) [39, 40]; IFG, insula, putamen, and caudate [31, 41]; and dorsal anterior cingulate cortex (dACC) and dlPFC [31, 34, 40]. Signal change value extraction was performed in MarsBaR version 0.43 for MATLAB [42], using bilateral spherical regions of interest (i.e., based upon the above literature and Harvard-Oxford atlases) centered at the following MNI coordinates. Spheres of 5 mm radius were created for nucleus accumbens (±9, 12, −6); putamen (±21, 9, −6); caudate (±12, 15, 6); and dACC (±6, 27, 27). Spheres of 10 mm radius were created for insula (±39, −6, 6); OFC (±36, 27, −15); IFG (±48, 24, 6); and dlPFC (±45, 30, 27). Delay Discounting Paradigm During the scan, participants completed an fMRI-compatible delay discounting paradigm developed by Claus and colleagues [31], whose manuscript includes a schematic diagram of the task trial presentation. The task involves a series of hypothetical intertemporal monetary choices, and for each trial, participants must select whether they would prefer to receive a smaller monetary reward sooner or a larger reward later (e.g., $35 today or $50 in 2 weeks). Trials vary in the magnitude of the monetary rewards and the length of delay for the large reward and are individually tailored for each participant. Each trial in the event-related design lasts 6 seconds, and the contrasts of interest (detailed below) are calculated based upon the time period during which the participant makes the decision as well as a 2-second period following the decision when the chosen option is highlighted on the screen. Trials are separated by a variable intertrial interval. Before entering the MRI scanner, participants complete a series of computerized practice trials, which orient them to the task and also permit an estimation of each participant’s discounting parameter (k value). This estimate then allows an algorithm to tailor the in-scanner trials to be either Easy (i.e., farther from the participant’s indifference point) or Hard (i.e., closer to the indifference point) for the participant as well as to bias the participant to choose the more immediate or more delayed option, based upon their individual preferences and indifference point. These two factors (difficulty and time preference) are crossed in this task to create four types of trials: Easy Now, Hard Now, Easy Later, and Hard Later. Thus, for example, an Easy Now trial is designed such that the participant would be biased to select the smaller, more immediate reward, and the decision is easier for them to make because the choice options are selected to be further from the participant’s indifference point. In the scanner, participants complete a total of 18 trials per condition, separated across two functional imaging runs. This approach ensures appropriate variability in response choices and permits calculation of key neuroimaging contrasts. The contrast of interest was Hard > Easy (collapsed across time preference), which highlights brain regions implicated in the choices that should be more, rather than less, effortful based upon the participant’s specific delay preferences. Brain activation during Hard > Easy delay discounting trials has been associated with other health conditions (e.g., [32–34]). Importantly, this contrast was selected because it may help to highlight the neural substrates of the type of cognitive processing that underlies decision-making when youth are faced with an opportunity to engage in risky sexual behavior, which likely entails highly motivated emotional and cognitive processing when faced with a salient immediate reward that carries the risk of substantial and detrimental long-term impact (i.e., sexually transmitted infections, unwanted pregnancy). Latent Growth Curve Modeling Latent growth curve models, using brain activation during delay discounting to prospectively predict change over time in risky sexual behavior, were estimated using maximum likelihood estimation in Mplus 7 for Mac [43]. Neural activation was modeled as a latent factor, constructed from the extracted signal change values on the Hard > Easy contrast in each of the brain regions of interest. Models controlled for pubertal development status, which incorporates both age- and sex-related effects, as well as intervention condition. Results Self-report measures A total of 284 participants were enrolled in the parent trial. Of those, 48 did not have an MRI scan session, 7 did not have usable scan data, 19 were excluded for being outliers on the delay discounting task, and 33 were excluded for having motion in excess of 3 mm during the task. Thus, a total of 177 participants were included in the analysis, of whom 46 (26%) were female. Demographics and other self-report characteristics appear in Table 1. At baseline, participants ranged in age from 14–18 (mean 16.18, standard deviation [SD] = 1.07), with a mean pubertal development scale score of 3.22 (SD = 0.38), indicating that for the average participant, pubertal development had begun but was not complete, across several gender-specific metrics (e.g., body hair development, facial hair/menarche). Participants were 58.2% Hispanic or Latino, 23.2% non-Hispanic White, 5.1% African American, 3.4% Native American, 1.1% Asian or Pacific Islander, and 8.5% more than one race. The majority of participants (87.3%) indicated only other-sex sexual attraction and behavior; 11.6% of participants were bisexual and 1.2% of participants indicated only same-sex sexual attraction and behavior. Of those participants (n = 168) who received an intervention, 48.8% received the information-only sexual risk reduction intervention and 51.2% received the alcohol and sexual risk reduction intervention. Table 1 Demographics and participant characteristics Characteristic  Percentage of participants  Gender     Male  74.0   Female  26.0  Race/ethnicity     Hispanic/Latino  58.2   Non-Hispanic White  23.2   Black/African American  5.1   Native American  3.4   Asian/Pacific Islander  1.1   More than one race  8.5  Sexual Preferences     Only other-sex attraction  87.3   Only same-sex attraction  1.2   Bisexual  11.6  Intervention condition     Information-only sexual risk  46.3   Alcohol and sexual risk  48.6   No intervention  5.1  Self-reported STI diagnosis (baseline)  6.8    Have been/have gotten partner pregnant (baseline)  14.8    Characteristic  Mean  Std. Dev.  Age  16.18  1.07  Age of sexual debut  13.29  2.10  Pubertal development (baseline)  3.22  0.38  AUDIT (baseline)  7.30  6.92  Risky sex metric (0–30; higher is greater risk)       Baseline  6.09  6.68   3 Months  6.37  7.49   6 Months  5.46  7.27   9 Months  6.39  7.93   12 Months  6.27  7.93  Characteristic  Percentage of participants  Gender     Male  74.0   Female  26.0  Race/ethnicity     Hispanic/Latino  58.2   Non-Hispanic White  23.2   Black/African American  5.1   Native American  3.4   Asian/Pacific Islander  1.1   More than one race  8.5  Sexual Preferences     Only other-sex attraction  87.3   Only same-sex attraction  1.2   Bisexual  11.6  Intervention condition     Information-only sexual risk  46.3   Alcohol and sexual risk  48.6   No intervention  5.1  Self-reported STI diagnosis (baseline)  6.8    Have been/have gotten partner pregnant (baseline)  14.8    Characteristic  Mean  Std. Dev.  Age  16.18  1.07  Age of sexual debut  13.29  2.10  Pubertal development (baseline)  3.22  0.38  AUDIT (baseline)  7.30  6.92  Risky sex metric (0–30; higher is greater risk)       Baseline  6.09  6.68   3 Months  6.37  7.49   6 Months  5.46  7.27   9 Months  6.39  7.93   12 Months  6.27  7.93  View Large Table 1 Demographics and participant characteristics Characteristic  Percentage of participants  Gender     Male  74.0   Female  26.0  Race/ethnicity     Hispanic/Latino  58.2   Non-Hispanic White  23.2   Black/African American  5.1   Native American  3.4   Asian/Pacific Islander  1.1   More than one race  8.5  Sexual Preferences     Only other-sex attraction  87.3   Only same-sex attraction  1.2   Bisexual  11.6  Intervention condition     Information-only sexual risk  46.3   Alcohol and sexual risk  48.6   No intervention  5.1  Self-reported STI diagnosis (baseline)  6.8    Have been/have gotten partner pregnant (baseline)  14.8    Characteristic  Mean  Std. Dev.  Age  16.18  1.07  Age of sexual debut  13.29  2.10  Pubertal development (baseline)  3.22  0.38  AUDIT (baseline)  7.30  6.92  Risky sex metric (0–30; higher is greater risk)       Baseline  6.09  6.68   3 Months  6.37  7.49   6 Months  5.46  7.27   9 Months  6.39  7.93   12 Months  6.27  7.93  Characteristic  Percentage of participants  Gender     Male  74.0   Female  26.0  Race/ethnicity     Hispanic/Latino  58.2   Non-Hispanic White  23.2   Black/African American  5.1   Native American  3.4   Asian/Pacific Islander  1.1   More than one race  8.5  Sexual Preferences     Only other-sex attraction  87.3   Only same-sex attraction  1.2   Bisexual  11.6  Intervention condition     Information-only sexual risk  46.3   Alcohol and sexual risk  48.6   No intervention  5.1  Self-reported STI diagnosis (baseline)  6.8    Have been/have gotten partner pregnant (baseline)  14.8    Characteristic  Mean  Std. Dev.  Age  16.18  1.07  Age of sexual debut  13.29  2.10  Pubertal development (baseline)  3.22  0.38  AUDIT (baseline)  7.30  6.92  Risky sex metric (0–30; higher is greater risk)       Baseline  6.09  6.68   3 Months  6.37  7.49   6 Months  5.46  7.27   9 Months  6.39  7.93   12 Months  6.27  7.93  View Large The mean age of sexual debut was 13.29 (SD = 2.1; median and mode = 13). At baseline, 6.8% of participants endorsed that they had ever had a sexually transmitted infection, and 14.7% of participants indicated that they had either been pregnant (if female) or gotten someone pregnant (if male). At all five time points (Baseline, 3 Months, 6 Months, 9 Months, 12 Months), the risky sex scores ranged from 0 (no sexual intercourse in the past 3 months) to 30 (intercourse “almost every day”, “never” using condoms). The means (SD) at each time point were as follows: Baseline 6.09 (6.68), 3 Months 6.37 (7.49), 6 Months 5.46(7.27), 9 Months 6.39 (7.93), and 12 Months 6.27 (7.93). For reference, a score of 6 could represent the following behavior patterns: monthly intercourse, “almost never” using condoms, or weekly intercourse “sometimes” using condoms. Neural Activation During Delay Discounting Before proceeding to the primary latent growth analysis, we conducted a whole brain analysis on the Hard > Easy contrast to assess main effects of the task on brain activation. In other words, this analysis tests whether the task, when implemented in this sample, successfully elicited activation in expected brain regions. Fig. 1 displays the brain regions activated across all participants, without controlling for covariates, during the Hard > Easy contrast, with stronger intensities indicating greater activation on Hard trials than Easy trials. They include regions implicated in reward processing and decision making (e.g., caudate, putamen, insula, cingulate cortex, and frontal cortical regions). These results indicate engagement of brain regions that is consistent with hypotheses and prior research findings from fMRI delay discounting paradigms, supporting the validity of this task in this sample. Fig. 1 View largeDownload slide Brain activation during Hard > Easy intertemporal monetary choices (p < .01). Results from a whole brain analysis of the Hard > Easy contrast, without controlling for covariates (threshold p < .01, uncorrected). A list of all brain regions identified in this analysis can be provided by the first author upon request. Fig. 1 View largeDownload slide Brain activation during Hard > Easy intertemporal monetary choices (p < .01). Results from a whole brain analysis of the Hard > Easy contrast, without controlling for covariates (threshold p < .01, uncorrected). A list of all brain regions identified in this analysis can be provided by the first author upon request. Latent Growth Curve Models Extracted values from the right and left structures from each region of interest were averaged to create a single indicator for each structure in the latent variable models (e.g., left and right IFG were averaged for a summary “IFG” activation value). Preliminary analyses indicated a high degree of covariation among all structures, and model fit was significantly enhanced with the use of a single-latent variable for neural activation, as opposed to separate variables for cortical/subcortical regions or reward/control network regions. Thus, this single-latent variable for neural activation was utilized in all final models. Modification indices suggested a residual correlation between dlPFC and IFG. This modification was deemed acceptable, given the strong interconnections between these regions [44] as well as their spatial proximity, and thus was adopted for the final models. The group means of risky sexual behavior at each time point indicated a nonlinear trajectory, with a large increase from the 6-month to 9-month time points. When a model was estimated using all five time points, there was not a significant relationship between the latent neural activation variable and the latent slope of risky sexual behavior (p = .38). Given the nonlinear trajectory, models were then estimated without the last two time points (i.e., 9 months, 12 months). Fig. 2 displays the final latent growth curve model, which incorporates a latent neural activation variable that predicts the latent intercept and slope of risky sexual behavior at baseline, 3 months, and 6 months, controlling for intervention condition and pubertal development status. This model displayed an adequate fit to the data (CFI = 0.96; RMSEA = 0.076; SRMR = 0.059; X2 (60) = 121.77, p < .001), and all indicator loadings on the neural activation variable were statistically significant (.742 to .931, all p’s < .001). The neural activation latent variable was significantly negatively associated with the intercept for risky sexual behavior, indicating that greater relative activation in these regions of interest during Hard (vs. Easy) intertemporal choices is associated with lower mean levels risky sexual behavior at baseline. Additionally, the neural activation latent variable was positively associated with the slope of risky sexual behavior, such that greater neural activation during these choices was associated with greater increases, or less substantial decreases, in risky sexual behavior over time, from baseline to 6 months. Notably, in concert with our initial models, these results suggest that neural activity during delay discounting is associated with change in risky sexual behavior at 6 months following fMRI scanning, but over a period longer than 6 months, neural activation may no longer predict the trajectory of risky sexual behavior. Fig. 2 View largeDownload slide Latent growth model predicting 6-month change in risky sexual behavior. Only statistically significant paths are presented. Integer values indicate fixed loadings. *p ≤ .05; **p ≤ .01; ***p ≤ .001. ACC anterior cingulate cortex; PFC prefrontal cortex. Fig. 2 View largeDownload slide Latent growth model predicting 6-month change in risky sexual behavior. Only statistically significant paths are presented. Integer values indicate fixed loadings. *p ≤ .05; **p ≤ .01; ***p ≤ .001. ACC anterior cingulate cortex; PFC prefrontal cortex. Additionally, consistent with the overall outcomes for the trial, intervention condition significantly negatively predicted the slope of risky sexual behavior, such that individuals in the combined risky sex and alcohol intervention demonstrated a more substantial decrease in risky sexual behavior over time. Condition was not significantly associated with the risky sex intercept, and pubertal development was not significantly associated with the slope or intercept for risky sexual behavior. Discussion A vast literature has sought to elucidate the neural correlates of risk behavior engagement in adolescence, and this research has been incredibly fruitful in its contributions to our understanding of the confluence of neurodevelopmental factors in this unique life stage and characteristic increases in risk-taking behavior. A wealth of extant research has highlighted the mechanistic role that brain development, especially differential development and engagement of reward and cognitive control circuitry, plays in the etiology of risky behavior engagement. However, to date, remarkably little research has examined the neurocognitive correlates of risky sexual behavior [45], which emerges in adolescence and can have severe health consequences. The present research demonstrates the first investigation of the neural substrates of delay discounting that are associated with risky sexual behavior in adolescence. Though prior research has provided evidence that a brief behavioral measure of delay discounting propensity is associated with prior risky sexual behavior among adolescents and young adults [30], the present findings demonstrate these relationships using a brain-based measure of activation during a delay discounting task and prospective design. We found that brain activation during delay discounting is associated with both mean levels as well as change in risky sexual behavior over time, from only in the 3 months prior to fMRI scanning, as well as also 3 and 6 months later. The relationship between brain activation during Hard (as compared to Easy) choices and change in risky sexual behavior over time was positive, which may suggest that less efficient or mature [4] neural processing during more cognitively demanding decisions may be a risk factor for increased risky sexual behavior or less receptivity to sexual risk reduction interventions. These findings indicate both concurrence with and divergence from the prior studies that have examined neural correlates of delay discounting in adolescence. Specifically, our findings largely align with these other results in highlighting the roles of key reward or valuation (e.g., ventral striatum, insula, caudate) and control (e.g., IFG, dlPFC) regions in intertemporal monetary decisions among adolescents [46, 47]. However, those two prior studies also provided indication of oppositional relationships between reward and control circuitry, such that successful delay of gratification was associated with greater relative activation of the control regions and lesser relative activation of the reward regions when examining Later versus Now choices. In contrast, our results highlight that risky sexual behavior is associated with a neural factor characterized by reward and control region activations of the same valence. This divergence may be explained by the different decision-making processes highlighted by the contrasts of interest (Hard vs. Easy as compared to Later vs. Now) as well as our specific focus on risky sexual behavior (as opposed to pure task performance) as an outcome, but future research should nonetheless continue to examine these differences. The present findings highlight the contribution of both reward and cognitive control regions to intertemporal monetary choices and risky sexual behavior. Thus, they align with a robust literature that emphasizes the importance of both reward and control circuitry in decision-making and risk-taking during the unique adolescent period of neurodevelopment and behavioral exploration. Many theoretical approaches in this domain include the premise that adolescent risk behavior is driven in part by an inability of later-developing frontal control circuity to inhibit relatively more mature mesolimbic circuitry. In contrast to this oppositional account, the present findings indicate that risky sexual behavior engagement was associated with a latent variable that includes both reward and control regions loading in the same, rather than opposing, directions. This result thus aligns with contemporary models of motivated cognition and processing during adolescence, in which corticostriatal interactions (i.e., components of both systems) are key drivers of reward learning and decision-making in motivated contexts [10], as well as work suggesting that a singular valuation system, which requires coordinated rather than competing relationships between subcortical reward circuitry and neocortical control regions, underlies motivated decisions like intertemporal monetary choices [48]. Importantly, understanding cognitive and neural factors that contribute to adolescents’ decisions and behavior can improve our intervention techniques designed to promote safe and healthy behavior during this crucial developmental window. Notably, many interventions designed to decrease risk behavior in this age range focus on social-cognitive theories of health behavior, which frequently fail to address motivational and affective factors, such as reward processing, as well as the role of cognitive control (and deficiencies therein) [45]. That is to say, many of our extant intervention techniques, including those employed in the interventions utilized in the present study, do not address the distinguishing factors of adolescent neurodevelopment, cognition, and decision-making. As demonstrated in the present study, these neurocognitive factors, which prospectively predict risky sex for up to 6 months over and above intervention effects, are crucial contributors to behavior engagement. Thus, integrating knowledge of the unique features of reward processing, cognitive control, and decision-making during adolescence into health behavior theory and subsequently future interventions may greatly enhance their efficacy. Limitations The specialized high-risk sample is both a strength and a limitation of the present research. Specifically, the population of youth involved in the juvenile justice system engages in risk behaviors to a greater degree than their age peers who are not involved in the justice system, and thus these results may not be representative of the adolescent population at large. Nevertheless, these high-risk youth also demonstrate the greatest need for behavioral risk reduction, and thus an increased understanding of their specific neurocognitive and behavioral patterns may have an outsize effect on public health and intervention efforts. An additional limitation is the nature of the items that contribute to the risky sexual behavior metric. First, the questions assessing intercourse and condom use frequency are self-report measures of behavior and are thus subjective reports that are subject to recall and response biases. However, obtaining objective measures of these behaviors is obviously infeasible, for myriad practical, legal, and ethical reasons. In addition, the response options for these items are subjective and non-specific quantitative concepts of frequency (e.g., “sometimes”, “almost always”). These response options may be interpreted differently by different participants and do not offer great numerical specificity. However, in our prior and ongoing research with juvenile justice involved adolescents, we have found that numeracy is generally low in our participants, and thus response options using percentages or fractions are difficult for participants to interpret and result in results that are less reliable than those anchored with these less specific verbal labels. Continued translational research at the interface between cognitive neuroscience and clinical and intervention sciences stands to contribute to improved treatment and prevention outcomes for adolescents by targeting interventions for this unique developmental population and helping to identify which treatments will be most effective [49]. More broadly, biopsychosocial frameworks, which consider and integrate social, cognitive, affective, motivational, neurobiological, genetic, and environmental influences on behavior and health, can enhance the efficacy of our interventions in adolescent and other specialized populations [50]. Thus, this enhanced understanding of neural contributions to adolescent risky sexual behavior is situated within a broader context of scientific approaches to understanding and targeting the diverse biobehavioral factors that can be harnessed to promote public health by reducing risk behavior engagement. Acknowledgement Funding: This research was funded by a grant from the National Institute on Alcohol Abuse and Alcoholism (R01AA013844) to Angela D. Bryan. Casey K. Gardiner is funded by a National Science Foundation Graduate Research Fellowship (DGE 1144083), and Rachel E. Thayer is funded by the National Institute on Drug Abuse (R36DA040020). Compliance with Ethical Standards Conflict of interest: The authors declare that they have no conflict of interest. Ethical Approval: All research procedures were conducted in accordance with ethical standards, including the Declaration of Helsinki and its later amendments. The Institutional Review Board at the University of New Mexico and the federal Office of Human Research Protection approved all aspects of the study, and a certificate of confidentiality was obtained to protect participants and encourage full disclosure. Both adolescent participant assent and parental consent via digital audio recording over the telephone were obtained prior to involvement in the study. References 1. Kann L, Kinchen S, Shanklin SLet al.  ; Centers for Disease Control and Prevention (CDC). Youth risk behavior surveillance–United States, 2013. MMWR Suppl . 2014; 63( 4): 1– 168. 2. Satterwhite CL, Torrone E, Meites Eet al.   Sexually transmitted infections among US women and men: prevalence and incidence estimates, 2008. Sex Transm Dis . 2013; 40( 3): 187– 193. 3. Steinberg L. A social neuroscience perspective on adolescent risk-taking. Dev Rev . 2008; 28( 1): 78– 106. 4. Luna B, Padmanabhan A, O’Hearn K. What has fMRI told us about the development of cognitive control through adolescence? Brain Cogn . 2010; 72( 1): 101– 113. 5. Geier C, Luna B. The maturation of incentive processing and cognitive control. Pharmacol Biochem Behav . 2009; 93( 3): 212– 221. 6. Casey BJ, Jones RM, Hare TA. The adolescent brain. Ann N Y Acad Sci . 2008; 1124(1): 111– 126. 7. Ernst M, Pine DS, Hardin M. Triadic model of the neurobiology of motivated behavior in adolescence. Psychol Med . 2006; 36( 3): 299– 312. 8. Victor EC, Hariri AR. A neuroscience perspective on sexual risk behavior in adolescence and emerging adulthood. Dev Psychopathol . 2016; 28( 2): 471– 487. 9. Casey BJ. Beyond simple models of self-control to circuit-based accounts of adolescent behavior. Annu Rev Psychol . 2015; 66: 295– 319. 10. Somerville LH, Jones RM, Casey BJ. A time of change: behavioral and neural correlates of adolescent sensitivity to appetitive and aversive environmental cues. Brain Cogn . 2010; 72( 1): 124– 133. 11. Brenhouse HC, Sonntag KC, Andersen SL. Transient D1 dopamine receptor expression on prefrontal cortex projection neurons: relationship to enhanced motivational salience of drug cues in adolescence. J Neurosci . 2008; 28( 10): 2375– 2382. 12. Cunningham MG, Bhattacharyya S, Benes FM. Increasing interaction of amygdalar afferents with GABAergic interneurons between birth and adulthood. Cereb Cortex . 2008; 18( 7): 1529– 1535. 13. Katoh-Semba R, Takeuchi IK, Semba R, Kato K. Distribution of brain-derived neurotrophic factor in rats and its changes with development in the brain. J. Neurochem . 1997; 69(1): 34– 42. 14. Bourgeois JP, Goldman-Rakic PS, Rakic P. Synaptogenesis in the prefrontal cortex of rhesus monkeys. Cereb Cortex . 1994; 4( 1): 78– 96. 15. Galvan A, Hare T, Voss H, Glover G, Casey BJ. Risk-taking and the adolescent brain: who is at risk? Dev Sci . 2007; 10( 2): F8– F14. 16. Ewing SWF, Houck JM, Bryan AD. Neural activation during response inhibition is associated with adolescents’ frequency of risky sex and substance use. Addict. Behav . 2015; 44: 80– 87. 17. Aklin WM, Lejuez CW, Zvolensky MJ, Kahler CW, Gwadz M. Evaluation of behavioral measures of risk taking propensity with inner city adolescents. Behav Res Ther . 2005; 43( 2): 215– 228. 18. Lejuez CW, Aklin WM, Zvolensky MJ, Pedulla CM. Evaluation of the Balloon Analogue Risk Task (BART) as a predictor of adolescent real-world risk-taking behaviours. J Adolesc . 2003; 26( 4): 475– 479. 19. Lejuez CW, Aklin W, Daughters S, Zvolensky M, Kahler C, Gwadz M. Reliability and validity of the youth version of the Balloon Analogue Risk Task (BART-Y) in the assessment of risk-taking behavior among inner-city adolescents. J Clin Child Adolesc Psychol . 2007; 36( 1): 106– 111. 20. Chiu C.-Y P, Tlustos SJ, Walz NC, Holland SK, Eliassen JC, Bernard L, Wade SL. Neural correlates of risky decision making in adolescents with and without traumatic brain injury using the balloon analog risk task. Dev. Neuropsychol . 2012; 37(2): 176– 183. 21. Braams BR, van Duijvenvoorde AC, Peper JS, Crone EA. Longitudinal changes in adolescent risk-taking: a comprehensive study of neural responses to rewards, pubertal development, and risk-taking behavior. J Neurosci . 2015; 35( 18): 7226– 7238. 22. Qu Y, Galvan A, Fuligni AJ, Lieberman MD, Telzer EH. Longitudinal changes in prefrontal cortex activation underlie declines in adolescent risk taking. J Neurosci . 2015; 35( 32): 11308– 11314. 23. Kirby KN. One-year temporal stability of delay-discount rates. Psychon Bull Rev . 2009; 16( 3): 457– 462. 24. Audrain-McGovern J, Rodriguez D, Epstein LH, Cuevas J, Rodgers K, Wileyto EP. Does delay discounting play an etiological role in smoking or is it a consequence of smoking? Drug Alcohol Depend . 2009; 103( 3): 99– 106. 25. Kirby KN, Petry NM, Bickel WK. Heroin addicts have higher discount rates for delayed rewards than non-drug-using controls. J Exp Psychol Gen . 1999; 128( 1): 78– 87. 26. MacKillop J, Amlung MT, Few LR, Ray LA, Sweet LH, Munafò MR. Delayed reward discounting and addictive behavior: a meta-analysis. Psychopharmacology (Berl) . 2011; 216( 3): 305– 321. 27. Reynolds B. A review of delay-discounting research with humans: relations to drug use and gambling. Behav Pharmacol . 2006; 17(8): 651– 667. 28. MacKillop J. Integrating behavioral economics and behavioral genetics: delayed reward discounting as an endophenotype for addictive disorders. J Exp Anal Behav . 2013; 99( 1): 14– 31. 29. MacKillop J, Celio MA, Mastroleo NRet al.   Behavioral economic decision making and alcohol-related sexual risk behavior. AIDS Behav . 2015; 19( 3): 450– 458. 30. Chesson HW, Leichliter JS, Zimet GD, Rosenthal SL, Bernstein DI, Fife KH. Discount rates and risky sexual behaviors among teenagers and young adults. J. Risk Uncertain . 2006; 32(3): 217– 230. 31. Claus ED, Kiehl KA, Hutchison KE. Neural and behavioral mechanisms of impulsive choice in alcohol use disorder. Alcohol Clin Exp Res . 2011; 35( 7): 1209– 1219. 32. Monterosso JR, Ainslie G, Xu J, Cordova X, Domier CP, London ED. Frontoparietal cortical activity of methamphetamine-dependent and comparison subjects performing a delay discounting task. Hum Brain Mapp . 2007; 28( 5): 383– 393. 33. Hoffman WF, Schwartz DL, Huckans MSet al.   Cortical activation during delay discounting in abstinent methamphetamine dependent individuals. Psychopharmacology (Berl) . 2008; 201( 2): 183– 193. 34. Kishinevsky FI, Cox JE, Murdaugh DL, Stoeckel LE, Cook EW3rd, Weller RE. fMRI reactivity on a delay discounting task predicts weight gain in obese women. Appetite . 2012; 58( 2): 582– 592. 35. Schmiege SJ, Broaddus MR, Levin M, Bryan AD. Randomized trial of group interventions to reduce HIV/STD risk and change theoretical mediators among detained adolescents. J Consult Clin Psychol . 2009; 77( 1): 38– 50. 36. Carskadon MA, Acebo C. A self-administered rating scale for pubertal development. J Adolesc Health . 1993; 14(3): 190– 195. 37. Scott A, Courtney W, Wood D, De la Garza R, Lane S, Wang R, King M, Roberts J, Turner JA, Calhoun VD. COINS: an innovative informatics and neuroimaging tool suite built for large heterogeneous datasets. Front. Neuroinformatics . 2011; 5: 33. 38. Welsh RC, Chen AC, Taylor SF. Low-frequency BOLD fluctuations demonstrate altered thalamocortical connectivity in schizophrenia. Schizophr Bull . 2010; 36( 4): 713– 722. 39. Karoly HC, Bryan AD, Weiland BJ, Mayer A, Dodd A, Feldstein Ewing SW. Does incentive-elicited nucleus accumbens activation differ by substance of abuse? An examination with adolescents. Dev Cogn Neurosci . 2015; 16: 5– 15. 40. Yokum S, Gearhardt AN, Harris JL, Brownell KD, Stice E. Individual differences in striatum activity to food commercials predict weight gain in adolescents. Obesity (Silver Spring) . 2014; 22( 12): 2544– 2551. 41. Karoly HC, Weiland BJ, Sabbineni A, Hutchison KE. Preliminary functional MRI results from a combined stop-signal alcohol-cue task. J Stud Alcohol Drugs . 2014; 75( 4): 664– 673. 42. Brett M, Anton J.-L, Valabregue R, Poline J.-B. Region of interest analysis using the MarsBar toolbox for SPM 99. Neuroimage . 2002; 16(2): S497. 43. Muthén LK, Muthén BO. Mplus: Statistical Analysis With Latent Variables: User’s Guide . LA: Muthén & Muthén; 2005. 44. Morawetz C, Bode S, Baudewig J, Kirilina E, Heekeren HR. Changes in effective connectivity between dorsal and ventral prefrontal regions moderate emotion regulation. Cereb Cortex . 2016; 26( 5): 1923– 1937. 45. Ewing SWF, Ryman SG, Gillman AS, Weiland BJ, Thayer RE, Bryan AD. Developmental cognitive neuroscience of adolescent sexual risk and alcohol use. AIDS Behav . 2016; 20(1): 97– 108. 46. Christakou A, Brammer M, Rubia K. Maturation of limbic corticostriatal activation and connectivity associated with developmental changes in temporal discounting. Neuroimage . 2011; 54( 2): 1344– 1354. 47. Stanger C, Elton A, Ryan SR, James GA, Budney AJ, Kilts CD. Neuroeconomics and adolescent substance abuse: individual differences in neural networks and delay discounting. J Am Acad Child Adolesc Psychiatry . 2013; 52(7): 747– 755.e6. 48. Monterosso JR, Luo S. An argument against dual valuation system competition: cognitive capacities supporting future orientation mediate rather than compete with visceral motivations. J Neurosci Psychol Econ . 2010; 3( 1): 1– 14. 49. Ewing SWF, Tapert SF, Molina BS. Uniting adolescent neuroimaging and treatment research: recommendations in pursuit of improved integration. Neurosci Biobehav Rev . 2016; 62: 109– 114. 50. Magnan RE, Callahan TJ, Ladd BO, Claus ED, Hutchison KE, Bryan AD. Evaluating an integrative theoretical framework for HIV sexual risk among juvenile justice involved adolescents. J AIDS Clin Res . 2013; 4(6): 217. © Society of Behavioral Medicine 2018. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
TI - Neural activation during delay discounting is associated with 6-month change in risky sexual behavior in adolescents
JF - Annals of Behavioral Medicine
DO - 10.1093/abm/kax028
DA - 2018-05-01
UR - https://www.deepdyve.com/lp/oxford-university-press/neural-activation-during-delay-discounting-is-associated-with-6-month-bCc5dG0rHH
SP - 356
EP - 366
VL - 52
IS - 5
DP - DeepDyve
ER -