Abstract
Background
On a theoretical level, impulsivity represents a multidimensional construct associated with acting without foresight, inefficient inhibitory response control, and alterations in reward processing. On an empirical level, relationships and changes in associations between different measures of impulsivity from adolescence into young adulthood and their relation to neural activity during inhibitory control and reward anticipation have not been fully understood.
Methods
We used data from IMAGEN, a longitudinal multicenter, population-based cohort study in which 2034 healthy adolescents were investigated at age 14, and 1383 were reassessed as young adults at age 19. We measured the construct of trait impulsivity using self-report questionnaires and neurocognitive indices of decisional impulsivity. With functional magnetic resonance imaging, we assessed brain activity during inhibition error processing using the stop signal task and during reward anticipation in the monetary incentive delay task. Correlations were analyzed, and mixed-effect models were fitted to explore developmental and predictive effects.
Results
All self-report and neurocognitive measures of impulsivity proved to be correlated during adolescence and young adulthood. Further, pre-supplementary motor area and inferior frontal gyrus activity during inhibition error processing was associated with trait impulsivity in adolescence, whereas in young adulthood, a trend-level association with reward anticipation activity in the ventral striatum was found. For adult delay discounting, a trend-level predictive effect of adolescent neural activity during inhibition error processing emerged.
Conclusions
Our findings help to inform theories of impulsivity about the development of its multidimensional nature and associated brain activity patterns and highlight the need for taking functional brain development into account when evaluating neuromarker candidates.
Keywords
The ability to decide and act quickly without hesitation can be advantageous in some settings. However, when persistently expressed, this can also have negative consequences in many daily life situations and is then referred to as impulsivity (
1
,2
). Impulsivity, often defined as action without foresight, is reflected in a lack of self-regulatory capacity associated with overt difficulties in response inhibition and altered reward processing (3
). It underlies a broad range of human behaviors and a variety of different clinically relevant mental health conditions (4
,5
). Definitions range from rather general multifaceted personality accounts to more specific operationalizations, focusing, for example, on indices derived from neurocognitive testing (2
). Impulsivity should thus be viewed as a multidimensional construct encompassing a broad variety of components and facets (6
,7
). It is therefore not surprising that a range of methods exists to assess impulsivity (4
,8
). However, on an empirical level, it is still not clear if and how those different rather specific measures of impulsivity are related, and if and how associations change over time, particularly during critical developmental periods, such as adolescence. Furthermore, in this context, studies that relate different measures of impulsivity to brain activity patterns of related neural processes are lacking. This knowledge would empirically inform theoretical conceptualizations of the construct and enable a deeper characterization and understanding of its multidimensional nature, the various methods of assessment, associated neural processes, and developmental changes. Further, it might be relevant for potential clinical application with regard to the identification of relevant biomarkers for indicating vulnerability or for predicting treatment response.The current study addresses these open issues within a large-scale multicenter study exploring a population-based cohort of adolescents followed longitudinally into young adulthood (
9
). In line with Dalley and Robbins (6
), for the multidimensional concept of impulsivity, we distinguished trait impulsivity assessed via self-report, questionnaire-based methods from more objective neurocognitive measures. Thereby, trait impulsivity is defined as a predisposition for rapid, but often premature actions without appropriate foresight (6
), reflecting preferences for impulsive responding and immediate rewards (10
,11
). It tends to explain impulsive behaviors such as impaired response inhibition as a consequence of dysfunctional processing of delayed rewards (3
). For objective neurocognitive measures of impulsivity, we differentiated between two indices of decisional impulsivity (6
) assessing temporal discounting of delayed rewards (delay discounting) and delay aversion, respectively (12
). Thereby, decisional impulsivity is reflected in a preference for smaller, immediate rewards over later but larger ones (13
).Whereas some component processes of impulsivity share common variance, thus suggesting overlapping psychological mechanisms, on the level of assessment, different measures of impulsivity often fail to intercorrelate substantially (
14
). Therefore, there might also be a nonoverlap (at least partially) between different facets of impulsivity representing rather distinct psychological component processes and/or distinct facets of the construct assessed via a specific method of measurement (6
,15
, 16
, 17
). Additionally, the period of time in the life span, and thus specific characteristics of the investigated samples, might have co-determined so far rather inconsistent previous results in this respect. Earlier studies on impulsivity mainly relied on adult samples and mostly used cross-sectional designs, while we still know very little about potential intercorrelations of different measures of impulsivity facets during critical maturational phases from a longitudinal perspective (18
,19
). Especially adolescence, with its often acknowledged substantial behavioral and brain changes (20
,21
), might be an interesting developmental period to focus on (22
,23
).Within the last few years, research efforts have tried to unravel the neurobiological underpinnings of impulsivity and its latent phenotypes (
24
,25
), resulting in a vast amount of literature on brain regions involved in aspects of the multifaceted construct. A prominent focus has been on the limbic corticostriatal systems (1
). According to Kozak et al. (26
), two of the relevant brain circuits involved in impulsivity are the cortical control system, including frontal regions for regulatory processes, and the reward system. Among potential brain correlates of interest, particularly the prefrontal cortex, the inferior frontal gyrus (IFG), the pre-supplementary motor area (pre-SMA), and further subcortical regions such as the striatum might play a role as indicated by previous studies [e.g. (18
,22
,27
,28
)]. For example, during inhibitory response control and inhibition error processing, those areas have been associated with neurocognitive measures of impulsivity (29
, 30
, 31
, 32
). Further, during reward processing, frontostriatal deviations were related to measures of trait impulsivity previously (18
). However, it is still unclear if and how those associations with neurobiological indices change over time, e.g., during adolescence (33
), and if those brain activity patterns have a predictive value for later changes in self-report and/or neurocognitive measures of impulsivity (35
). Our study addressed these gaps by capturing brain activity data during inhibition error processing and reward anticipation and relating them to self-report measures of trait impulsivity and neurocognitive indices of decisional impulsivity. Associations were explored longitudinally during adolescence until young adulthood in a large population-based sample using cross-correlations, mixed-effect models, and predictive modeling.Methods and Materials
Participants
Anonymized data for the present study were obtained from the IMAGEN project (
35
), a multicenter study including 8 sites (London, Nottingham, Dublin, Mannheim, Dresden, Berlin, Hamburg, and Paris). Healthy participants were longitudinally assessed at baseline during adolescence (at around age 14) and reassessed at follow-up in young adulthood (at around age 19). Ethics approval was provided by the local ethical committees for each participating center, and written informed consent was obtained from parents or guardians. Furthermore, verbal assent was obtained from the adolescents. As young adults, participants provided written informed consent. Exclusion criteria were serious medical conditions, previous trauma with loss of consciousness, any magnetic resonance imaging (MRI) contraindications, or IQ <70. Details regarding the study protocol and data acquisition have been published previously (9
). Participants were included in the current analyses if they had at least one facet of either trait impulsivity or decisional impulsivity assessed via self-report questionnaires or neurocognitive testing, respectively, and one functional MRI (fMRI) task of interest at baseline. Finally, data from 2034 healthy adolescents (mean [SD] age = 13.96 [0.45] years; IQ = 108.15 [13.53]; 49.4% male) and 1383 of these participants as young adults (age = 19.09 [0.77] years) were included.Assessment of Trait Impulsivity and Neurocognitive Indices of Decisional Impulsivity
Trait impulsivity was assessed at baseline and follow-up using the following scales: the extraversion subscale from the Revised NEO Personality Inventory (
36
), the impulsivity subscale from the Substance Use Risk Profile Scale (37
), and the impulsiveness versus reflection subscale from the Temperament and Character Inventory–Revised (38
). To assess delay discounting as a facet of decisional impulsivity, the Kirby delay discounting task (39
) was used at baseline and follow-up. Further, for measuring delay aversion, the Cambridge Gambling Task (CGT) delay aversion subscale from the Cambridge Cognition Neuropsychological Test Automated Battery (CANTAB) was included at baseline only. For more details on the scales and indices used, see Supplement A.Assessment of Related Neural Activity—fMRI Paradigms
The stop signal task (SST) (
21
,40
) was implemented to assess neural activity related to inhibitory control, more specifically, inhibition error processing in pre-SMA and IFG (see Supplement B for a detailed description and Figure S1 for an example outline). The monetary incentive delay (MID) task (40
,41
) was used to explore striatal activity during reward anticipation (see Supplement B and Figure S2). Both tasks were assessed at baseline and follow-up.For the current analyses, we focused on specific task contrasts and brain regions selected a priori based on previous findings [e.g., from the IMAGEN project (
35
,42
)] for comparison purposes. In addition, the ventral striatum (VS) as a core region related to reward processing (anticipation) (43
) has been the focus of a recent meta-analysis from Plichta and Scheres (18
) exploring the relationship of brain activity in this region with trait impulsivity in adults.Data Acquisition
Imaging data were acquired at each of the 8 sites with 3T MRI scanners by different manufacturers (Siemens, Philips Healthcare, GE Healthcare, Bruker). Full details of the MRI acquisition protocols and quality checks have been published previously (
9
). fMRI images were acquired using an echo-planar imaging sequence. For each subject, 444 volumes were acquired for the SST, and 300 volumes were acquired for the MID task. For both tasks, each volume consisted of 40 slices (2.4-mm slice thickness, 1-mm gap), and to provide reliable imaging of subcortical areas, echo time was optimized (echo time = 30 ms, repetition time = 2.2 s). The same scanning protocol was used at all sites.Data Preparation
Data were z-standardized, and confirmatory factor analyses (
44
) were conducted on the indices for trait impulsivity (NEO Five-Factor Inventory, Substance Use Risk Profile Scale, Temperament and Character Inventory–Revised) and delay discounting K1–K3 values for the Kirby delay discounting task, respectively, to form one latent factor using the R software package lavaan, version 0.6-6 (http://cran.uni-muenster.de/web/packages/lavaan/lavaan.pdf). Predicted values were calculated for each participant estimating the factor scores on each construct. As only one value from the CGT delay aversion subscale was used, no factor analysis was conducted for this impulsivity index.fMRI data were analyzed with SPM8 (https://www.fil.ion.ucl.ac.uk/spm/) and MATLAB, version 2011b (The MathWorks, Inc.). A detailed description of fMRI data preprocessing has been published previously (
45
) (see also Supplement B). The “stop failure versus stop success” contrast was analyzed for each participant to measure neural activity associated with unsuccessful stopping and inhibition error processing [inhibition error detection–related activity, in line with (42
,46
)]. This cognitive process is highly relevant for selecting adequate response tendencies and adapting goal-directed behavior (47
) and thereby directly linked to the theoretical concept of impulsivity. For the MID task, the “anticipation hit big win versus anticipation hit no win” contrast was used for each participant as an index of neural activity associated with anticipation of a large reward (43
). Region-of-interest (ROI) masks were derived from WFU PickAtlas, version 3.0.5 (48
): the bilateral IFG and pre-SMA (separately) for the SST (49
) and the bilateral VS for the MID task (50
,51
). Via the Region of Interest Extraction Toolbox, version 2.1 (https://www.nitrc.org/projects/rex/), mean ROI activity values were exported for each participant. Analogous to the manifest variables assessing trait impulsivity and delay discounting, values were z-standardized, and a confirmatory factor analysis was conducted for the SST mean activity values to estimate predicted values for each participant forming one factor representing unsuccessful stopping activity related to inhibitory control in both ROIs. Only neuronal activity from the SST was included for the core analysis. Behavioral performance data were deliberately omitted, as behavioral and brain measures from the same task are per se dependent, and including both levels of data would have biased our data structure and analyses (nevertheless, descriptions and correlations with fMRI brain activity for the SST performance data are presented in Supplement B). Data distribution and the presence of extreme values were explored using Stem-and-Leaf plots in IBM SPSS Statistics for Windows, version 24 (IBM Corp.).Statistical Analyses
To explore associations between measures of trait impulsivity and the two indices of decisional impulsivity (at baseline and follow-up, respectively) and their relationship with brain functional activity, linear partial correlation analyses were performed using IBM SPSS, version 24, controlling for age, sex, IQ, and site. As we explored associations between three or two facets/indices of impulsivity and with brain functional activity at baseline and follow-up, respectively, false discovery rate (FDR)–corrected p values are reported to address issues of multiple testing. As the CGT was assessed only at baseline, no correlational analyses could be conducted for follow-up data.
To assess changes in relationships from adolescence to young adulthood, linear mixed-effect models were fitted to the data using the nlme package in R software, version 3.5.1 (https://www.r-project.org/). Random intercepts and slopes were included. Measures of trait impulsivity and decisional impulsivity were entered as dependent variables separately. Brain functional measures (during inhibition error processing and reward anticipation, respectively) were used as independent variables. Of special interest for exploring developmental changes in associations was the interaction term between visit (baseline vs. follow-up) and brain activity. Furthermore, these models were subsequently used for prediction purposes. Baseline-corrected predictors were additionally entered into the model to explicitly control for confounding effects of previous measurements in longitudinal data.
Age, sex, IQ, and site were included as control variables of no interest for all mixed-effect model analyses. FDR-corrected p values (for each of the four interaction effects of interest and for the two models on predictive effects) are reported. Relevant effects of control variables on constructs of interest are presented in Figures S3–S6. As the CGT delay aversion subscale measuring probabilistic discounting was assessed only at baseline, models could not be fitted for this facet as dependent variable. For linear mixed-effect models, extreme values were not excluded owing to the possibility that they represent true random variation that might be of special interest with regard to implications for psychiatric populations. Measures analyzed in the current work have already been addressed in some previous publications for different purposes. Thereby, often a cross-sectional approach was adopted, and variables were analyzed in association with further scales and tasks implemented within the IMAGEN project not included in the current work [e.g. (
21
,42
,43
,49
,52
, 53
, 54
, 55
, 56
, 57
, 58
, 59
, 60
, 61
, 62
)]. Therefore, no correction for multiple testing was applied with regard to those earlier studies but only for statistical tests conducted within the current work.Results
Descriptive Statistics
Descriptive statistics for included variables at baseline and follow-up are presented in Table 1. Furthermore, Figure S7 shows the distribution of (predicted) values for each of the constructs after z-score standardization. During adolescence, there was a significant positive association between all dimensions of trait impulsivity and decisional impulsivity: trait impulsivity and delay discounting (rpartial = 0.094, pFDR-corrected = .007, df = 1315), trait impulsivity and delay aversion (rpartial = 0.113, pFDR-corrected < .007, df = 1105), and delay discounting and delay aversion (rpartial = 0.090, pFDR-corrected = .011, df = 1102). During young adulthood, a significant positive linear relationship was observed between trait impulsivity and delay discounting (rpartial = 0.100, pFDR-corrected = .009, df = 917).
Table 1Descriptive Statistics for Analyzed Measures (All Available Data)
| Scale | Baseline (Age 14) | Follow-up (Age 19) | ||
|---|---|---|---|---|
| n | Mean (SD) | n | Mean (SD) | |
| NEO-FFI | 2019 | 30.03 (5.60) | 1448 | 29.44 (5.81) |
| SURPS | 2014 | 2.44 (0.45) | 1428 | 2.21 (0.43) |
| TCI-R | 2011 | 26.03 (4.27) | 511 | 25.54 (3.35) |
| Kirby DDT K1 | 2027 | 0.04 (0.04) | 1434 | 0.04 (0.04) |
| Kirby DDT K2 | 2027 | 0.03 (0.03) | 1434 | 0.02 (0.03) |
| Kirby DDT K3 | 2027 | 0.02 (0.03) | 1434 | 0.02 (0.03) |
| CANTAB CGT Delay Aversion | 1670 | 0.25 (0.14) | NA | NA |
| MID Task (ROI: VS) | 1430 | 0.26 (0.32) | 749 | 0.22 (0.27) |
| SST (ROI: Pre-SMA) | 1609 | −0.04 (0.75) | 712 | −0.05 (0.74) |
| SST (ROI: IFG) | 1609 | −0.03 (0.64) | 712 | −0.04 (0.64) |
CANTAB, Cambridge Cognition Neuropsychological Test Automated Battery; DDT, delay discounting task; CGT, Cambridge Gambling Task; IFG, inferior frontal gyrus; MID, monetary incentive delay; NA, not available; NEO-FFI, NEO Five-Factor Inventory; pre-SMA, pre-supplementary motor area; ROI, region of interest; SURPS, Substance Use Risk Profile Scale; SST, stop signal task; TCI-R, Temperament and Character Inventory–Revised; VS, ventral striatum.
a Weighted mean ROI activity across all participants. Not all data were assessed for each subject included; see selection criteria in Methods and Materials (therefore, N < 2034).
Neural Correlates of Trait Impulsivity and Neurocognitive Indices of Decisional Impulsivity at Baseline (Age 14)
During adolescence, higher brain activity in pre-SMA and IFG during inhibition error processing was related to lower trait impulsivity (rpartial = −0.075, pFDR-corrrected = .042, df = 1063) (Figure 1) (for nonsignificant associations, see Figure S8). A trend-level effect was found for delay aversion (rpartial = −0.069, pFDR-corrrected = .091, df = 884). Figures S10–S16 present the results for the relationships in adolescence excluding extreme values, resulting in similar associations.
Figure 1Significant brain responsivity impulsivity associations at baseline and follow-up. Functional magnetic resonance imaging (fMRI) reward anticipation represents mean region of interest activity in ventral striatum. fMRI inhibitory control reflects weighted mean region of interest activity (confirmatory factor analysis) in pre-supplementary motor area and inferior frontal gyrus.
Neural Correlates of Trait Impulsivity and Neurocognitive Indices of Decisional Impulsivity at Follow-up (Age 19)
During young adulthood, there was a trend-level effect (positive association) between brain activity in the VS during reward anticipation with trait impulsivity (rpartial = 0.094, pFDR-corrrected = .092, df = 453) (Figure 1) (for nonsignificant associations, see Figure S9). Again, a similar association was observed when extreme values were removed before analysis (Figures S17–S21).
Changes in Associations From Baseline (Age 14) to Follow-up (Age 19)
Changes in associations between brain activity and facets of impulsivity are displayed in Tables 2 and 3. There was a trend-level effect for the interaction between visit and reward anticipation activity in the VS for trait impulsivity (F281 = 5.94, β = 0.12, pFDR-corrrected = .062), but not for delay discounting (pFDR-corrrected > .05). No significant changes in association strength were observed for neural activity in pre-SMA and IFG during inhibition error processing either for trait impulsivity or for delay discounting (pFDR-corrrected > .05). Figure 2 shows the differences in associations at baseline and follow-up. Plots of residuals and plots of random effects for mixed-effect models are shown in Figures S22 and S23.
Table 2Linear Mixed-Effect Model Results for Change in Association Between Trait Impulsivity/Delay Discounting and Neural Activity During Inhibition-Error Processing
| Effects | Trait Impulsivity | Delay Discounting (Decisional Impulsivity) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Random Effects | SD | r | SD | r | ||||||||||
| Intercept | 1.61 | – | 1.43 | – | ||||||||||
| Visit | 0.96 | −0.90 | 0.87 | −0.86 | ||||||||||
| Residual | 0.54 | – | 0.51 | – | ||||||||||
| Fixed Effects | Estimate | SE | df | t | pt | F | p Value | Estimate | SE | df | t | pt | F | p Value |
| Intercept | 0.02 | 0.98 | 1130 | 0.02 | .99 | 3.41 | .07 | 3.08 | 0.96 | 1126 | 3.22 | <.01 | 1.77 | .18 |
| Visit | <0.01 | 0.05 | 347 | 0.18 | .86 | 0.08 | .77 | 0.07 | 0.05 | 340 | 1.41 | .16 | 2.07 | .15 |
| fMRI Inhibition Error Processing | 0.55 | 1.04 | 347 | 0.53 | .60 | 9.09 | .003 | −0.78 | 1.00 | 340 | −0.78 | .44 | 2.05 | .15 |
| Age | <0.001 | <0.001 | 1130 | 0.33 | .74 | 0.02 | .88 | <−0.001 | <0.001 | 1126 | −2.03 | .04 | 2.54 | .11 |
| Sex (as Factor) | −0.004 | 0.06 | 1130 | −0.07 | .95 | 0.09 | .76 | −0.18 | 0.05 | 1126 | −3.36 | <.001 | 10.43 | .001 |
| IQ | −0.002 | 0.002 | 1130 | −0.93 | .35 | 3.94 | .05 | −0.01 | 0.002 | 1126 | −5.27 | <.0001 | 27.00 | <.0001 |
| Site (as Factor) | – | – | – | – | – | 3.19 | .002 | – | – | – | – | >.10 | 1.26 | .27 |
| Site 2 | 0.09 | 0.10 | 1130 | 0.85 | .39 | |||||||||
| Site 3 | 0.15 | 0.13 | 1130 | 1.22 | .22 | |||||||||
| Site 4 | −0.18 | 0.12 | 1130 | −1.48 | .14 | |||||||||
| Site 5 | −0.18 | 0.10 | 1130 | −1.81 | .07 | |||||||||
| Site 6 | −0.07 | 0.11 | 1130 | −0.69 | .49 | |||||||||
| Site 7 | −0.09 | 0.10 | 1130 | −0.84 | .40 | |||||||||
| Site 8 | −0.34 | 0.12 | 1130 | −2.75 | <.01 | |||||||||
| Visit × fMRI Inhibition Error Processing | −0.01 | 0.06 | 347 | −0.23 | .82 | 0.19 | .66 | 0.01 | 0.06 | 340 | 0.22 | .83 | 0.002 | .97 |
| fMRI Inhibition Error Processing × Age | <−0.001 | <0.001 | 347 | −0.71 | .48 | 0.12 | .73 | <0.001 | <0.001 | 340 | 0.40 | .69 | 0.01 | .91 |
| fMRI Inhibition Error Processing × Sex | 0.03 | 0.06 | 347 | 0.50 | .62 | 0.33 | .57 | 0.06 | 0.05 | 340 | 1.10 | .27 | 0.47 | .50 |
| fMRI Inhibition Error Processing × IQ | <0.001 | 0.002 | 347 | 0.29 | .77 | 0.62 | .43 | 0.003 | 0.002 | 340 | 1.40 | .16 | 1.28 | .26 |
| fMRI Inhibition Error Processing × Site (as Factor) | – | – | – | – | >.10 | 0.41 | .90 | – | – | – | – | – | 2.17 | .04 |
| fMRI inhibition error processing × site 2 | 0.03 | 0.09 | 340 | 0.31 | .76 | |||||||||
| fMRI inhibition error processing × site 3 | 0.23 | 0.12 | 340 | 1.99 | <.05 | |||||||||
| fMRI inhibition error processing × site 4 | 0.21 | 0.12 | 340 | 1.78 | .08 | |||||||||
| fMRI inhibition error processing × site 5 | −0.06 | 0.10 | 340 | −0.54 | .59 | |||||||||
| fMRI inhibition error processing × site 6 | −0.08 | 0.10 | 340 | −0.81 | .42 | |||||||||
| fMRI inhibition error processing × site 7 | 0.06 | 0.11 | 340 | 0.59 | .55 | |||||||||
| fMRI inhibition error processing × site 8 | −0.04 | 0.12 | 340 | −0.32 | .75 | |||||||||
Sex and site were included as factors with sex 1 and site 1 as reference, respectively.
fMRI, functional magnetic resonance imaging.
a p ≤ .1 (based on p values from F statistic).
b p ≤ .05 (based on p values from F statistic).
c p ≤ .01 (based on p values from F statistic).
d p ≤ .001 (based on p values from F statistic).
e All site effects not significant: p > .10 (based on p values from F statistic).
Table 3Linear Mixed-Effect Model Results for Change in Association Between Trait Impulsivity/Delay Discounting and Neural Activity During Reward Anticipation
| Effects | Trait Impulsivity | Delay Discounting (Decisional Impulsivity) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Random Effects | SD | r | SD | r | ||||||||||
| Intercept | 1.52 | – | 1.61 | – | ||||||||||
| Visit | 0.92 | −0.89 | 0.90 | −0.93 | ||||||||||
| Residual | 0.52 | – | 0.47 | – | ||||||||||
| Fixed Effects | Estimate | SE | df | t | pt | F | p Value | Estimate | SE | df | t | pt | F | p Value |
| Intercept | 1.14 | 0.99 | 1093 | 1.15 | .25 | 0.94 | .33 | 2.08 | 0.89 | 1091 | 2.33 | .02 | 2.07 | .15 |
| Visit | −0.07 | 0.05 | 281 | −1.33 | .19 | 4.43 | .04 | −0.07 | 0.05 | 272 | −1.41 | .16 | 5.24 | .02 |
| fMRI Reward Anticipation | 0.30 | 0.98 | 281 | 0.31 | .76 | 1.00 | .32 | −0.54 | 0.90 | 272 | −0.60 | .55 | 0.75 | .39 |
| Age | <−0.001 | <0.001 | 1093 | −0.54 | .59 | 0.30 | .58 | <−0.001 | <0.001 | 1091 | −0.87 | .39 | 0.30 | .58 |
| Sex (as Factor) | −0.04 | 0.06 | 1093 | −0.66 | .51 | 0.41 | .52 | −0.23 | 0.05 | 1091 | −4.58 | <.001 | 20.27 | <.0001 |
| IQ | −0.004 | 0.002 | 1093 | −1.98 | <.05 | 8.95 | .003 | −0.01 | 0.002 | 1091 | −5.55 | <.001 | 32.69 | <.0001 |
| Site (as Factor) | – | – | – | – | >.10 | 2.89 | .005 | – | – | – | – | >.10 | 1.11 | .35 |
| Visit × fMRI Reward Anticipation | 0.12 | 0.06 | 281 | 2.12 | .03 | 5.94 | .02 | 0.10 | 0.05 | 272 | 2.06 | .04 | 3.68 | .06 |
| fMRI Reward Anticipation × Age | <−0.001 | <0.001 | 281 | −0.59 | .55 | 0.81 | .37 | <0.001 | <0.001 | 272 | 0.33 | .74 | 0.01 | .92 |
| fMRI Reward Anticipation × Sex | −0.07 | 0.05 | 281 | −1.26 | .21 | 1.37 | .24 | −0.06 | 0.05 | 272 | −1.13 | .26 | 1.26 | .26 |
| fMRI Reward Anticipation × IQ | 0.001 | 0.002 | 281 | 0.65 | .52 | 0.21 | .64 | <0.001 | 0.002 | 272 | 0.44 | .66 | 0.06 | .81 |
| fMRI Reward Anticipation × Site (as Factor) | – | – | – | – | >.10 | 0.47 | .86 | – | – | – | – | >.10 | 0.60 | .76 |
Sex and site were included as factors with sex 1 and site 1 as reference, respectively.
fMRI, functional magnetic resonance imaging.
a p ≤ .1 (based on p values from F statistic).
b p ≤ .05 (based on p values from F statistic).
c p ≤ .01 (based on p values from F statistic).
d p ≤ .001 (based on p values from F statistic).
e All site effects not significant: p > .10 (based on p values from F statistic).
Figure 2Plots of changes of brain responsivity impulsivity relationships from adolescence to young adulthood. (A) Association between trait impulsivity and functional magnetic resonance imaging (fMRI) inhibitory control. (B) Association between temporal discounting and fMRI inhibitory control. (C) Association between trait impulsivity and fMRI reward anticipation. (D) Association between temporal discounting and fMRI reward anticipation. Red dots and lines represent data in adolescence (baseline). Turquoise dots and lines show data in young adulthood (follow-up). Δ indicates change (in association), also highlighted by arrows; ∗p ≤ .05, n.s., not significant (significance value refers to change in association).
Predictive Value of Neural Activation Patterns at Baseline (Age 14) for Impulsivity at Follow-up (Age 19)
Using mixed-effect models, we explored whether brain activity during adolescence predicted either trait impulsivity or delay discounting during young adulthood. For delay discounting, a trend-level predictive effect of neural activity in pre-SMA and IFG during inhibition error processing was found (F488 = 5.92, β = 0.12, pFDR-corrrected = .061). Tables 4 and 5 present the results from the mixed-effect model analyses including baseline-corrected predictors.
Table 4Linear Mixed-Effect Model Results for Predicting Trait Impulsivity Based on Neural Activity
| Effects | Trait Impulsivity | ||||||
|---|---|---|---|---|---|---|---|
| Random Effects | SD | r | |||||
| Intercept | 0.64 | – | |||||
| Visit | 0.32 | 0 | |||||
| Residual | 0.24 | – | |||||
| Fixed Effects | Estimate | SE | df | t | pt | F | p Value |
| Intercept | 3.35 | 1.46 | 500 | 2.29 | .02 | 0.02 | .89 |
| Trait Impulsivity (Baseline) | 0.25 | 0.04 | 500 | 5.93 | <.001 | 40.89 | <.0001 |
| fMRI Reward Anticipation (Baseline) | −0.04 | 0.04 | 500 | −0.86 | .39 | 0.44 | .51 |
| fMRI Inhibition Error Processing (Baseline) | −0.07 | 0.05 | 500 | −1.60 | .11 | 3.61 | .06 |
| Age | <−0.001 | <0.001 | 500 | −1.98 | <.05 | 3.58 | .06 |
| Sex (as Factor) | −0.07 | 0.08 | 500 | −0.83 | .41 | 0.45 | .51 |
| IQ | −0.006 | 0.003 | 500 | −1.68 | .09 | 3.31 | .07 |
| Site (as Factor) | – | – | 500 | – | <.10 | 0.43 | .88 |
Sex and site were included as factors with sex 1 and site 1 as reference, respectively.
fMRI, functional magnetic resonance imaging.
a p ≤ .1 (based on p values from F statistic).
b p ≤ .001 (based on p values from F statistic).
Table 5Linear Mixed-Effect Model Results for Predicting Delay Discounting Based on Neural Activity
| Effects | Delay Discounting (Decisional Impulsivity) | ||||||
|---|---|---|---|---|---|---|---|
| Random Effects | SD | r | |||||
| Intercept | 0.72 | – | |||||
| Visit | 0.36 | 0 | |||||
| Residual | 0.27 | – | |||||
| Fixed Effects | Estimate | SE | df | t | pt | F | p Value |
| Intercept | 4.85 | 1.52 | 488 | 3.19 | .002 | 0.76 | .38 |
| Delay Discounting (Baseline) | 0.27 | 0.04 | 488 | 6.31 | <.001 | 49.02 | <.0001 |
| fMRI Reward Anticipation (Baseline) | −0.003 | 0.05 | 488 | −0.07 | .94 | 0.04 | .84 |
| fMRI Inhibition Error Processing (Baseline) | 0.12 | 0.05 | 488 | 2.56 | .01 | 5.92 | .02 |
| Age | <−0.001 | <0.001 | 488 | −2.36 | .02 | 4.53 | .03 |
| Sex (as Factor) | −0.23 | 0.09 | 488 | −2.61 | <.01 | 5.57 | .02 |
| IQ | −0.01 | 0.003 | 488 | −3.37 | <.001 | 13.36 | <.001 |
| Site (as Factor) | – | – | – | – | <.10 | 1.03 | .41 |
Sex and site were included as factors with sex 1 and site 1 as reference, respectively.
fMRI, functional magnetic resonance imaging.
a p ≤ .05 (based on p values from F statistic).
b p ≤ .001 (based on p values from F statistic).
Post Hoc Analyses
Simple-slope analyses were conducted for the four mixed models exploring changes from adolescence (baseline) to young adulthood (follow-up). Results indicate that decreases in impulsivity measures depend on fMRI brain activity during reward anticipation in the two respective models (Supplement H). In further post hoc analyses, partial correlations between baseline and follow-up measurements were calculated to explore associations across time controlling for age, IQ, sex, and site. Especially, we found low, nonsignificant correlations for neural activity measures between baseline and follow-up (see Supplement H for details).
Discussion
Summary of Results and Interpretation
Within our large population-based sample of adolescents assessed longitudinally at age 14 and assessed again as young adults (age 19), we found significant associations between all measures of trait and decisional impulsivity (delay discounting and delay aversion) in adolescence. In young adulthood, a significant correlation was identified for trait impulsivity and delay discounting, as delay aversion was not assessed at follow-up. In line with previous results (
15
,63
,64
), correlations between those measures of impulsivity were rather small, indicating that although different operationalizations of component processes of impulsivity share common variance, they are characterized by requiring distinct psychological processes reflecting different subdomains of impulsivity. This conclusion is further supported by the finding that measures of different facets of impulsivity distinguished within the current work show distinct patterns of associated brain activity. During adolescence, the exploration of the relationship between trait impulsivity and two indices of decisional impulsivity with neural activity resulted in a significant, but weak negative association between neural activity in pre-SMA and IFG during inhibition error processing and trait impulsivity. Further, a nominally significant negative correlation with delay aversion was identified. During young adulthood, we found a trend for a positive association between neural activity in VS during reward anticipation and trait impulsivity. As indicated by these results, brain activity in pre-SMA and IFG during inhibition error processing might represent a possible candidate network for trait impulsivity in adolescence. However, neural correlates of inhibition error processing are not associated with the delay discounting subdomain of decisional impulsivity. This might be due to stronger associations of executive, higher-order cognitive processes of inhibition and error processing with trait-related aspects of impulsivity, rather than with cognitive aspects of decision making. Nevertheless, there is an association with delay aversion, which might be related to the fact that this construct reflects an aversion to delay owing to learned or expected negative emotions associated with waiting, thereby additionally involving affective components (12
). Further replication is needed exploring the sensitivity and specificity of those associations in other longitudinal samples (65
) that also distinguish between delay discounting and delay aversion as two different expressions of decisional impulsivity and to prove whether they result from the same underlying neurobiological processes or also involve activity in distinct neural regions (66
). As shown by the correlations with behavioral performance data from the SST, higher activity during inhibition error processing is associated with fewer omission errors (for both adolescence and young adulthood) and higher reaction times (go and stop signals) as well as more commission errors (for adolescence), especially with the latter being indicative of impulsive behavior.Our findings indicate no relationship between facets of trait impulsivity and decisional forms of impulsivity with VS activity during reward anticipation in adolescence. However, we found a trend-level positive association between VS activity during reward anticipation and trait impulsivity in young adulthood, indicating a striatal hyperactivation with higher trait impulsivity scores. This finding is in line with previous meta-analytical results for healthy young adult populations (
18
), further supporting the validity of the current study results. Consequently, reward anticipation–related activity in the VS might represent a potentially relevant biomarker candidate for characterizing adult trait impulsivity. Especially, the facet of trait impulsivity is broadly associated with different brain activity patterns across the critical developmental period from adolescence to young adulthood. The current results indicate that trait impulsivity seems to represent a rather stable core construct within the broad and heterogeneous impulsivity concept.An important objective of this study was to explicitly analyze developmental trajectories in the associations of interest. We found that relationships between neural activity and measures of facets of impulsivity change across the developmental period from adolescence to young adulthood. Specifically, a trend-level significant change emerged for the association between reward anticipation activity in the VS with trait impulsivity from adolescence to young adulthood. These developmental effects are probably related to substantial changes on a neural level, whereas dimensions of trait impulsivity and decisional impulsivity remain relatively stable during the developmental period of interest as shown by post hoc simple-slope and correlational analyses. This finding is in line with results from previous studies that reported on ongoing developmental changes, especially with regard to neural reward processing, during the sensitive period of adolescence and young adulthood (
23
,67
).Finally, after controlling for multiple testing, a trend-level predictive effect of neural activity in pre-SMA and IFG during inhibition error processing in adolescence for delay discounting in young adulthood was found to have a potential practical value with regard to the prediction of development and treatment response.
Relevance of Results and Potential Clinical Implications
Impulsivity is a highly prevalent characteristic of normal as well as altered deviant human behavior, being implicated in a broad range of psychiatric disorders (
1
,2
). Therefore, the current findings are of practical relevance—although the implications are restricted owing to the rather low effect sizes. Our results have implications for the conceptualization of the multidimensional nature of impulsivity, its developmental course, and associated neural activity patterns. Further, current findings might be relevant for our understanding of the etiology, diagnosis, and treatment of impulsivity-related behavioral problems and mental health disorders.First, on a theoretical level, current results strengthen previous findings that impulsivity represents a rather heterogeneous construct being reflected in trait-focused self-report measures and further neurocognitive testing indices that on one hand share common variance, but nevertheless represent distinct psychological aspects (
6
,15
, 16
, 17
). Although not all key subdomains of impulsivity have been addressed within the current work, these findings underline the need for an informed, theoretically driven fragmentation of the construct in future research and clinical practice, including detailed assessments, analyses, and comparisons of component processes of the complex concept.Our study identified that brain activity in pre-SMA and IFG during inhibition error processing might be one candidate biomarker network characterizing the facet of trait impulsivity in adolescents. Furthermore, reward anticipated activity in the VS might possibly be a useful characteristic for characterizing adult trait impulsivity. Current results again underline the nonunitary of the concept of impulsivity and highlight that different facets of impulsivity are differentially related to distinct neural activity patterns depending on maturational stage. Relationships change substantially during the critical developmental period from adolescence to young adulthood, resulting in distinct patterns of associations at baseline and follow-up. The identified emergence of an association between reward anticipation activity in the VS and trait impulsivity in young adulthood is consistent with previous findings that report on ongoing dramatic changes with regard to neural processing until early adulthood (
67
) and results from an earlier meta-analysis (18
). These findings highlight the need for taking maturational processes into account when assessing impulsivity within and across different developmental periods and when exploring biomarker candidates [in line with (68
)].However, given the small effect sizes identified, the current results suggest a rather restricted value for practical application with regard to clinical characterization. The identified neural correlates are not yet ready to be used for diagnostic purposes (i.e., for the characterization of behavioral symptoms) or as indices of vulnerability. Nevertheless, the neural correlates identified within the current work might already be relevant for inclusion in machine learning approaches (
69
). With regard to the predictive value of neural activity for later impulsivity, results show only a trend-level effect after controlling for multiple testing not allowing for further conclusions either about early markers of vulnerability or about their sensitivity for predicting treatment response, especially as a healthy, community-based sample was analyzed.Limitations and Future Directions
A few limitations of the current work have to be mentioned. First, healthy individuals were recruited for the sample, excluding patients with diagnoses of impulsivity-related disorders. Therefore, variance within the facets of trait and decisional impulsivity might be restricted. Associations might change in extreme ranges of impulsivity dimensions. Future studies are needed that include larger samples with individuals having more extreme and clinically relevant scores on impulsivity measures for broader conclusions that additionally explore nonlinearity in those relationships. In addition, other, probably more valid, scales would be available to assess aspects of trait impulsivity [for overview, see, e.g. (
27
)], which might in future studies add further relevant knowledge to the conceptualization of trait impulsivity by probing additional facets involving, for example, aspects of emotion processing and executive dysfunction. Further, effect sizes identified within current analyses are rather small (owing to the large sample size and high power), and results obtained within mixed-effect model analyses do not explain a substantial amount of variance, questioning the relevance of the biological characteristics explored in the sense of a biomarker. This needs to be taken into account when drawing conclusions for further clinical applications. Future studies are needed that replicate current findings while allowing for conclusions with regard to early vulnerability and predictive effects. Furthermore, identified associations are correlational in nature, not yet allowing for conclusions on causal relationships. Subsequent studies are warranted replicating current findings and explicitly proving biomarker criteria [as proposed by Thome et al. (65
)]. In addition, in our study, there was a focus on specific, a priori defined neural processes of interest, also with regard to the fMRI contrasts and ROIs selected. However, others might be relevant [e.g. (34
)]. Currently, this study should be seen as a starting point, and future studies extending these results are warranted. Furthermore, using confirmatory factor analyses, data were reduced by estimating predicted values. However, a higher validity of assessment might have been obtained by combining more than one measure per concept. Also, recent articles question the test-retest reliability of the fMRI blood oxygen level–dependent signal activity [e.g. (70
, 71
, 72
)]. The resulting discussion is especially relevant when interpreting findings from longitudinal multicenter datasets focusing on developmental processes, and these should be addressed in future studies. These studies should explicitly assess test-retest reliability via adequate indices (e.g., intraclass correlation coefficient), compare indices, and then provide a framework for interpretation. Finally, earlier changes during the period from childhood to adolescence could not be analyzed. Future longitudinal projects should start assessing brain-behavior relationships in childhood (and continue assessment until late adulthood) to explore developmental trajectories across the life span.Conclusions
The current study contributes important knowledge to our understanding of the multidimensional nature of the transdiagnostic impulsivity construct, various assessment methods, and associated neural and developmental processes. In this study, intercorrelations among measures of facets of impulsivity, developmental trajectories, and neurobiological correlates were explored. We found trend-level significant associations between measures of trait and decisional impulsivity as well as with brain activity in pre-SMA and IFG during inhibition error processing and in VS during reward anticipation from adolescence to young adulthood. Associations between self-report and neurocognitive measures of impulsivity with brain activity change substantially until young adulthood and are probably due to changes on the level of neural processing, resulting in distinct developmental patterns. These findings are relevant with regard to the conceptualization of the impulsivity construct and related neurobiological markers across the life span. The current results highlight the need for taking brain developmental processes into account when exploring brain activity correlates of different operationalizations of impulsivity. However, the relationships identified are rather small. Future studies are needed for replication and extension.
Acknowledgments and Disclosures
This work was supported by the research consortium on ADHD, ESCA-Life, funded by the German Federal Ministry of Education and Research (Grant No. FKZ 01EE1408E [to TB]), European Union–funded FP6 Integrated Project IMAGEN (Reinforcement-related behaviour in normal brain function and psychopathology) (Grant No. LSHM-CT-2007-037286), Horizon 2020–funded European Research Council Advanced Grant STRATIFY (Brain network based stratification of reinforcement-related disorders) (Grant No. 695313 ), Human Brain Project (Grant Nos. HBP SGA 2, 785907 , and HBP SGA 3, 945539 ), Medical Research Council Grant c-VEDA (Consortium on Vulnerability to Externalizing Disorders and Addictions) (Grant No. MR/N000390/1 ), National Institutes of Health (NIH) (A decentralized macro and micro gene-by-environment interaction analysis of substance use behavior and its brain biomarkers) (Grant No. R01DA049238 ), National Institute for Health Research Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London, Bundesministerium für Bildung und Forschung (Grant Nos. 01GS08152, 01EV0711, and Forschungsnetz AERIAL 01EE1406A and 01EE1406B), Deutsche Forschungsgemeinschaft (Grant Nos. SM 80/7-2 , SFB 940 , TRR 265 , and NE 1383/14-1 ), Medical Research Foundation and Medical Research Council (Grant Nos. MR/R00465X/1 and MR/S020306/1 ), NIH-funded ENIGMA (Grant Nos. 5U54EB020403-05 and 1R56AG058854-01 ), Agence Nationale de la Recherche (Grant Nos. ANR-12-SAMA-0004 and AAPG2019—GeBra ), Eranet Neuron (Grant Nos. AF12-NEUR0008-01—WM2NA and ANR-18-NEUR00002-01—ADORe ), Fondation de France (Grant No. 00081242 ), Fondation pour la Recherche Médicale (Grant No. DPA20140629802 ), Mission Interministérielle de Lutte-contre-les-Drogues-et-les-Conduites-Addictives, Assistance-Publique-Hôpitaux-de-Paris and Institut National de la Santé et de la Recherche Médicale (interface grant), Paris Sud University IDEX 2012, Fondation de l’Avenir (Grant No. AP-RM-17-013), Fédération pour la Recherche sur le Cerveau, NIH, Science Foundation Ireland (Grant No. 16/ERCD/3797 ), USA (Axon, Testosterone and Mental Health during Adolescence) (Grant No. RO1 MH085772-01A1), and NIH Consortium (Grant No. U54 EB020403), supported by a cross-NIH alliance that funds Big Data to Knowledge Centres of Excellence.
The study data have been published online (September 17 2021) as part of AK’s dissertation thesis (https://doi.org/10.11588/heidok.00030332).
TB served in an advisory or consultancy role for ADHS Digital, InfectoPharm, Lundbeck, Medice, Neurim Pharmaceuticals, Oberberg GmbH, Roche, and Takeda; received conference support or speaker’s fee from Medice and Takeda; and received royalties from Hogrefe, Kohlhammer, CIP Medien, and Oxford University Press. Gareth J. Barker has received honoraria from GE Healthcare for teaching on scanner programming courses. DB serves as an unpaid scientific consultant for a European Union–funded neurofeedback trial. LP served in an advisory or consultancy role for Roche and Vifor Pharma; received speaker’s fee from Shire; and received royalties from Hogrefe, Kohlhammer, and Schattauer. The present work is unrelated to the above grants and relationships. All other authors report no biomedical financial interests or potential conflicts of interest.
Supplementary Material
- Supplementary Data
References
- Inhibition and impulsivity: Behavioral and neural basis of response control.Prog Neurobiol. 2013; 108: 44-79
- Trait impulsivity and the externalizing spectrum.Annu Rev Clin Psychol. 2017; 13: 343-368
- Behavioral models of impulsivity in relation to ADHD: Translation between clinical and preclinical studies.Clin Psychol Rev. 2006; 26: 379-395
- Impulsivity and vulnerability to psychopathology.in: Beauchaine T.P. Hinshaw S.P. Child and Adolescent Psychopathology. John Wiley & Sons, Hoboken, NJ2013: 197-226
- Impulsivity in gambling disorder and problem gambling: A meta-analysis.Neuropsychopharmacology. 2019; 44: 1354-1361
- Fractionating impulsivity: Neuropsychiatric implications.Nat Rev Neurosci. 2017; 18: 158-171
- Fractionation of impulsive and compulsive trans-diagnostic phenotypes and their longitudinal associations.Aust N Z J Psychiatry. 2019; 53: 896-907
- Trans-diagnostic measurement of impulsivity and compulsivity: A review of self-report tools.Neurosci Biobehav Rev. 2021; 120: 455-469
- The IMAGEN study: Reinforcement-related behaviour in normal brain function and psychopathology.Mol Psychiatry. 2010; 15: 1128-1139
- Neural substrates of trait impulsivity, anhedonia, and irritability: Mechanisms of heterotypic comorbidity between externalizing disorders and unipolar depression.Dev Psychopathol. 2016; 28: 1177-1208
- Disentangling symptoms of externalizing disorders in children using multiple measures and informants.Psychol Assess. 2021; 33: 1065-1079
- Delay and reward choice in ADHD: An experimental test of the role of delay aversion.Neuropsychology. 2009; 23: 367-380
- Delay or probability discounting in a model of impulsive behavior: Effect of alcohol.J Exp Anal Behav. 1999; 71: 121-143
- The latent structure of impulsivity: Impulsive choice, impulsive action, and impulsive personality traits.Psychopharmacology (Berl). 2016; 233: 3361-3370
- Toward a theory of distinct types of “impulsive” behaviors: A meta-analysis of self-report and behavioral measures.Psychol Bull. 2014; 140: 374-408
- Dimensions of impulsive behavior: Personality and behavioral measures.Pers Individ Dif. 2006; 40: 305-315
- When self-report diverges from performance: The usage of BIS-11 along with neuropsychological tests.Psychiatry Res. 2014; 218: 236-243
- Ventral-striatal responsiveness during reward anticipation in ADHD and its relation to trait impulsivity in the healthy population: A meta-analytic review of the fMRI literature.Neurosci Biobehav Rev. 2014; 38: 125-134
- Neural dynamic responses of monetary and social reward processes in adolescents.Front Hum Neurosci. 2020; 14: 141
- Neural substrates of inhibitory control maturation in adolescence.Trends Neurosci. 2019; 42: 604-616
- GABRB1 single nucleotide polymorphism associated with altered brain responses (but not performance) during measures of impulsivity and reward sensitivity in human adolescents.Front Behav Neurosci. 2017; 11: 24
- Adolescent impulsivity phenotypes characterized by distinct brain networks.Nat Neurosci. 2012; 15: 920-925
- Heritability and longitudinal stability of impulsivity in adolescence.Behav Genet. 2012; 42: 378-392
- Brain micro-architecture and disinhibition: A latent phenotyping study across 33 impulsive and compulsive behaviours.Neuropsychopharmacology. 2021; 46: 423-431
- The latent genetic structure of impulsivity and its relation to internalizing psychopathology.Psychol Sci. 2020; 31: 1025-1035
- The neurobiology of impulsivity and substance use disorders: Implications for treatment.Ann N Y Acad Sci. 2019; 1451: 71-91
- Dissociated neural substrates underlying impulsive choice and impulsive action.Neuroimage. 2016; 134: 540-549
- Recent insights into the neurobiology of impulsivity.Curr Addict Rep. 2014; 1: 309-319
- Neural activity changes associated with impulsive responding in the sustained attention to response task.PLoS One. 2013; 8e67391
- A large scale (N=102) functional neuroimaging study of response inhibition in a Go/NoGo task.Behav Brain Res. 2013; 256: 529-536
- Neural correlates of failed inhibitory control as an early marker of disordered eating in adolescents.Biol Psychiatry. 2019; 85: 956-965
- Individual differences in stop-related activity are inflated by the adaptive algorithm in the stop signal task.Hum Brain Mapp. 2018; 39: 3263-3276
- Functional connectivity predicts individual development of inhibitory control during adolescence.Cereb Cortex. 2021; 31: 2686-2700
- Brain functional connectome-based prediction of individual decision impulsivity.Cortex. 2020; 125: 288-298
- The IMAGEN study: A decade of imaging genetics in adolescents.Mol Psychiatry. 2020; 25: 2648-2671
- Stability and change in personality assessment: The revised NEO Personality Inventory in the year 2000.J Pers Assess. 1997; 68: 86-94
- Validation of a four-factor model of personality risk for substance abuse and examination of a brief instrument for assessing personality risk.Addict Biol. 2002; 7: 329-346
- The Temperament and Character Inventory-Revised.Center for Psychobiology of Personality, Washington University, St Louis, MO1999
- Delay-discounting probabilistic rewards: Rates decrease as amounts increase.Psychon Bull Rev. 1996; 3: 100-104
- Mapping motor inhibition: Conjunctive brain activations across different versions of go/no-go and stop tasks.Neuroimage. 2001; 13: 250-261
- FMRI visualization of brain activity during a monetary incentive delay task.Neuroimage. 2000; 12: 20-27
- From gene to brain to behavior: Schizophrenia-associated variation in AMBRA 1 alters impulsivity-related traits.Eur J Neurosci. 2013; 38: 2941-2945
- Do ADHD-impulsivity and BMI have shared polygenic and neural correlates?.Mol Psychiatry. 2021; 26: 1019-1028
- The neuropsychiatry of impulsivity.Curr Opin Psychiatry. 2007; 20: 255-261
- Analytical strategies for large imaging genetic datasets: Experiences from the IMAGEN study.Ann N Y Acad Sci. 2013; 1282: 92-106
- Dissociation of response inhibition and performance monitoring in the stop signal task using event-related fMRI.Hum Brain Mapp. 2007; 28: 1347-1358
- Error detection in the stop signal task.Neuroimage. 2010; 53: 664-673
- An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets.Neuroimage. 2003; 19: 1233-1239
- Sex differences in COMT polymorphism effects on prefrontal inhibitory control in adolescence.Neuropsychopharmacology. 2014; 39: 2560-2569
- The anticipation and outcome phases of reward and loss processing: A neuroimaging meta-analysis of the monetary incentive delay task.Hum Brain Mapp. 2018; 39: 3398-3418
- Accumbens functional connectivity during reward mediates sensation-seeking and alcohol use in high-risk youth.Drug Alcohol Depend. 2013; 128: 130-139
- Personality and substance use: Psychometric evaluation and validation of the Substance Use Risk Profile Scale (SURPS) in English, Irish, French, and German adolescents.Alcohol Clin Exp Res. 2015; 39: 2234-2248
- Predicting change trajectories of neuroticism from baseline brain structure using whole brain analyses and latent growth curve models in adolescents.Sci Rep. 2020; 10: 1207
- Personality, attentional biases towards emotional faces and symptoms of mental disorders in an adolescent sample.PLoS One. 2015; 10e0128271
- A phenotypic structure and neural correlates of compulsive behaviors in adolescents.PLoS One. 2013; 8e80151
- Common structural correlates of trait impulsiveness and perceptual reasoning in adolescence.Hum Brain Mapp. 2013; 34: 374-383
- Differential predictors for alcohol use in adolescents as a function of familial risk.Transl Psychiatry. 2021; 11: 157
- Mapping adolescent reward anticipation, receipt, and prediction error during the monetary incentive delay task.Hum Brain Mapp. 2019; 40: 262-283
- Polygenic risk of psychosis and ventral striatal activation during reward processing in healthy adolescents.JAMA Psychiatry. 2016; 73: 852-861
- Substance use initiation, particularly alcohol, in drug-naive adolescents: Possible predictors and consequences from a large cohort naturalistic study.J Am Acad Child Adolesc Psychiatry. 2021; 60: 623-636
- A target sample of adolescents and reward processing: Same neural and behavioral correlates engaged in common paradigms?.Exp Brain Res. 2012; 223: 429-439
- Blunted ventral striatal responses to anticipated rewards foreshadow problematic drug use in novelty-seeking adolescents.Nat Commun. 2017; 8: 14140
- Measurement of constructs using self-report and behavioral lab tasks: Is there overlap in nomothetic span and construct representation for impulsivity?.Clin Psychol Rev. 2011; 31: 965-982
- Associations between self-reported impulsivity and a latent variable of impulsive action constructed from three laboratory tasks.J Exp Psychopathol. 2019; 102043808719861894
- Biomarkers for attention-deficit/hyperactivity disorder (ADHD). A consensus report of the WFSBP task force on biological markers and the World Federation of ADHD.World J Biol Psychiatry. 2012; 13: 379-400
- Delay aversion in attention deficit/hyperactivity disorder: An empirical investigation of the broader phenotype.Neuropsychologia. 2009; 47: 446-456
- The effects of age on reward magnitude processing in the monetary incentive delay task.Neuroimage. 2020; 207: 116368
- The importance of a developmental perspective in psychiatry: What do recent genetic-epidemiological findings show?.Mol Psychiatry. 2020; 25: 1631-1639
- Machine learning for precision psychiatry: Opportunities and challenges.Biol Psychiatry Cogn Neurosci Neuroimaging. 2018; 3: 223-230
- Reliability in adolescent fMRI within two years—a comparison of three tasks.Sci Rep. 2017; 7: 2287
- Test-retest reliability of longitudinal task-based fMRI: Implications for developmental studies.Dev Cogn Neurosci. 2018; 33: 17-26
- Test-retest reliability of neural correlates of response inhibition and error monitoring: An fMRI study of a stop-signal task.Front Neurosci. 2021; 15: 624911
Article Info
Publication History
Published online: February 16, 2022
Accepted:
February 4,
2022
Received in revised form:
February 3,
2022
Received:
October 5,
2021
Footnotes
Members of the IMAGEN consortium: Tobias Banaschewski, Gareth J. Barker, Arun L.W. Bokde, Sylvane Desrivières, Herta Flor, Antoine Grigis, Hugh Garavan, Penny Gowland, Andreas Heinz, Rüdiger Brühl, Jean-Luc Martinot, Marie-Laure Paillère Martinot, Eric Artiges, Frauke Nees, Dimitri Papadopoulos Orfanos, Herve Lemaitre, Tomáš Paus, Luise Poustka, Sarah Hohmann, Sabina Millenet, Juliane H. Fröhner, Lauren Robinson, Michael N. Smolka, Henrik Walter, Jeanne M. Winterer, Robert Whelan, Gunter Schumann.
Identification
Copyright
© 2022 Society of Biological Psychiatry.
User License
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
00042-8&id=gr1.jpg){kind=link}
00042-8&id=gr2.jpg){kind=link}