Patients with cocaine use disorder exhibit reductions in delay discounting with episodic future thinking cues regardless of incarceration history

Highlights • Episodic future thinking (EFT) has been shown to reduce impulsive decision-making.• This EFT effect has previously been demonstrated in substance use disorders (SUDs).• No previous work has examined EFT with respect to incarceration history.• Findings show no impact of incarceration history on the EFT effect in SUD patients.• EFT-based interventions may be useful in SUDs, regardless of incarceration history.


Introduction
Choosing immediate gratification over delayed reward is a common decision.However, the rate at which future reward loses value depends upon individual differences (Odum, 2011).Psychiatric conditions characterized by behavioral disinhibition (e.g., Substance Use Disorders (SUDs), Attention-Deficit/Hyperactivity Disorder, Borderline Personality Disorder) are associated with a tendency to favor immediate gratification (Levitt et al., 2020).Delay discounting (DD), the phenomenon whereby a reward's subjective value decreases with delays, measures this tendency-with more impulsive individuals exhibiting steeper discounting rates, (i.e., reward loses value more rapidly with increasing temporal distance).Importantly, DD is associated with real-world behaviors in community samples, as well as patient populations characterized by pathological disinhibition.For example, steeper discounting rates have been linked to greater reactive thinking (i.e., emotionalimpulsive thought processes) in formerly incarcerated adults (Varghese et al., 2014), which appears to mediate the relationship between past criminal behavior and future offending (Walters, 2016).
As an analogue for real-world impulsive behavior, DD provides a proving ground for novel intervention strategies, of which episodic future thinking (EFT) manipulations appear especially promising (Rung & Madden, 2018;Scholten et al., 2019).EFT, imagining oneself in future contexts, may reduce impulsive decision-making by increasing the psychological proximity of delayed rewards.These mental simulations combine sensory details drawn from episodic memories with personal semantic information (Schacter et al., 2017), with the latter playing a more integral role with increasing temporal distance (La Corte & Piolino, 2016) and supporting variably idealized projections of the future self (Sellitto, 2020).In DD, EFT has been evoked by including cues or prompts referencing personally-relevant future events (e.g., plans, goals, anniversaries) and has been shown to reduce discounting rates in community samples and clinical populations, including SUDs (Bickel et al., 2016;Rung & Madden, 2018;Scholten et al., 2019).Importantly, such EFT manipulations have been shown to support future-conscious, real-world behaviors, such as reduced tobacco use and food intake (see Scholten et al., 2019), and may similarly have implications for substance use relapse and criminal recidivism.
While reduced discounting under EFT conditions has been demonstrated in SUD populations, including alcohol, cannabis, and cocaine use disorder (CUD; Forster, Steinhauer, et al., 2021;Mellis et al., 2019;Snider et al., 2016;Sofis et al., 2020), it has not yet been evaluated relative to personal incarceration history.Similar to individuals with SUDs, previously incarcerated individuals exhibit higher trait impulsivity (Pratt & Cullen, 2000), which contributes to diminished future orientation and steeper DD.Indeed, incarcerated individuals display more pronounced DD than non-offenders, even when controlling for substance use history (Arantes et al., 2013).Research on recently released inmates has further shown steeper DD among individuals with higher rates of reincarceration (Varghese et al., 2014).As substance use history has been associated with even greater impulsivity among incarcerated individuals (Cuomo et al., 2008), the effectiveness of EFT manipulations in individuals with a history of both SUD and incarceration is of special interest.
To examine whether the EFT effect on DD varies by incarceration history in individuals with SUDs, we utilized a personalized DD paradigm to measure discounting behavior with and without cues to evoke EFT.Participants were treatment-seeking individuals with CUD, with approximately half the sample reporting a history of incarceration.It was hypothesized that EFT cues would reduce discounting behavior in individuals with CUD, regardless of incarceration history.

Participants
Participants were treatment-seeking U.S. military Veterans, aged 18-75, diagnosed with CUD and recruited through the VA Pittsburgh Healthcare System (VAPHS).All participants provided written, informed consent to participate, as approved by the VAPHS Institutional Review Board, and reported cocaine use within the past 60 days, as well as normal or corrected-to-normal a.udiovisual acuity.Exclusion criteria included history of severe traumatic brain injury, neurological disease, or convulsive therapy, and moderate-to-severe cognitive impairment.Participants tested negative for non-prescribed, controlled substances via urine-and/or saliva-based point-of-care testing and negative for alcohol via breath-or saliva-based testing prior to assessment.Compensation was provided for time and effort.Thirty-five individuals with CUD (91% male) were included, with 19 participants reporting a history of incarceration and 16 participants reporting no incarceration history.Additional demographic and clinical characteristics are presented in Table 1.

Procedures and measures
Participants completed assessments prior to randomization into a trial of Contingency Management at VAPHS (see Forster, Forman, et al., 2021, for details).Assessments included the Addiction Severity Index-Lite, wherein incarceration history was queried.Methods for the personalized DD task were generally as described in Forster, Steinhauer, et al. (2021).In brief, an interview was conducted to identify positivelyor neutrally-valenced, personally-meaningful future events at latencies approximating 1-week, 2-weeks, 1-month, 3-months, 6-months, and 1year from the testing date.As previously described (see Forster, Steinhauer, et al., 2021) and in contrast to other similar work (Chiou & Wu, 2016;Mellis et al., 2019;Snider et al., 2016;Sofis et al., 2020) episodic specificity induction techniques were not utilized to support EFT during the interview or computerized cognitive task.
Events were rated on a 7-point Likert scale for personal relevance, valence, and arousal/excitement (see Supplementary Materials) and included in a personalized DD task, programmed using E-Prime (Psychology Software Tools, Pittsburgh, PA) and administered via desktop computer.On each trial, participants decided between $10 available "today" or a different amount, available at some delay.Details of immediate and delayed options were summarized on the left and right side of the display, respectively (see Supplementary Fig. 1), and participants responded by pressing the corresponding left-versus right-most button on a response pad.Twelve blocks of trials, representing two conditions (EFT, Standard) and 6 target latencies, were presented in a randomly-determined order.Latencies for the EFT condition reflected dates of future events identified.Improving upon our previous work (Forster, Steinhauer, et al., 2021), Standard latencies were pseudo-randomly determined to approximate EFT latencies and ensure conditions were matched for chronological primacy.A staircase algorithm determined delayed values to systematically approximate the "switch point" at which preference shifted to the larger, delayed reward (maximum amount: $50).A more rigorous control for the EFT effect was also implemented by displaying dates for Standard latencies in place of personally-relevant event descriptors used in the EFT condition (see Discussion).Additional minor methodological differences, not directly relevant to current findings, will be detailed elsewhere.

Data analysis
Determination of "switch points," modeling of DD functions, and evaluation of model fit were conducted using methods previously described in Forster, Steinhauer, et al. (2021).We additionally implemented a modified version of the algorithm for detection of nonsystematic DD responses described by Johnson & Bickel (2008; see Supplementary Materials).Because EFT cues introduce variability in response contexts across latencies (potentially engendering nonsystematicity), results are presented with and without omission of these datapoints.The discounting rate parameter (k) was computed separately for EFT and Standard conditions for each participant.A natural log transformation was applied due to skewness in the distribution of k values, yielding log(k).Due to the non-normal distribution of the data, a non-parametric Aligned Rank Transform (ART) repeatedmeasures analysis of variance (rANOVA) was used for within (EFT, Standard) and between (incarceration history, no incarceration history) group statistical tests (Wobbrock et al., 2011) and post hoc comparisons (Elkin et al., 2021).Descriptive statistics further summarize subjective ratings, individually-determined task and discounting model parameters, and relevant participant characteristics for incarcerated and nonincarcerated subgroups (see Table 1).

Results
Discounting model fit was evaluated using root-mean-square-error and was comparable to that reported in our earlier work, and similar for EFT (M = 9.6, SD = 4.3) and Standard (M = 9.6, SD = 5.5) conditions, with no significant main or interaction effects identified through ART rANOVA.The same statistical approach was utilized for comparison of log(k) values and yielded a significant effect of condition (F(1,33) = 5.93, p = 0.02).The effect of incarceration history and interaction between condition and incarceration were non-significant.The significant effect of condition was driven by reduced discounting in EFT versus Standard conditions, as indicated by lower log(k) values, for those with and without a history of incarceration (see Fig. 1).However, within and between group post hoc comparisons returned no significant results, following correction for multiple comparisons.
Following exclusion of 4 non-systematic responders (see Supplementary Materials), the analysis again yielded a significant effect of condition only (F(1,29) = 6.06, p = 0.02).As before, log(k) values were lower for EFT relative to Standard for both subgroups but post hoc comparisons isolating the EFT effect within each subgroup did not survive correction for multiple comparisons.However, post hoc comparison of Standard discounting rates between groups yielded a significant result (p = 0.047, Tukey-adjusted), consistent with steeper discounting in the Standard condition for individuals with a history of incarceration.

Discussion
Considering EFT as an intervention to reduce impulsive decisionmaking, we hypothesized that EFT cues would significantly reduce DD in treatment-seeking individuals with CUD, regardless of incarceration history.Results support this hypothesis.While significantly steeper Standard discounting was identified in individuals with a history of incarceration when excluding non-systematic responders, DD did not differ when descriptors of personally-relevant future events were included.While requiring replication, our findings offer preliminary evidence that EFT manipulations can reduce discounting in individuals with CUD, regardless of personal incarceration history.Importantly, this was the case despite our relatively "hands-off" approach to evoking EFT through short textual cues describing events identified by participants.Previous work in clinical populations employed episodic specificity induction to prime or enhance elaborative mental simulation of future events and larger, more consistent EFT effects may be obtained through these methods (see, Forster, Steinhauer, et al., 2021, for additional discussion).
This work replicates our previous finding of the EFT effect in a pilot sample of 14 treatment-seeking individuals with CUD (Forster, Steinhauer, et al., 2021) and provides the first demonstration of reduced DD through an EFT manipulation wherein incarceration history has been specifically considered.Reactive, impulsive thought processes are a hallmark of criminal thinking that appears to be an instrumental link in the concatenation of predisposing and precipitating factors that lead to recidivism.While existing cognitive-behavioral interventions may serve to target reactive thinking (Walters, 2016), integrating EFT-based approaches could expand treatment options, especially in subpopulations with both criminal and substance use histories.EFT cues may be especially promising as just-in-time adaptive interventions to reduce impulsive decision-making in high-risk contexts for substance use and/ or criminal behavior.Indeed, previous work in university students showed that EFT reduced hypothetical decisions to engage in delinquent behaviors (Wu et al., 2017) and such benefits may extend to real-world decisions with potential to save lives.
Beyond contributing novel evidence that the EFT effect is robust to incarceration history, the current work also adds to the literature by utilizing a Standard condition designed to more rigorously control for EFT framing effects.In our previous study (as in the seminal work of, Peters & Büchel, 2010), Standard delays included hashtags in place of textual cues used to describe future events in the EFT condition.This design, however, framed Standard delays as latencies, while EFT delays were framed as future reference points-failing to sufficiently control for a possible 'date-delay framing'-like effect, wherein reduced discounting occurs when latencies are merely replaced with corresponding dates (DeHart & Odum, 2015).Our current findings suggest the EFT effect is present, even when controlling for 'date-delay framing.'Nevertheless, the current work is limited by its modest sample size and subtle differences in the EFT effect by incarceration history could emerge in larger samples with greater power to detect such effects.While this study cannot definitively rule out differences in the strength of the EFT effect by incarceration history, it serves to inspire future inquiry in this important area of research, which may also include consideration of current justice-involvement versus lifetime history.

Conclusions
Our findings replicate previous work demonstrating reduced DD in the context of EFT cues in individuals with CUD and further suggest this effect manifests in individuals with and without a history of incarceration.EFT-based interventions may therefore hold promise as a strategy to reduce impulsive decision-making in CUD, regardless of incarceration history.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
This work was supported by funding from IK2 CX001807/CX/CSRD VA (PI: Forster).The contents do not represent the views of the Department of Veterans Affairs, Department of Defense, or the United States Government.CRediT authorship contribution statement Taylor M. Torres: Conceptualization, Investigation, Data curation, Writingoriginal draft.Stuart R. Steinhauer: Methodology, Writingreview & editing, Supervision.Steven D. Forman: Methodology, Writingreview & editing, Supervision.Sarah E. Forster: Funding acquisition, Conceptualization, Methodology, Software, Investigation, Data curation, Formal analysis, Validation, Visualization, Writingoriginal draft, Writingreview & editing.

Fig. 1 .
Fig. 1.Plots depicting the median value of subjective reward by delay latency, as well as corresponding discounting functions and predicted values are presented separately for subgroups with and without a history of incarceration (A).A steeper discounting function is evident for Standard versus Episodic Future Thinking (EFT) conditions in both subgroups.Median values depicted were computed for the full sample and do not reflect omission of non-systematic responders.Individual values for the log-transformed discounting parameter (log(k)) by condition also reflect reduced discounting (i.e., lower values of log(k)) in EFT relative to Standard conditions for both subgroups (B).Violin plots summarize results by condition for individuals with (left) and without (right) a history of incarceration.Values for EFT and Standard conditions for each participant are presented as linked datapoints, without omission of non-systematic responders.