A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty

A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty View ORCID Profile Ziwei Cheng , View ORCID Profile Nadja R. Ging-Jehli , Maisy Tarlow , Joonhwa Kim , Henry W. Chase , Manan Arora , Lisa Bonar , Ricki Stiffler , Alex Grattery , Simona Graur , View ORCID Profile Michael J. Frank , Mary L. Phillips , Amitai Shenhav doi: https://doi.org/10.1101/2025.07.12.663075 Ziwei Cheng 1 Department of Psychology, University of California Berkeley , Berkeley, CA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Ziwei Cheng For correspondence: ziweicheng{at}berkeley.edu amitai{at}berkeley.edu Nadja R. Ging-Jehli 2 Carney Institute for Brain Science, Department of Cognitive and Psychological Sciences, Brown University , Providence, RI, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Nadja R. Ging-Jehli Maisy Tarlow 2 Carney Institute for Brain Science, Department of Cognitive and Psychological Sciences, Brown University , Providence, RI, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Joonhwa Kim 2 Carney Institute for Brain Science, Department of Cognitive and Psychological Sciences, Brown University , Providence, RI, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Henry W. Chase 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Manan Arora 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Lisa Bonar 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Ricki Stiffler 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Alex Grattery 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Simona Graur 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Michael J. Frank 2 Carney Institute for Brain Science, Department of Cognitive and Psychological Sciences, Brown University , Providence, RI, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Michael J. Frank Mary L. Phillips 3 Department of Psychiatry, University of Pittsburgh , Pittsburgh, PA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site Amitai Shenhav 1 Department of Psychology, University of California Berkeley , Berkeley, CA, United States Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: ziweicheng{at}berkeley.edu amitai{at}berkeley.edu Abstract Full Text Info/History Metrics Supplementary material Preview PDF Abstract To behave adaptively, people need to integrate information about probabilistic outcomes and balance drives to approach positive outcomes and avoid negative outcomes. However, questions remain about how uncertainty in positive and negative outcomes influence approach-avoid decision-making dynamics. To fill this gap, we developed a novel Probabilistic Approach Avoidance Task (PAAT) and characterized behavior in this task using sequential sampling models In this task, participants (Study 1: blinded mixed clinical sample N=34; Study 2: online nonpsychiatric sample N = 58) made a series of choices between pairs of options, each consisting of variable probabilities of reaching a positive outcome (monetary reward) and of reaching a negative outcome (aversive image). Participants tended to choose options that maximized the likelihood of reward and minimized the likelihood of aversive outcomes. Moreover, the weights they placed on each of these differed for choices where these likelihoods were in opposition (i.e., the riskier option was also more rewarding; incongruent trials) relative to when these were aligned (congruent trials). Computational modeling revealed that the relative influence of rewarding and aversive outcomes on choice was captured by differences in the rate of decision-relevant information accumulation. These modeling results were validated with a series of model comparisons and posterior predictive checks, demonstrating that our sequential sampling models reliably captured our behavioral data. Together, these findings improve our understanding of the influence of motivational conflict, outcome type, and levels of uncertainty on approach-avoid decision-making. Everyday behavior is guided both by the drive to seek desirable outcomes (approach motivation) and the drive to prevent undesirable outcomes (avoidance motivation) ( Atkinson, 1957 ; Schneirla, 1959 ; Elliot, 2013 ). Oftentimes, these motivations compete with one another – to increase their chances of achieving a positive outcome, a person must risk higher likelihoods of an aversive outcome occurring. In such cases of approach-avoid conflict (AAC), the person must simultaneously weigh the valence, magnitude, and likelihood of expected outcomes to make a decision ( Kirlic et al., 2017 ; Ging-Jehli et al., 2024 ). Suboptimal evaluation or integration of these factors may lead to maladaptive motivated behavior and prolonged negative mood ( Loijen et al., 2020 ; Trew, 2011 ; Angelakis & Pseftogianni, 2021 ; Akbari et al., 2022 ). Thus, understanding the cognitive mechanisms underpinning approach-avoid decision-making is of great theoretical and translational importance. Previous studies have sought to fill this gap with a variety of tasks in which participants face trade-offs between positive and negative outcomes, primarily in cases where those outcomes are expected to occur with certainty if the relevant action is selected (e.g., Aupperle et al., 2011 ; Ironside et al., 2020 ; Klaassen et al., 2021 ; Talmi et al., 2009 ; Zorowitz et al., 2019 ; Ging-Jehli et al., 2024 ; Table 1 ). Real-world approach-avoid conflicts, by contrast, typically involve outcomes that are uncertain ( Monosov, 2020 ). Some work has begun to address this by having outcomes (e.g., shock versus rewards) occur with a fixed probability, and varying the magnitude of those outcomes ( Ging-Jehli et al., 2024 ; Klaassen et al., 2021 ), or by varying the uncertainty and magnitude of approach-related outcomes, but not avoidance-related outcomes (e.g., Sierra-Mercado et al., 2015 ; Ironside et al., 2020 ; Bach 2015 ). However, this still leaves open the question of how approach-avoidance decisions vary as a function of outcome likelihood, that is, as rewarding and aversive outcomes vary from having a high likelihood of occurring to having a high likelihood of not occurring. View this table: View inline View popup Table 1. Descriptions of PAAT and previous relevant behavioral paradigms that investigated Approach-Avoid Conflict. To address these gaps and obtain computational insights into the cognitive mechanisms of approach-avoid decision-making under outcome uncertainty, the current study developed a Probabilistic Approach-Avoid Task (PAAT; Figure 1 ). On each trial in this task, participants choose between 2 options, with each option carrying some likelihood of a positive outcome (monetary reward) and some likelihood of a negative outcome (aversive image). The probabilities of positive and negative outcomes varied independently. Across trials, the task also varied whether approach and avoidance motivations were in alignment (i.e., safer options were also more likely to be rewarded; congruent trials) or were in opposition (i.e., risky options were more likely to be rewarded; incongruent trials). This additional manipulation allowed us to examine the evaluation of reward and aversive outcome probability under different contexts, comprehensively characterizing choice dynamics across conflict domain and conflict levels ( Talmi et al., 2009 ; Aupperle et al., 2011 ; Weaver et al., 2020 ). Download figure Open in new tab Figure 1. Probabilistic Approach Avoidance Task. (A) On each trial, each of 2 options had a probability of positive (money) and a negative outcome (aversive image). Participants chose which pair of wheels they prefer in the choice phase, one wheel was chosen at random to remain on the screen from the chosen pair, and participants viewed the outcome determined by the wheel spin. (B) In incongruent trials, the risky option had a higher probability of a rewarding outcome, whereas for congruent trials the risky option had a lower probability of a rewarding outcome. The task also included forced choice trials (24 total) where participants had only one option (a pair of a positive outcome probability and a negative outcome probability) and were cued to choose the option. These trials would be relevant for future neuroimaging analysis to understand outcome probability processing and were excluded from current behavioral analysis. To characterize the processes and dynamics underlying these probabilistic approach-avoid decisions, we fit the data with sequential sampling models (SSMs). By jointly accounting for response times (RT) and response frequencies, SSMs provide an algorithmic account of how evidence in value-based decision accumulates during the decision process (e.g., choosing to approach vs. avoid) ( Westbrook et al., 2020 ; Frank, 2025 ; Turner et al., 2018 ). These models were originally primarily applied to characterize perceptual decision-making processes and have since been shown to accurately capture behavior across a wide range of domains, including value-based decision-making tasks in neuroscience and economics ( Gluth et al., 2012 ; Krajbich & Rangel, 2011 ; Milosavljevic et al., 2010 ; Krajbich, 2019 ; Ging-Jehli et al., 2023, 2025 ). Previous studies have also demonstrated within-trial accumulation of option values or incentives in neural recordings ( Rich & Wallis, 2016 ; Doi et al., 2020 ), and evidence accumulation models that formalize these latent dynamics can capture more information than classic analyses or other distribution models that primarily focus on response times ( Clithero, 2018 ; Forstmann et al., 2016 ; Fengler et al., 2021 ; Ging-Jehli et al., 2021 ). Moreover, they have been applied in clinical contexts, such as psychiatric diagnosis and assessment (e.g., Ging-Jehli et al., 2021 , 2022 ; Wiecki et al., 2015 ; Pedersen et al., 2021 ). Through these applications, the modeling parameters have gained well-established psychological interpretations ( Forstmann et al., 2016 ; Ging-Jehli et al., 2021 ), enabling us to formulate testable hypotheses about the influence of task variables (e.g., uncertainty, outcome type) on specific decision parameters. To effectively establish the underlying decision process and disentangle the influence of each task variable, we compared SSMs with different specifications about the decision dynamics (e.g., drift diffusion model, linear collapse model). Based on past findings ( Talmi et al., 2009 ; Klaassen et al., 2021 ; Ging-Jehli et al., 2024 ), we predicted that people would evaluate and integrate information about the motivational context, probability of reward, and probability of aversive outcome into decision-making. Across two studies (Study 1: blinded balanced sample of patients with obsessive-compulsive disorder and nonpsychiatric controls; Study 2: online nonpsychiatric sample), we found that decisions on incongruent trials were determined by both the probability of reward and the probability of aversive outcomes, whereas congruent trials were primarily influenced by the probability of reward.Our computational models were able to account for both choices and RT distributions in the task, and the best-fitting model revealed that the probability that a choice would lead to reward increased drift rates in favor of that choice, whereas the probability of a choice leading to an aversive outcome had the opposite effect. Collectively, these findings help to validate our new paradigm and provide insights into the relative contributions of outcome type and likelihood in approach-avoid decision making. Methods and Materials Participants Study 1 Thirty-four adults were recruited from Pittsburgh, PA area and all participants provided informed consent to participate in the study (Gender: 71% females; Age: 25.2 ± 4.5 years; OCD diagnoses: 50%; Supplement Table S1). The sample was part of an ongoing study aimed at recruiting a balanced sample of patients with obsessive-compulsive disorder (OCD) and individuals without mental health conditions (see Supplement exclusion and inclusion criteria). The diagnostic status of the participants is blinded to the investigators, and clinical variables are not considered in the current investigation. Study 2 Power analysis ( Murayama et al., 2022 ) on Study 1 found that 43 participants were required to detect the smallest effect of interest (motivation congruency effect on response time; t = 2.57) with 80% power, and we collected 58 participants online through Prolific. 8 participants were excluded for choosing the option with lower reward and higher aversive probability on more than 20% of congruent trials or completing fewer than 80% of trials. The remaining 50 participants (Gender: 46% Female; Age: 37±10 years; Supplement Table S1) were included in all of our analyses. Task We developed a novel Probabilistic Approach Avoidance Task (PAAT, Figure 1 ) to parametrically vary uncertainty in positive and negative outcomes respectively. On each two-alternative choice trials, participants chose between pairs of options, and each option included a probability of a rewarding outcome ($0.50 for Study 1 and $0.10 for Study 2 monetary gain indicated by a green token) and a probability of an aversive outcome (aversive image). Outcome probabilities were represented as wedges in a wheel, where green represented rewarding outcome probability and red represented aversive outcome probability. Each trial has 3 main phases: choice, anticipation and outcome. After participants chose an option in each trial (choice phase), one of the wheels (green or red) in the chosen pair would be randomly selected (50% probability) to remain on the screen (choice shown). A pointer appeared at the top of the selected wheel, which spun for 4 seconds (anticipation phase), and participants were shown the outcome determined by the wheel spin for 2 seconds (outcome phase). The possible outcomes of the green wheel included reward (if stopped at the green wedge) or no reward (if stopped at the empty wedge), and the possible outcomes of the red wheel included aversive image (if stopped at the red wedge) or neutral image (if stopped at the empty wedge). The response deadline during the choice period was 6 seconds in each trial, and participants would view an aversive image if they did not make a response in time. Across trials, we varied the probability of reward and aversive outcomes associated with each option (10-90% probability range at 20% increments) such that the options varied in the extent to which one was more likely to yield a reward and/or more likely to yield an aversive outcome. We counter-balanced the relative probabilities between the left and right options across trials. Trials in which one option had both a higher probability of reward and a higher probability of an aversive outcome were classified as incongruent , placing the approach-avoid motivations into conflict. In contrast, trials where one option had a higher probability of reward and a lower probability of an aversive outcome were classified as congruent . Procedures Study 1 Participants completed a Probabilistic Approach and Avoidance Task (PAAT) as part of a larger ongoing clinical study. Before the task started, participants received detailed instructions and rated stimuli from the International Affective Picture System ( Lang et al., 2008 ) for their valence and arousal (same images across participants; see Supplement for analysis of IAPS ratings). The present study was centered on the behavioral data from PAAT. Additional experimental investigations using overlapping samples and targeting independent hypotheses will be reported elsewhere. The task included a total of 120 trials, divided into 6 blocks of 20 trials each. Within each block, participants completed 16 two-alternative choice trials and 4 forced-choice trials ( Figure 1 ). Due to a change in task version, 18 participants completed the task where the percentage of incongruent trials ranged between 42% and 63%, and 16 participants completed the task where the percentage was 69%. The average percentage of incongruent and congruent trials in the sample was 60% and 40%. Supplementary analyses (Table S2) showed that results (e.g., effects of trial-level probabilities on task behavior) did not differ with these ratios on congruent (ps > .06) or incongruent trials (ps > .07), and the main analyses collapsed data across all participants. Participants were informed that they would receive a monetary bonus based on rewards gained in the task. At the end of the study, all participants were debriefed and received a $25 bonus for completing the session. Study 2 Participants completed PAAT online (implemented within the PsiTurk framework). Same as Study 1, participants received detailed instructions and rated IAPS stimuli for their valence and arousal. The task included a total of 96 (60 incongruent and 36 congruent) trials, divided into 6 blocks of 16 two-alternative choice trials each. At the end of the study, all participants were debriefed and received a monetary bonus based on rewards gained in the task. Behavioral Analysis On each trial, we defined the risky option as the option with a higher probability of an aversive outcome (and safe option as the one with a lower probability of an aversive outcome). For incongruent trials, the risky option was the one that had a higher probability of a reward, whereas for congruent trials the safe option also entailed higher likelihood of reward. We examined trial-level decision and response time in two ways: (1) using trial-wise probability of rewarding and aversive outcomes ( Pr(Rew) risky , Pr(Rew) safe , Pr(Avr) risky , and Pr(Avr) safe ) and (2) using the relative differences in probability of rewarding or aversive outcome between the risky and safe options ( Rel Rew = Pr(Rew) risky , - Pr(Rew) safe ; Rel Avr = Pr(Avr) risky - Pr(Avr) safer ). Model comparisons (in both regression and computational models) found that the relative values captured variability in behavior across PAAT trials with fewer parameters, and we focus on Rel Rew and Rel Avr in the main analysis. Detailed regression model comparisons and computational model comparisons, including alternative model specifications and results, are provided in the Supplementary Materials (Table S4 and S5). Across Rel Rew (ranges from -0.8 to 0.8) and Rel Avr (ranges from 0.1 to 0.8), we presented variations in choice frequencies, average response time (RT), and RT quantiles across trials. In addition, we fitted linear mixed-effects regression models with the lme4 package ( Bates et al., 2015 ) in R (R Core Team, 2023) to quantify the influence of outcome type and degree of outcome uncertainty on approach decisions (choosing the risky option) and RT. We fitted the regression models separately on incongruent and congruent trials, as well as on all trials with congruency included as an additional predictor (binary-coded with incongruent trials as the reference group). To model trial-wise decision (choose risky = 1, choose safe = 0), we included Rel Rew and Rel Avr as predictors. To model trial-level (log-transformed) RT, we included Rel Avr and trial-level decision (choose risky =1, choose safe=0) as predictors. Given that Rel Rew ranges from -0.8 to 0.8, and that RT may be influenced by the similarity between options, we also included the absolute value of Rel Rew (| Rel Rew |). Computational Analysis To quantify the dynamics of decision-making processes for approach-avoidance, we used the HDDM toolbox and applied different versions of Sequential Sampling Models that provided an algorithmic account of how evidence accumulation contributes to a binary approach vs. avoid decision ( Fengler et al., 2021 , 2022 ; Wiecki et al., 2013 ). Specifically, we applied simple drift diffusion model (DDM; Ratcliff, 1978 ), full-DDM ( Ratcliff & McKoon, 2008 ), linear collapse model (LCM; Cisek et al., 2009 ), and Ornstein-Uhlenbeck model (OUM; Busemeyer & Townsend, 1993 ). Each model conceptualized that on each trial, participants may begin with an initial bias towards one decision and when the options were presented, they evaluated the presented options to accumulate evidence until there was sufficient information to reach a decision threshold. This process would be captured by four main parameters with well-established psychological significance and distinct effects on behavioral patterns: drift rate (v) that reflects how quickly and effectively information is integrated to reach a decision, boundary separation (a) that reflects the required amount of evidence for reaching decisions, starting point bias (z) that reflects the inherent bias towards a particular response boundary prior to observing the trial-specific contingencies, and non-decision time (t) that reflects time spent outside the decision process (e.g., perceptual encoding or motor execution). The full DDM includes additional inter-trial parameter variability ( sv, sz, st ); the OUM includes an additional leak ( g ) on drift rate such that the accumulated evidence gradually decays back towards a neutral point; and the LCM includes an additional collapse rate parameter ( θ ) that allows for the decision boundary to decline linearly over time. The candidate models differed in how they represent the dynamics of evidence accumulation over time, and each model reflected distinct theoretical hypotheses about the decision-making process. All variants of these models allowed four main parameters to vary: (1) the rate at which evidence accumulated towards a decision (drift rate, v ); (2) the amount of evidence required to make a decision (decision threshold, a ; (3) a bias towards a particular response (starting point, z ); and (4) elements of response time unrelated to the decision process (non-decision time, t ). In addition, variants of the model allowed for the decision boundary to either remain static throughout the trial or to decline linearly over time (linear collapse model; LCM; Ging-Jehli, Kuhn et al., 2024). For each model, we used hierarchical regression to test the impact of trial-by-trial variation in Rel Pas , Rel Neg , and Cangruency (binary coded as 0=incongruent trial and 1=congruent trial) on the decision parameters. Continuous model predictors were normalized before entering the regressions. Models were run with 3 chains, each with 10,000 samples (including 5,000 samples as burn-in). The Gelman-Rubin Ȓ statistic was used to assess model convergence (Ȓ<= 1.1 indicated model convergence; Gelman & Rubin, 1992 ). To assess the significance of the fitted parameters, we computed q-values by comparing the proportion of the Bayesian posterior distributions against 0. Alternative modeling approaches that used Cumulative Prospect Theory models to investigate approach-avoidance decision dynamics in PAAT are reported in the Supplement. Nevertheless, the predictive performance and parameter recovery of the alternative models were worse compared to SSMs (Figure S4-6). Model Validation and Selection For each fitted model, we generated posterior predictive checks (PPC) to visualize potential misfits of the model by simultaneously considering response frequency and RT distribution for the responses. The best-fitting model was selected based on deviance information criterion (DIC) and PPC. Parameter Recovery After identifying and validating the best-fitting model, we performed simulation-based parameter recovery by sampling from normal distributions N (pasteriar mean, pasteriar sd) of the fitted parameters and fitting the model on synthetic data to determine whether the parameters can be recovered ( Wiecki et al, 2013 ; Wilson & Collins, 2019 ). Results To examine how sensitive approach-avoid decisions were to the likelihood of a given outcome, we parametrically varied the probabilities of rewarding and aversive outcomes between two options, one riskier (higher likelihood of a negative outcome) and one safer (lower likelihood of a negative outcome). To examine how these choice dynamics varied as a function of motivational congruency (i.e., the alignment between approach and avoid motivations), we also varied whether the safer option (lower likelihood of a negative outcome) was associated with a higher probability of a reward than the riskier option (motivationally congruent ), or whether the safer option was associated with a lower probability of reward (motivationally incongruent ). Collectively, this task enabled us to examine the influence of these varying outcome probabilities on the dynamics of these approach-avoid decisions. Model-Agnostic Analyses Study 1 Focusing first on the choice participants made on incongruent trials (where rewarding and aversive outcomes supported divergent choices), mixed-effects regressions showed that participants were more likely to select the risky option the more it increased their chance of reward compared to the safer option (higher relative reward probability; B = 2.80, 95% CI = [1.84 3.76], p<.001). At the same time, participants were less likely to choose the riskier option the more it increased their chances of an aversive outcome (higher relative aversive probability; B = -0.97, 95% CI = [-1.24 -0.69], p .10). Collapsing across all trials, we saw that participants were overall more likely to choose the safe option on congruent trials (when the safe option was more rewarding) compared to incongruent trials (when the safe option was less rewarding; B = -3.60, 95% CI = [-5.33 -1.86], p<.001; Table 2 ). Download figure Open in new tab Figure 2. Study 1 behavioral patterns in PAAT. (A) Mean rates for choosing the risky option and average response times for presented congruent (with negative, left side of the dashed line) and incongruent trials (with positive, right side of the dashed line). In the choice heatmap (top panel), the red arrow indicated the trend of increasingly choosing the risky option as the increased and decreased. In the Response Time Heatmap (bottom panel), the red arrow indicated the trend of longer response time as the absolute value of decreased. (B) Mean rates for choosing the risky option as or varied, and average response times as varied. See behavioral patterns replications from Study 2 in Supplement Figure S7. View this table: View inline View popup Download powerpoint Table 2. Study 1 mixed model results for approach choices and RT (log-transformed) using incongruent trials, congruent trials, and all trials. See Study 2 mixed model results in Supplement Table S7. Response times revealed that participants were faster to choose on incongruent trials the greater the difference in reward likelihoods between the safer and riskier option (larger | Rel Rew |; B = -0.05, 95% CI = [-0.08 -0.03], p<.001). We did not find a comparable effect for negative outcome probabilities ( Rel Avr ; B = -0.00, 95% CI = [-0.02 0.02], p = .785) ( Figure 2B ). RTs also differed depending on which option they chose, such that participants responded faster when choosing the risky relative to the safe option (B = -0.08, 95% CI = [-0.13 -0.03], p =.002). On congruent trials, we observed similar patterns such that participants chose faster as the safer option increased the chance of reward compared to the risky option (B = -0.07, 95% CI = [-0.09 -0.05], p <.001). Collapsing across all trials, participants responded faster on congruent trials (B = -0.13, 95% CI = [-0.23 -0.03], p =0.010), and participants responded slower when choosing the risky option on congruent trials compared to incongruent trials (B congruency X choice = 0.11, 95% CI = [0.05 0.18], p=.001; Table 2 ). We also investigated how outcome probabilities for each option, individual differences in reactivity to IAPS images, and recent trial outcomes may influence task behavior (Supplement - Methods). Using the outcome probability of each option as predictors (instead of the relative values) found comparable effects, such that participants preferred an option (risky or safe option) as its trial-wise probabilities of rewarding outcome increased or trial-wise probabilities of aversive outcome decreased (Table S3). Neither recent trial history nor individual differences in IAPS image ratings influenced choice behavior or subjective estimates of outcome probabilities (Supplement Results - IAPS Images). Study 2 Using data from an independent online sample, we replicated the influences of higher relative reward probability (B =1.50, 95% CI = [1.10 1.91], p<.001) and higher relative aversive probability (B = -0.44, 95% CI = [-0.61 -0.28], p<.001) on incongruent trials, as observed in Study 1 (which consists of a mixed clinical sample). On congruent trials, higher relative reward probability also influenced choices (B =1.51, 95% CI = [0.30 2.71], p=0.01). Similar to the findings from the discovery sample, participants were also more likely to choose the safe option on congruent trials compared to incongruent trials (B =-2.88, 95% CI = [-3.71 - 2.06], p<.001). We observed similar variations in RTs depending on difference in reward likelihoods and choices: participants were faster to choose the greater the difference in reward likelihoods between the safer and riskier option on incongruent (B = -0.33, 95% CI = [-0.43 - 0.24], p < .001) and congruent (B = -0.33, 95% CI = [-0.46 -0.19], p <.001) trials. Participants were faster when choosing the risky relative to the safe option on incongruent trials (B = -0.11, 95% CI = [-0.15 -0.08], p < .001), and were slower when choosing the risky option on congruent trials compared to incongruent trials (B congruency X choice = 0.10, 95% CI = [0.05 0.15], p<.001; Table S7). Model-Based Analyses We have thus far shown that variability in outcome probabilities, valence, and motivational congruency influence both the timing and outcomes of the decisions. Sequential sampling models (SSMs) enable the joint modeling of these two dependent variables (choices and RTs), to quantitatively capture changes in latent decision parameters across trials. To quantify the dynamics of approach-avoidance decision-making, we therefore fit SSMs to choice behavior in the PAAT task, with the aim of identifying the model parameterization that best accounts for these data. Study 1 We found that the best-fitting model included trial-wise influences of relative reward probability, relative aversive probability, and their interaction with motivational congruency on drift rate ( Table 3 ). Specifically, we found that, on incongruent trials, drift rate towards the risky option increased as a function of the relative probability of the rewarding outcome (B = 0.63, 95 % CI = [0.50 0.75], q-value <.001) and decreased as a function of the relative probability of the aversive outcome (B = -0.30, 95 % CI = [-0.37 -0.23], q-value <.001) ( Figure 3 ). On congruent trials, drift rate increased with higher relative reward probability to a similar degree as on incongruent trials (B = 0.49, 95 % CI = [0.33 0.64], q-value <.001). By contrast, the negative effect of the relative aversive probability was smaller and not significant (B = -0.04, 95 % CI = [-0.11 0.03], q-value = .11). Across all trials, decisions were also characterized by a biased starting point for evidence accumulation, manifesting as a default bias towards choosing the risky option ( z intercept = 0.53, 95 % CI = [0.51 0.55], q-value <.001). Supplementary analysis using individual outcome probabilities (i.e., probability of positive and negative outcome associated with each option) found similar effects, such that drift rate towards an option (risky or safe) increased as the option’s probability of reward increased, or as the option’s probability of aversive outcome decreased (Figure S3). Download figure Open in new tab Figure 3. Study 1 group-level posterior distributions of the LCM parameters. Left: higher relative reward probability of the risky option influenced drift rate towards the risky option (on both incongruent and congruent trials; dark and light greens). Higher relative aversive outcome probability influenced drift rate towards the safe option (particularly on incongruent trials; dark red). Right: visualizations of the decision bias (purple), decision boundary (blue), boundary collapse rate (yellow), and non-decision time (grey) parameter distributions. See group-level posterior distributions of the LCM parameters replications from Study 2 in Supplement S8. View this table: View inline View popup Download powerpoint Table 3. Different LCMs with the best-fitting model (selected based on DIC and PPC) represented in the first row. DIC refers to the deviance information criterion. Our best-fitting model also incorporated a threshold that collapsed linearly over the course of a decision, but was not otherwise modulated by relative reward or aversive probability or motivational congruency (see Table S4 for model comparisons). This model provided a better fit to the data than equivalent models with static thresholds and/or additional parameters than can sometimes capture similar patterns in the data (e.g., models with leaks in evidence accumulation or with inter-trial parameter variability; Table S4). Test of model validation showed that the best-fitting model showed good posterior predictive performance and parameter recovery ( Figure 4 ). For instance, simulated data captured the qualitative patterns seen on incongruent trials, whereby participants chose the risky option less frequently and had longer RTs for risky choices as conflict increased (filled red, blue, and green squares). Download figure Open in new tab Figure 4. Study 1 Computational model validation. (A). Posterior predictive check of the best-fitting model. The x-axis represents choice frequency of choosing an option (filled shapes represent choosing the risky option and unfilled shapes represent choosing the safe option). Vertical columns moving from bottom to top represent the 0.1, 0.3, 0.5, 0.7, and 0.9 quantile response times. Empirical data are represented as squares and simulated data are represented as crosses. Ellipse widths (represented by dotted lines) index standard deviation of posterior predictive distributions and indicate model estimation uncertainty. Trials were equally binned (separately for congruent and incongruent trials) into five conditions (lowest, low conflict, high, higher, and highest conflict) based on the relative probability of positive outcome. (B). Simulation results comparing generative parameters (x-axis) with recovered parameters (y-axis). See computational model validation results replications from Study 2 in Supplement Figure S9. Study 2 The best-fitting model from the discovery sample (Study 1, mixed clinical sample) also captured data from the replication sample (Study 2, online sample) well ( Table 3 ). We found that relative probability of the rewarding outcome (B = 0.59, 95 % CI = [0.48 0.69], q-value <.001) and relative probability of the aversive outcome (B = -0.16, 95 % CI = [-0.20 - 0.11], q-value <.001) modulated drift rate towards the risky option on incongruent trials, and relative probability of the rewarding outcome also modulated drift rate towards the risky option on congruent trials (B = 0.40, 95 % CI = [0.27 0.52], q-value <.001). Similar to Study 1, model validations showed that the best-fitting model had good posterior predictive performance and parameter recovery (rs = 0.26-0.90; Figure S9). Discussion The current study introduced a novel behavioral paradigm, the Probabilistic Approach-Avoidance Task (PAAT), that systematically manipulated outcome uncertainty and motivational conflict to explore how humans integrate probabilities of positive and negative outcomes during approach-avoid decision-making. We validated the paradigm by demonstrating that choice behavior in the PAAT is influenced by the probability of rewarding outcomes, the probability of aversive outcomes, and motivational congruency. We also leveraged a computational model of decision dynamics to quantify the impact of probabilistic outcomes on dissociable decision parameters. Results revealed that the default tendency to choose a risky or safe option varied across individuals. In addition, as outcome uncertainty varied, individuals adjusted evidence accumulation towards a risky or safe option to appropriately balance the drives to approach positive outcomes and avoid negative outcomes. Finally, we established that the computational model provided recoverable, subject-specific parameter estimates. Together, our model-based analysis of PAAT provided an algorithmic account of how evidence accumulation varies with outcome type and outcome uncertainty to contribute to approach-avoid decision making. Our work builds on previous studies of valued-based risky decision-making and decision-making under approach-avoid conflict, which have shown that people are sensitive to relative value differences and gamble congruency, and the speed and direction of evidence accumulation vary with decision outcomes ( Gelskov et al., 2015 ; Botvinik-Nezer et al., 2019 ; Peters & D’Esposito, 2020 ; Chen et al., 2022 ; Ironside et al., 2020 ; Pedersen et al., 2021 ; Ging-Jehli et al., 2024 ). We extend these findings to show that similar dynamics govern the influence of uncertainty about reward and aversive outcomes, particularly in cases where the two outcomes produce motivational conflict (incongruent trials). Across both samples (mixed clinical and online), participants on average displayed a default tendency towards approach before any trial-level option information was presented (i.e., starting point bias towards the riskier option). After the options were revealed, participants integrated information about potential outcomes by accumulating evidence: the relative probability of reward served as evidence supporting approach, while the relative probability of an aversive outcome supported avoidance. Although the present study was limited in examining associations between OCD symptoms and default approach tendencies or outcome integration, the modeling results provided an algorithmic account of how evidence accumulation processes contribute to approach-avoid decision-making under outcome uncertainty, with relevance for future clinical follow-up studies. On congruent trials, where approach and avoidance motivation align, people tended to choose the most valuable option (the one with higher probability of rewarding outcome and lower probability of aversive outcome) on the vast majority of trials. In spite of these choices being near ceiling, response times revealed that decision-making still varied systematically with outcome likelihoods. Our model-based analyses demonstrated that the relative probability of positive outcomes still influenced drift rate parametrically and to a similar degree on these congruent trials as on incongruent trials. The influence of aversive outcomes on drift rate was directionally consistent with their influence on incongruent trials, but smaller and non-significant. It is possible that this sample was under-powered to detect the latter effect (particularly given that there were fewer congruent than incongruent trials), so it will be important to test whether this asymmetry emerges in follow-up work. If so, it may suggest that participants either adjust their weighting of negative outcomes on congruent trials, or that they sequentially consider positive outcomes prior to negative outcomes across all trials, a possibility that bears testing with measures of attention (e.g., eye tracking) and relevant variants of SSMs (e.g., Krajbich et al., 2010 ; Ting & Gluth, 2024 ). Previous studies including congruent and incongruent trials in approach-avoidance tasks demonstrated that these trial types were helpful in identifying effective and ineffective avoidance, and the associations between behavioral flexibility and psychiatric symptoms ( Weaver et al., 2020 ). Our model comparisons likewise demonstrated improved fits for including parameters that quantify the influence of relative probabilities in both trial types relative to a model that assume an overall shift in evidence accumulation as trial type varied. Therefore, the parametric manipulation of outcome uncertainty in PAAT allowed for in-depth analysis of how individuals adjust their choices under outcome uncertainty in different motivational contexts. The variability we observe in decision parameters during approach-avoid conflict offers valuable insights into potential individual differences in the decision dynamics under outcome uncertainty, and point toward potentially fruitful directions for understanding maladaptive behavioral patterns, such as persistent avoidance of potential unlikely negative outcomes at the expense of foregoing potential rewards. For example, obsessive-compulsive symptoms may be associated with biased responding (e.g., bias towards choosing safe) and/or biased evaluation on the PAAT (e.g., heightened sensitivity to potential aversive outcomes; Gillan et al., 2014 ; Luigjes et al., 2016 ); anxiety symptoms may be associated with heightened sensitivity to potential aversive outcomes (larger absolute regression coefficient of aversive outcome on drift rate; Perkins et al., 2020 ); and depressive symptoms may be associated with reduced sensitivity to potential rewards (smaller absolute regression coefficient of rewarding outcome on drift rate; Pedersen et al., 2021 ; Ging-Jehli et al., 2024 ). Thus, investigating the relationship between diverse symptom domains and computational parameters derived from the PAAT task represents a promising avenue for future research to understand flexible or inflexible approach-avoid decisions in response to uncertain rewarding vs. aversive consequences. While the current provides critical validation for the PAAT task as a tool for modeling variability in decision dynamics based on choice behavior under approach-avoid conflict, sequential sampling models applied to our task may not generalize as a model across approach-avoid tasks (e.g., Bach, 2015 ; Zorowitz et al., 2019 ; Letkiewicz et al., 2023 ) and the source of evidence accumulation in the approach-avoid choices remains an open question. Therefore, critical next steps for future work will be to leverage neural activity measured during these decisions, as well as during the anticipation and outcome phases, to obtain a more comprehensive understanding of how individuals process the potential or actual occurrence of rewarding versus aversive outcomes ( Critchley et al., 2001 ; Oldham et al., 2018 ; Ben-Zion & Levy, 2025 ). In addition, the current study was limited by the blinding diagnostic status in investigating the relationships between OCD diagnosis or symptom severity and approach-avoidance conflict behavior. Future studies in larger, clinically diverse samples may combine PAAT with neural circuit-level computational models to identify reliable neurocomputational biomarkers. Relatedly, while the current analyses focused on the relative differences in probability of rewarding or aversive outcome between the options, future studies can examine the impact of individual probability attributes (cf. Supplement Table S3), and the brain regions involved in signaling one or more of these variables. This approach would help link neural mechanisms with specific decision parameters and translate neuroscientific findings into a principled understanding of behavior and mental health (Montague et al., 2011; Wiecki et al., 2015 ; Huys et al., 2016 ). Authors’ contributions Conceptualization: Ziwei Cheng, Nadja R. Ging-Jehli, Michael J. Frank, Mary L. Phillips, Amitai Shenhav Data curation: Maisy Tarlow, Joonhwa Kim, Henry W. Chase, Manan Arora, Lisa Bonar, Ricki Stiffler, Alex Grattery, Simona Graur Formal Analysis: Ziwei Cheng, Nadja R. Ging-Jehli Supervision: Michael J. Frank, Mary L. Phillips, Amitai Shenhav Writing – Original Draft Preparation: Ziwei Cheng, Nadja R. Ging-Jehli, Michael J. Frank, Amitai Shenhav Writing – Review & Editing: all co-authors Funding Grant P50MH106435 from the National Institute of Mental Health (M.J.F., M.L.P., A.S.) Declarations Conflicts of interest Ethics approval This research received approval from the Institutional Review Board of University of Pittsburgh. Consent to participate All participants included in the study provided informed consent. Acknowledgement We are grateful to Ilya Monosov and Steve Rasmussen for helpful discussions and valuable input. Funder Information Declared National Institute of Mental Health , P50MH106435 Footnotes ↵ ** shared senior author This version of the manuscript has been revised to update an additional replication sample; Table 1 and 3 revised; Supplemental files updated. References Abivardi , A. , Khemka , S. , & Bach , D. R . ( 2020 ). Hippocampal representation of threat features and behavior in a human approach–avoidance conflict anxiety task . Journal of Neuroscience , 40 ( 35 ), 6748 – 6758 . OpenUrl Abstract / FREE Full Text ↵ Akbari , M. , Seydavi , M. , Hosseini , Z. S. , Krafft , J. , & Levin , M. E . ( 2022 ). Experiential avoidance in depression, anxiety, obsessive-compulsive related, and posttraumatic stress disorders: A comprehensive systematic review and meta-analysis . Journal of Contextual Behavioral Science , 24 , 65 – 78 . OpenUrl ↵ Angelakis , I. , & Pseftogianni , F . ( 2021 ). Association between obsessive-compulsive and related disorders and experiential avoidance: A systematic review and meta-analysis . Journal of psychiatric research , 138 , 228 – 239 . OpenUrl CrossRef PubMed ↵ Atkinson , J. W . ( 1957 ). Motivational determinants of risk-taking behavior . Psychological Review , 64 ( 6, Pt.1 ), 359 – 372 . doi: 10.1037/h0043445 OpenUrl CrossRef ↵ Aupperle , R. L. , Sullivan , S. , Melrose , A. J. , Paulus , M. P. , & Stein , M. B . ( 2011 ). A reverse translational approach to quantify approach-avoidance conflict in humans . Behavioural brain research , 225 ( 2 ), 455 – 463 . OpenUrl CrossRef PubMed ↵ Bach , D. R . ( 2015 ). Anxiety-like behavioural inhibition is normative under environmental threat-reward correlations . PLoS computational biology , 11 ( 12 ), e1004646 . OpenUrl ↵ Bates , D. , Mächler , M. , Bolker , B. , & Walker , S . ( 2015 ). Fitting Linear Mixed-Effects Models Using lme4 . Journal of Statistical Software , 67 , 1 – 48 . doi: 10.18637/jss.v067.i01 OpenUrl CrossRef PubMed ↵ Ben-Zion , Z. , & Levy , I . ( 2025 ). Representation of Anticipated Rewards and Punishments in the Human Brain . Annual Review of Psychology , 76 . ↵ Botvinik-Nezer , R. , Iwanir , R. , Holzmeister , F. et al. fMRI data of mixed gambles from the Neuroimaging Analysis Replication and Prediction Study . Sci Data 6 , 106 ( 2019 ). doi: 10.1038/s41597-019-0113-7 OpenUrl CrossRef PubMed ↵ Busemeyer , J. R. , & Townsend , J. T . ( 1993 ). Decision field theory: A dynamic-cognitive approach to decision making in an uncertain environment . Psychological Review , 100 ( 3 ), 432 – 459 . doi: 10.1037/0033-295X.100.3.432 OpenUrl CrossRef PubMed Web of Science Camille , N. , Coricelli , G. , Sallet , J. , Pradat-Diehl , P. , Duhamel , J. R. , & Sirigu , A . ( 2004 ). The involvement of the orbitofrontal cortex in the experience of regret . Science , 304 ( 5674 ), 1167 – 1170 . OpenUrl Abstract / FREE Full Text ↵ Chen , H. Y. , Lombardi , G. , Li , S. C. , & Hare , T. A . ( 2022 ). Older adults process the probability of winning sooner but weigh it less during lottery decisions . Scientific Reports , 12 ( 1 ), 11381 . OpenUrl PubMed ↵ Cisek , P. , Puskas , G. A. , & El-Murr , S . ( 2009 ). Decisions in Changing Conditions: The Urgency-Gating Model . Journal of Neuroscience , 29 ( 37 ), 11560 – 11571 . doi: 10.1523/JNEUROSCI.1844-09.2009 OpenUrl Abstract / FREE Full Text ↵ Clithero , J. A . ( 2018 ). Response times in economics: Looking through the lens of sequential sampling models . Journal of Economic Psychology , 69 , 61 – 86 . OpenUrl ↵ Critchley , H. D. , Mathias , C. J. , & Dolan , R. J . ( 2001 ). Neural activity in the human brain relating to uncertainty and arousal during anticipation . Neuron , 29 ( 2 ), 537 – 545 . OpenUrl CrossRef PubMed Web of Science ↵ Doi , T. , Fan , Y. , Gold , J. I. , & Ding , L . ( 2020 ). The caudate nucleus contributes causally to decisions that balance reward and uncertain visual information . elife , 9 , e56694 . OpenUrl CrossRef PubMed ↵ Elliot , A. J. , Eder , A. B. , & Harmon-Jones , E . ( 2013 ). Approach–avoidance motivation and emotion: Convergence and divergence . Emotion Review , 5 ( 3 ), 308 – 311 . OpenUrl CrossRef ↵ Fengler , A. , Bera , K. , Pedersen , M. L. , & Frank , M. J . ( 2022 ). Beyond Drift Diffusion Models: Fitting a Broad Class of Decision and Reinforcement Learning Models with HDDM . Journal of Cognitive Neuroscience , 34 ( 10 ), 1780 – 1805 . doi: 10.1162/jocn_a_01902 OpenUrl CrossRef PubMed ↵ Fengler , A. , Govindarajan , L. N. , Chen , T. , & Frank , M. J . ( 2021 ). Likelihood approximation networks (LANs) for fast inference of simulation models in cognitive neuroscience . eLife , 10 , e65074 . doi: 10.7554/eLife.65074 OpenUrl CrossRef ↵ Forstmann , B. U. , Ratcliff , R. , & Wagenmakers , E. J . ( 2016 ). Sequential sampling models in cognitive neuroscience: Advantages, applications, and extensions . Annual review of psychology , 67 ( 1 ), 641 – 666 . OpenUrl CrossRef PubMed ↵ Frank , M. J . ( 2025 ). Adaptive Cost-Benefit Control Fueled by Striatal Dopamine . Annual Review of Neuroscience , 48 ( 1 ), 1 – 22 . doi: 10.1146/annurev-neuro-112723-025228 OpenUrl CrossRef PubMed ↵ Gelman , A. , & Rubin , D. B . ( 1992 ). Inference from Iterative Simulation Using Multiple Sequences . Statistical Science , 7 ( 4 ), 457 – 472 . doi: 10.1214/ss/1177011136 OpenUrl CrossRef PubMed ↵ Gelskov , S. V. , Henningsson , S. , Madsen , K. H. , Siebner , H. R. , & Ramsøy , T. Z . ( 2015 ). Amygdala signals subjective appetitiveness and aversiveness of mixed gambles . Cortex , 66 , 81 – 90 . OpenUrl CrossRef PubMed ↵ Gillan , C. M. , Morein-Zamir , S. , Urcelay , G. P. , Sule , A. , Voon , V. , Apergis-Schoute , A. M. , Fineberg , N. A. , Sahakian , B. J. , & Robbins , T. W . ( 2014 ). Enhanced avoidance habits in obsessive-compulsive disorder . Biological psychiatry , 75 ( 8 ), 631 – 638 . doi: 10.1016/j.biopsych.2013.02.002 OpenUrl CrossRef PubMed Web of Science ↵ Ging-Jehli , N. R. , Ratcliff , R. , & Arnold , L. E . ( 2021 ). Improving neurocognitive testing using computational psychiatry—A systematic review for ADHD . Psychological Bulletin , 147 ( 2 ), 169 . OpenUrl CrossRef PubMed ↵ Ging-Jehli , N. R. , Arnold , L. E. , Roley-Roberts , M. E. , & DeBeus , R . ( 2022 ). Characterizing underlying cognitive components of ADHD presentations and co-morbid diagnoses: A diffusion decision model analysis . Journal of Attention Disorders , 26 ( 5 ), 706 – 722 . OpenUrl PubMed ↵ Ging-Jehli , N. R. , Kuhn , M. , Blank , J. M. , Chanthrakumar , P. , Steinberger , D. C. , Yu , Z. , Herrington , T. M. , Dillon , D. G. , Pizzagalli , D. A. , & Frank , M. J . ( 2024 ). Cognitive Signatures of Depressive and Anhedonic Symptoms and Affective States Using Computational Modeling and Neurocognitive Testing . Biological psychiatry. Cognitive neuroscience and neuroimaging , 9 ( 7 ), 726 – 736 . doi: 10.1016/j.bpsc.2024.02.005 OpenUrl CrossRef ↵ Ging-Jehli , N. R. , Cavanagh , J. F. , Ahn , M. , Segar , D. J. , Asaad , W. F. , & Frank , M. J . ( 2025 ). Basal ganglia components have distinct computational roles in decision-making dynamics under conflict and uncertainty . PLoS biology , 23 ( 1 ), e3002978 . OpenUrl CrossRef PubMed ↵ Gluth , S. , Rieskamp , J. , & Büchel , C . ( 2012 ). Deciding when to decide: time-variant sequential sampling models explain the emergence of value-based decisions in the human brain . Journal of Neuroscience , 32 ( 31 ), 10686 – 10698 . OpenUrl Abstract / FREE Full Text ↵ Huys , Q. J. , Maia , T. V. , & Frank , M. J . ( 2016 ). Computational psychiatry as a bridge from neuroscience to clinical applications . Nature neuroscience , 19 ( 3 ), 404 – 413 . OpenUrl CrossRef PubMed ↵ Ironside , M. , Amemori , K. I. , McGrath , C. L. , Pedersen , M. L. , Kang , M. S. , Amemori , S. , Frank , M. J. , Graybiel , A. M. , & Pizzagalli , D. A . ( 2020 ). Approach-Avoidance Conflict in Major Depressive Disorder: Congruent Neural Findings in Humans and Nonhuman Primates . Biological psychiatry , 87 ( 5 ), 399 – 408 . doi: 10.1016/j.biopsych.2019.08.022 OpenUrl CrossRef PubMed Khemka , S. , Barnes , G. , Dolan , R. J. , & Bach , D. R . ( 2017 ). Dissecting the function of hippocampal oscillations in a human anxiety model . Journal of Neuroscience , 37 ( 29 ), 6869 – 6876 . OpenUrl Abstract / FREE Full Text ↵ Kirlic , N. , Young , J. , & Aupperle , R. L . ( 2017 ). Animal to human translational paradigms relevant for approach avoidance conflict decision making . Behaviour research and therapy , 96 , 14 – 29 . OpenUrl CrossRef PubMed ↵ Klaassen , F. H. , Held , L. , Figner , B. , O’Reilly , J. X. , Klumpers , F. , de Voogd , L. D. , & Roelofs , K. ( 2021 ). Defensive freezing and its relation to approach–avoidance decision-making under threat . Scientific Reports , 11 ( 1 ), 12030 . OpenUrl PubMed Klaassen , F. H. , de Voogd , L. D. , Hulsman , A. M. , O’Reilly , J. X. , Klumpers , F. , Figner , B. , & Roelofs , K. ( 2024 ). The neurocomputational link between defensive cardiac states and approach-avoidance arbitration under threat . Communications Biology , 7 ( 1 ), 576 . OpenUrl PubMed ↵ Krajbich , I. , Armel , C. , & Rangel , A . ( 2010 ). Visual fixations and the computation and comparison of value in simple choice . Nature neuroscience , 13 ( 10 ), 1292 – 1298 . OpenUrl CrossRef PubMed Web of Science ↵ Krajbich , I. , & Rangel , A . ( 2011 ). Multialternative drift-diffusion model predicts the relationship between visual fixations and choice in value-based decisions . Proceedings of the National Academy of Sciences , 108 ( 33 ), 13852 – 13857 . OpenUrl Abstract / FREE Full Text ↵ Krajbich , I . ( 2019 ). Accounting for attention in sequential sampling models of decision making . Current opinion in psychology , 29 , 6 – 11 . OpenUrl PubMed ↵ Lang , P. J. , Bradley , M. M. , & Cuthbert , B. N . ( 2008 ). International Affective Picture System (IAPS): Affective ratings of pictures and instruction manual (Tech. Rep. A-8) . Gainesville : University of Florida . ↵ Letkiewicz , A. M. , Kottler , H. C. , Shankman , S. A. , & Cochran , A. L . ( 2023 ). Quantifying aberrant approach-avoidance conflict in psychopathology: A review of computational approaches . Neuroscience & Biobehavioral Reviews , 147 , 105103 . OpenUrl PubMed ↵ Loijen , A. , Vrijsen , J. N. , Egger , J. I. , Becker , E. S. , & Rinck , M . ( 2020 ). Biased approach-avoidance tendencies in psychopathology: A systematic review of their assessment and modification . Clinical psychology review , 77 , 101825 . OpenUrl CrossRef PubMed ↵ Luigjes , J. , Figee , M. , Tobler , P. N. , Van den Brink , W. , De Kwaasteniet , B. , Van Wingen , G. , & Denys , D. ( 2016 ). Doubt in the insula: risk processing in obsessive-compulsive disorder . Frontiers in human neuroscience , 10 , 283 . OpenUrl PubMed ↵ Milosavljevic , M. , Malmaud , J. , Huth , A. , Koch , C. , & Rangel , A . ( 2010 ). The drift diffusion model can account for the accuracy and reaction time of value-based choices under high and low time pressure . Judgment and Decision making , 5 ( 6 ), 437 – 449 . OpenUrl CrossRef ↵ Monosov , I. E . ( 2020 ). How outcome uncertainty mediates attention, learning, and decision-making . Trends in neurosciences , 43 ( 10 ), 795 – 809 . OpenUrl CrossRef PubMed Montague , P. R. , Dolan , R. J. , Friston , K. J. , & Dayan , P . ( 2012 ). Computational psychiatry . Trends in cognitive sciences , 16 ( 1 ), 72 – 80 . OpenUrl CrossRef PubMed Web of Science Morris , L. , & Mansell , W . ( 2018 ). A systematic review of the relationship between rigidity/flexibility and transdiagnostic cognitive and behavioral processes that maintain psychopathology . Journal of Experimental Psychopathology , 9 ( 3 ), 2043808718779431 . OpenUrl CrossRef Moughrabi , N. , Botsford , C. , Gruichich , T. S. , Azar , A. , Heilicher , M. , Hiser , J. , Crombie , K. M. , Dunsmoor , J. E. , Stowe , Z. , & Cisler , J. M . ( 2022 ). Large-scale neural network computations and multivariate representations during approach-avoidance conflict decision-making . NeuroImage , 264 , 119709 . doi: 10.1016/j.neuroimage.2022.119709 OpenUrl CrossRef PubMed ↵ Murayama , K. , Usami , S. , & Sakaki , M . ( 2022 ). Summary-statistics-based power analysis: A new and practical method to determine sample size for mixed-effects modeling . Psychological Methods , 27 ( 6 ), 1014 – 1038 . OpenUrl PubMed ↵ Oldham , S. , Murawski , C. , Fornito , A. , Youssef , G. , Yücel , M. , & Lorenzetti , V . ( 2018 ). The anticipation and outcome phases of reward and loss processing: A neuroimaging meta analysis of the monetary incentive delay task . Human brain mapping , 39 ( 8 ), 3398 – 3418 . OpenUrl CrossRef PubMed Pagnier , G. J. , Asaad , W. F. , & Frank , M. J . ( 2024 ). Double dissociation of dopamine and subthalamic nucleus stimulation on effortful cost/benefit decision making . Current Biology , 34 ( 3 ), 655 – 660 . OpenUrl CrossRef PubMed Payne , J. W. , & Braunstein , M. L . ( 1971 ). Preferences among gambles with equal underlying distributions . Journal of Experimental Psychology , 87 ( 1 ), 13 . OpenUrl CrossRef ↵ Pedersen , M. L. , Ironside , M. , Amemori , K. I. , McGrath , C. L. , Kang , M. S. , Graybiel , A. M. , Pizzagalli , D. A. , & Frank , M. J . ( 2021 ). Computational phenotyping of brain-behavior dynamics underlying approach-avoidance conflict in major depressive disorder . PLoS computational biology , 17 ( 5 ), e1008955 . doi: 10.1371/journal.pcbi.1008955 OpenUrl CrossRef PubMed ↵ Perkins , A. M. , Bley , J. G. , Cleare , A. J. , Young , A. H. , Corr , P. J. , Dohrenbusch , R. , & Ettinger , U . ( 2020 ). Threat-sensitivity in affective disorders: A case-control study . Journal of affective disorders , 266 , 595 – 602 . OpenUrl PubMed ↵ Peters , J. , & D’Esposito , M . ( 2020 ). The drift diffusion model as the choice rule in inter-temporal and risky choice: A case study in medial orbitofrontal cortex lesion patients and controls . PLoS computational biology , 16 ( 4 ), e1007615 . OpenUrl PubMed ↵ Ratcliff , R . ( 1978 ). A theory of memory retrieval . Psychological Review , 85 ( 2 ), 59 – 108 . doi: 10.1037/0033-295X.85.2.59 OpenUrl CrossRef Web of Science ↵ Ratcliff R , McKoon G . 2008 . The diffusion decision model: theory and data for two-choice decision tasks . Neural Computation 20 : 873 – 922 . DOI: 10.1162/neco.2008.12-06-420 , PMID: 18085991 OpenUrl CrossRef PubMed Web of Science ↵ Rich , E. L. , & Wallis , J. D . ( 2016 ). Decoding subjective decisions from orbitofrontal cortex . Nature neuroscience , 19 ( 7 ), 973 – 980 . OpenUrl CrossRef PubMed ↵ Schneirla , T. C. ( 1959 ). An evolutionary and developmental theory of biphasic processes underlying approach and withdrawal . Schlund , M. W. , Brewer , A. T. , Magee , S. K. , Richman , D. M. , Solomon , S. , Ludlum , M. , & Dymond , S . ( 2016 ). The tipping point: Value differences and parallel dorsal–ventral frontal circuits gating human approach–avoidance behavior . Neuroimage , 136 , 94 – 105 . OpenUrl CrossRef PubMed ↵ Sierra-Mercado , D. , Deckersbach , T. , Arulpragasam , A. R. , Chou , T. , Rodman , A. M. , Duffy , A. , McDonald , E. J. , Eckhardt , C. A. , Corse , A. K. , Kaur , N. , Eskandar , E. N. , & Dougherty , D. D . ( 2015 ). Decision making in avoidance-reward conflict: a paradigm for non-human primates and humans . Brain structure & function , 220 ( 5 ), 2509 – 2517 . doi: 10.1007/s00429-014-0796-7 OpenUrl CrossRef PubMed Smith , R. , Kirlic , N. , Stewart , J. L. , Touthang , J. , Kuplicki , R. , Khalsa , S. S. , … & Aupperle , R. L. ( 2021 ). Greater decision uncertainty characterizes a transdiagnostic patient sample during approach-avoidance conflict: a computational modelling approach . Journal of Psychiatry and Neuroscience , 46 ( 1 ), E74 – E87 . OpenUrl Abstract / FREE Full Text ↵ Talmi , D. , Dayan , P. , Kiebel , S. J. , Frith , C. D. , & Dolan , R. J . ( 2009 ). How humans integrate the prospects of pain and reward during choice . Journal of Neuroscience , 29 ( 46 ), 14617 – 14626 . OpenUrl Abstract / FREE Full Text ↵ Ting , C. C. , & Gluth , S . ( 2024 ). Unraveling information processes of decision-making with eye-tracking data . Frontiers in Behavioral Economics , 3 , 1384713 . doi: 10.3389/frbhe.2024.1384713 OpenUrl CrossRef ↵ Trew , J. L . ( 2011 ). Exploring the roles of approach and avoidance in depression: An integrative model . Clinical psychology review , 31 ( 7 ), 1156 – 1168 . OpenUrl CrossRef PubMed ↵ Turner , B. M. , Schley , D. R. , Muller , C. , & Tsetsos , K . ( 2018 ). Competing Theories of Multialternative, Multiattribute Preferential Choice . Psychological Review , 125 ( 3 ), 329 – 362 . doi: 10.1037/rev0000089 OpenUrl CrossRef PubMed ↵ Weaver , S. S. , Kroska , E. B. , Ross , M. C. , Sartin-Tarm , A. , Sellnow , K. A. , Schaumberg , K. , Kiehl , K. A. , Koenigs , M. , & Cisler , J. M . ( 2020 ). Sacrificing reward to avoid threat: Characterizing PTSD in the context of a trauma-related approach-avoidance conflict task . Journal of abnormal psychology , 129 ( 5 ), 457 – 468 . doi: 10.1037/abn0000528 OpenUrl CrossRef PubMed ↵ Westbrook , A. , van den Bosch , R. , Maatta , J. I. , Hofmans , L. , Papadopetraki , D. , Cools , R. , & Frank , M. J. ( 2020 ). Dopamine promotes cognitive effort by biasing the benefits versus costs of cognitive work . Science , 367 ( 6484 ), 1362 – 1366 . doi: 10.1126/science.aaz5891 OpenUrl Abstract / FREE Full Text ↵ Wiecki , T. V. , Sofer , I. , & Frank , M. J . ( 2013 ). HDDM: Hierarchical Bayesian estimation of the Drift-Diffusion Model in Python . Frontiers in Neuroinformatics , 7 . doi: 10.3389/fninf.2013.00014 OpenUrl CrossRef PubMed ↵ Wiecki , T. V. , Poland , J. , & Frank , M. J . ( 2015 ). Model-based cognitive neuroscience approaches to computational psychiatry: clustering and classification . Clinical Psychological Science , 3 ( 3 ), 378 – 399 . OpenUrl CrossRef ↵ Wilson , R. C. , & Collins , A. G. ( 2019 ). Ten simple rules for the computational modeling of behavioral data . ↵ Zorowitz , S. , Rockhill , A. P. , Ellard , K. K. , Link , K. E. , Herrington , T. , Pizzagalli , D. A. , Widge , A. S. , Deckersbach , T. , & Dougherty , D. D . ( 2019 ). The Neural Basis of Approach-Avoidance Conflict: A Model Based Analysis . eNeuro , 6 ( 4 ), ENEURO.0115 – 19.2019 . doi: 10.1523/ENEURO.0115-19.2019 OpenUrl Abstract / FREE Full Text View the discussion thread. Back to top Previous Next Posted February 15, 2026. Download PDF Supplementary Material Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty Ziwei Cheng , Nadja R. Ging-Jehli , Maisy Tarlow , Joonhwa Kim , Henry W. Chase , Manan Arora , Lisa Bonar , Ricki Stiffler , Alex Grattery , Simona Graur , Michael J. Frank , Mary L. Phillips , Amitai Shenhav bioRxiv 2025.07.12.663075; doi: https://doi.org/10.1101/2025.07.12.663075 Share This Article: Copy Citation Tools A Novel Approach-Avoidance Task to Study Decision Making Under Outcome Uncertainty Ziwei Cheng , Nadja R. Ging-Jehli , Maisy Tarlow , Joonhwa Kim , Henry W. Chase , Manan Arora , Lisa Bonar , Ricki Stiffler , Alex Grattery , Simona Graur , Michael J. Frank , Mary L. Phillips , Amitai Shenhav bioRxiv 2025.07.12.663075; doi: https://doi.org/10.1101/2025.07.12.663075 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neuroscience Subject Areas All Articles Animal Behavior and Cognition (7633) Biochemistry (17680) Bioengineering (13889) Bioinformatics (41928) Biophysics (21445) Cancer Biology (18585) Cell Biology (25491) Clinical Trials (138) Developmental Biology (13374) Ecology (19897) Epidemiology (2067) Evolutionary Biology (24308) Genetics (15606) Genomics (22496) Immunology (17736) Microbiology (40385) Molecular Biology (17175) Neuroscience (88583) Paleontology (666) Pathology (2830) Pharmacology and Toxicology (4822) Physiology (7641) Plant Biology (15149) Scientific Communication and Education (2045) Synthetic Biology (4293) Systems Biology (9822) Zoology (2271)