Hostname: page-component-745bb68f8f-mzp66 Total loading time: 0 Render date: 2025-02-08T05:07:35.311Z Has data issue: false hasContentIssue false
  • English
  • Français

Probabalistic reinforcement learning impairments predict negative symptom severity and risk for conversion in youth at clinical high-risk for psychosis

Published online by Cambridge University Press:  06 February 2025

Lauren Luther ,
Ian M. Raugh  and
Gregory P. Strauss
Lauren Luther
Affiliation:
Department of Psychology, University of Alabama at Birmingham, Birmingham, AL, USA Department of Psychology, University of Georgia, Athens, GA, USA
Ian M. Raugh
Affiliation:
Department of Psychology, University of Georgia, Athens, GA, USA Department of Psychiatry, Douglas Mental Health Institute, McGill University, Montréal, QC, Canada
Gregory P. Strauss*
Affiliation:
Department of Psychology, University of Georgia, Athens, GA, USA
*
Corresponding author: Gregory P. Strauss; Email: gstrauss@uga.edu
Rights & Permissions [Opens in a new window]

Abstract

Background

Elucidation of transphasic mechanisms (i.e., mechanisms that occur across illness phases) underlying negative symptoms could inform early intervention and prevention efforts and additionally identify treatment targets that could be effective regardless of illness stage. This study examined whether a key reinforcement learning behavioral pattern characterized by reduced difficulty learning from rewards that have been found to underlie negative symptoms in those with a schizophrenia diagnosis also contributes to negative symptoms in those at clinical high-risk (CHR) for psychosis.

Methods

CHR youth (n = 46) and 51 healthy controls (CN) completed an explicit reinforcement learning task with two phases. During the acquisition phase, participants learned to select between pairs of stimuli probabilistically reinforced with feedback indicating receipt of monetary gains or avoidance of losses. Following training, the transfer phase required participants to select between pairs of previously presented stimuli during the acquisition phase and novel stimuli without receiving feedback. These test phase pairings allowed for inferences about the contributions of prediction error and value representation mechanisms to reinforcement learning deficits.

Results

In acquisition, CHR participants displayed impaired learning from gains specifically that were associated with greater negative symptom severity. Transfer performance indicated these acquisition deficits were largely driven by value representation deficits. In addition to negative symptoms, this profile of deficits was associated with a greater risk of conversion to psychosis and lower functioning.

Conclusions

Impairments in positive reinforcement learning, specifically effectively representing reward value, may be an important transphasic mechanism of negative symptoms and a marker of psychosis liability.

Keywords

anhedoniaasocialityattenuated psychosis syndromeavolitionreward learningultra-high-risk
Type
Original Article
Information
Psychological Medicine , Volume 55 , 2025 , e28
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press

Introduction

Psychotic disorders (PD) are severe and debilitating mental illnesses that incur suffering, lead to high rates of functional disability, and present challenges for the healthcare system (Harvey & Strassing, Reference Harvey and Strassing2012; Kadakia et al., Reference Kadakia, Catillon, Fan, Williams, Marden, Anderson, Kirson and Dembek2022). Given that few individuals achieve recovery after the onset of PDs, there has been increased interest in the early identification and prevention of psychosis. PDs are often preceded by a prodromal (i.e., pre-illness) period characterized by functional decline and attenuated hallucinations and delusions that progressively worsen over time (Addington & Heinssen, Reference Addington and Heinssen2012). This period not only reflects a potential point of early intervention but also a critical window for investigating processes related to illness onset. The field’s current approach to identifying processes giving rise to psychosis has focused on tracking pluripotent (i.e., general psychopathology) risk factors, stable traits of PDs (e.g., cognitive deficits), and stable vulnerability markers that also occur in unaffected relatives (Addington, Farris, Devoe, & Metzak, Reference Addington, Farris, Devoe and Metzak2020; Cannon, Reference Cannon2005; Cannon et al., Reference Cannon, Bearden, Hollister, Rosso, Sanchez and Hadley2000). Although this approach has led to important advances, it has inherently resulted in poor specificity, a high false-positive rate, and not identified core pathophysiological processes that could serve as treatment targets (Fusar-Poli et al., Reference Fusar-Poli, Salazar De Pablo, Correll, Meyer-Lindenberg, Millan, Borgwardt, Galderisi, Bechdolf, Pfennig, Kessing, Van Amelsvoort, Nieman, Domschke, Krebs, Koutsouleris, McGuire, Do and Arango2020; Gold et al., Reference Gold, Corlett, Strauss, Schiffman, Ellman, Walker, Powers, Woods, Waltz, Silverstein and Mittal2020). To make progress in preventing the onset of PDs and the functional disability that accompanies them, the field needs novel approaches to identifying candidate mechanisms.

We propose that a novel approach to addressing this need involves investigating mechanisms underlying negative symptoms. There are several reasons why this alternative approach may be beneficial. First, among those at clinical high-risk (CHR) for psychosis, negative symptoms are one of the strongest predictors of functional outcome (Devoe, Braun, Seredynski, & Addington, Reference Devoe, Braun, Seredynski and Addington2020; Minichino et al., Reference Minichino, Francesconi, Carrión, Bevilacqua, Parisi, Rullo, Ando’, Biondi, Delle Chiaie and Cadenhead2017; Schlosser et al., Reference Schlosser, Campellone, Biagianti, Delucchi, Gard, Fulford, Stuart, Fisher, Loewy and Vinogradov2015), suggesting that a better understanding of their mechanisms has the potential to lead to new treatments. Second, negative symptoms are highly prevalent in the prodromal phase (Piskulic et al., Reference Piskulic, Addington, Cadenhead, Cannon, Cornblatt, Heinssen, Perkins, Seidman, Tsuang, Walker, Woods and McGlashan2012) and one of the earliest indicators of risk that emerge years before attenuated positive symptoms (Addington & Heinssen, Reference Addington and Heinssen2012). Third, negative symptoms are often the reason why CHR youth and their families make initial contact with the treatment system (Yung & McGorry, Reference Yung and McGorry1996) and are one of the strongest predictors of conversion (Alderman et al., Reference Alderman, Addington, Bearden, Cannon, Cornblatt, McGlashan, Perkins, Seidman, Tsuang, Walker, Woods and Cadenhead2015; Demjaha et al., Reference Demjaha, Valmaggia, Stahl, Byrne and McGuire2012; Healey et al., Reference Healey, Penn, Perkins, Woods, Keefe and Addington2018). These findings suggest that identifying mechanistic processes underlying negative symptoms has the potential to shift the identification of psychosis to earlier stages than has been possible, which is important because the majority of functional decline has already occurred by the time attenuated positive symptoms are present (Fusar-Poli et al., Reference Fusar-Poli, Borgwardt, Bechdolf, Addington, Riecher-Rössler, Schultze-Lutter, Keshavan, Wood, Ruhrmann, Seidman, Valmaggia, Cannon, Velthorst, De Haan, Cornblatt, Bonoldi, Birchwood, McGlashan, Carpenter and Yung2013). In particular, the volitional dimension of negative symptoms, as opposed to the expressive dimension, is highly predictive of poor functional outcomes in CHR (Pelletier-Baldelli, Strauss, Visser, & Mittal, Reference Pelletier-Baldelli, Strauss, Visser and Mittal2017; Strauss et al., Reference Strauss, Walker, Pelletier-Baldelli, Carter, Ellman, Schiffman, Luther, James, Berglund, Gupta, Ristanovic and Mittal2023). Although mechanisms underlying the volitional dimension have been identified in adults with PDs (Marder & Umbricht, Reference Marder and Umbricht2023; Sarkar, Reference Sarkar2015), little research has examined whether these same mechanisms also predict motivational symptoms in CHR.

Influential conceptual models developed for adults with PDs propose that volitional negative symptoms result from dysfunctional cortico-striatal circuitry and abnormalities in reward processing that impact the decision-making processes needed to initiate goal-directed activities (Kring & Barch, Reference Kring and Barch2014; Strauss, Waltz, & Gold, Reference Strauss, Waltz and Gold2014). One key reward processing area identified as being impaired in PDs and linked to negative symptoms is reinforcement learning (RL) (Waltz, Frank, Wiecki, & Gold, Reference Waltz, Frank, Wiecki and Gold2011; Waltz & Gold, Reference Waltz and Gold2007), which involves learning which stimuli and behaviors are linked to reward outcomes. Two interactive and complementary neural systems support RL. The first is mediated by the prefrontal cortex and involves rapid learning processes, such as the ability to update value representations on a trial-by-trial basis using reward feedback (Frank & Claus, Reference Frank and Claus2006; Gonzalez et al., Reference Gonzalez, Kramar, Garagoli, Rossato, Weisstaub, Cammarota and Medina2013). This system is critical for facilitating goal-directed action because it allows individuals to respond to changes in reinforcement contingencies within the environment. The second system, which is mediated by the basal ganglia, governs gradual reinforcement learning that occurs over a number of trials (Frank, Loughry, & O’Reilly, Reference Frank, Loughry and O’Reilly2001; Hélie, Ell, & Ashby, Reference Hélie, Ell and Ashby2015; Packard & Knowlton, Reference Packard and Knowlton2002). Critically, both the fast and slow RL systems make use of prediction errors (PEs), which occur when there are mismatches between obtained and expected reward outcomes. PEs are divided into those that are positive (i.e., transient bursts of dopamine cell activity that occur for better than expected outcomes) and negative (i.e., transient cessations in dopamine cell activity that occur when outcomes are worse than expected) (Cohen & Frank, Reference Cohen and Frank2009; Schultz, Reference Schultz1998).

Several studies have investigated the integrity of the rapid and gradual RL systems in PDs and their associations with negative symptoms. Individuals with PDs have consistently demonstrated impairments in rapid learning on explicit RL tasks, consistent with deficits in making adjustments in decision-making based on trial-by-trial feedback (Barch et al., Reference Barch, Carter, Gold, Johnson, Kring, MacDonald, Pizzagalli, Ragland, Silverstein and Strauss2017; Cheng et al., Reference Cheng, Tang, Li, Lau and Lee2012; Cicero, Martin, Becker, & Kerns, Reference Cicero, Martin, Becker and Kerns2014; Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Koch et al., Reference Koch, Schachtzabel, Wagner, Schikora, Schultz, Reichenbach, Sauer and Schlösser2010; Reinen et al., Reference Reinen, Van Snellenberg, Horga, Abi-Dargham, Daw and Shohamy2016; Strauss et al., Reference Strauss, Waltz and Gold2014; Waltz, Frank, Robinson, & Gold, Reference Waltz, Frank, Robinson and Gold2007; Yilmaz, Şimşek, & Gonul, Reference Yilmaz, Şimşek and Gonul2012). These rapid learning deficits are associated with reduced activation of the orbitofrontal cortex and greater negative symptom severity (Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011). In contrast, more gradual implicit RL governed by the basal ganglia appears intact in PDs (Collins et al., Reference Collins, Brown, Gold, Waltz and Frank2014, Reference Collins, Albrecht, Waltz, Gold and Frank2017); however, recruitment of regions other than the basal ganglia has been demonstrated, indicating that normal implicit RL may be accompanied by aberrant patterns of compensation (AhnAllen et al., Reference AhnAllen, Liverant, Gregor, Kamholz, Levitt, Gulliver, Pizzagalli, Koneru and Kaplan2012; Barch et al., Reference Barch, Carter, Gold, Johnson, Kring, MacDonald, Pizzagalli, Ragland, Silverstein and Strauss2017; Heerey, Bell-Warren, & Gold, Reference Heerey, Bell-Warren and Gold2008; Pratt et al., Reference Pratt, Barch, Carter, Gold, Ragland, Silverstein and MacDonald2021; Reiss et al., Reference Reiss, Campbell, Leslie, Paulus, Ryner, Polimeni, Foot and Sareen2006; Weickert et al., Reference Weickert, Goldberg, Callicott, Chen, Apud, Das, Zoltick, Egan, Meeter, Myers, Gluck, Weinberger and Mattay2009). Furthermore, RL paradigms that manipulate positive and negative feedback have shown that individuals with PDs typically have intact learning from negative feedback but impaired learning from positive feedback (Cheng et al., Reference Cheng, Tang, Li, Lau and Lee2012; Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007). This pattern of performance has been specifically associated with greater volitional negative symptom severity and can be considered a neurobehavioral recipe for reduced goal-directed activity (i.e., individuals are capable of learning what not to do to avoid negative outcomes but not what to do to obtain rewards) (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011, Reference Strauss, Waltz and Gold2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011). Multiple processes may contribute to impaired learning from positive feedback in PDs. For example, consistent with aberrant positive prediction error signaling, neuroimaging studies have shown that reduced activation of the ventral striatum occurs in response to reward outcomes and that these abnormalities are associated with greater negative symptoms (Culbreth et al., Reference Culbreth, Moran, Kandala, Westbrook and Barch2020; Prettyman et al., Reference Prettyman, Kable, Didier, Shankar, Satterthwaite, Davatzikos, Bilker, Elliott, Ruparel and Wolf2021; Wolf et al., Reference Wolf, Satterthwaite, Kantrowitz, Katchmar, Vandekar, Elliott and Ruparel2014). Alternatively, computational modeling studies that are capable of simulating the effects of the basal ganglia and orbitofrontal cortex suggest that prediction error signaling may be intact and that poor learning from positive feedback results from orbitofrontal cortex-driven deficits in value representation (Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011; Waltz & Gold, Reference Waltz, Gold, Simpson and Balsam2015).

Despite their importance, few studies have examined RL mechanisms underlying negative symptoms in CHR. Like PDs, there is evidence for intact implicit RL in CHR on tasks relying on gradual learning and basal ganglia-driven mechanisms (Spilka et al., Reference Spilka, Raugh, Berglund, Visser and Strauss2023); however, unlike PDs, individual differences in performance in gradual reinforcement learning are associated with negative symptom severity in CHR (Spilka et al., Reference Spilka, Raugh, Berglund, Visser and Strauss2023). Studies examining rapid RL and learning from positive versus negative feedback on explicit RL tasks have produced more mixed results. For example, there is some evidence for deficits in rapid RL and impaired learning from gains that are associated with greater negative symptoms and blunted prediction error signals in the ventral striatum, dorsal anterior cingulate cortex, and ventromedial prefrontal cortex (Millman et al., Reference Millman, Gallagher, Demro, Schiffman, Reeves, Gold, Rakhshan Rouhakhtar, Fitzgerald, Andorko, Redman, Buchanan, Rowland and Waltz2020; Schmidt et al., Reference Schmidt, Antoniades, Allen, Egerton, Chaddock, Borgwardt, Fusar-Poli, Roiser, Howes and McGuire2017; Waltz et al., Reference Waltz, Demro, Schiffman, Thompson, Kline, Reeves, Xu and Gold2015). However, other studies have reported intact explicit RL (Strauss et al., Reference Strauss, Datta, Armstrong, Raugh, Kraguljac and Lahti2021). Furthermore, aberrant prediction error signaling in the dorsal anterior cingulate cortex and explicit RL deficits have also been associated with greater severity of depression (Millman et al., Reference Millman, Gallagher, Demro, Schiffman, Reeves, Gold, Rakhshan Rouhakhtar, Fitzgerald, Andorko, Redman, Buchanan, Rowland and Waltz2020), calling into question whether associations between RL and negative symptoms reflect secondary processes due to mood diagnoses that are highly comorbid in CHR (Fusar-Poli et al., Reference Fusar-Poli, Nelson, Valmaggia, Yung and McGuire2014). Thus, it is currently unclear whether RL deficits play the same role in CHR in the development and maintenance of negative symptoms as they do in PDs.

To evaluate the nature of RL and its association with negative symptoms in CHR, the current study administered an explicit probabilistic RL task (PRLT) previously used in PDs to identify explicit RL components linked to negative symptoms (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012). The paradigm consists of two task phases. The first phase is termed acquisition and involves simultaneously learning 4 stimulus pairs over a number of trials. Two of the stimulus pairs assess gain learning, where the correct choice leads to a probabilistically reinforced monetary reward on 90% or 80% of trials, and incorrect choices lead to no reward. The other 2 pairs assess loss avoidance, where the correct choice allows for the avoidance of a monetary loss on 90% or 80% of trials, and an incorrect response results in a monetary loss. Importantly, both the gain and loss avoidance pairs generally result in a positive prediction error when participants select the optimal response (i.e., gain reward or avoid losing reward, respectively). This manipulation is critical for the second portion of the task, called the transfer phase, where participants select between two stimuli in the absence of feedback. Stimuli presented in the acquisition phase are presented along with novel stimuli, allowing for critical pairings such as the frequent winner versus frequent loss avoider (i.e., both stimuli are associated with positive prediction errors but only the frequent winner requires representation of positive value). Additional pairings allow for an enhanced understanding of the role of prediction error and value representation processes contributing to RL impairments: frequent winners versus infrequent winners, frequent winners versus frequent losers, and frequent loss avoiders versus infrequent winners. Past PD studies using this task have found evidence for impaired rapid learning and a selective deficit in learning from gains but intact loss avoidance learning during the acquisition phase (Gold, Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011). In the transfer phase, individuals with a PD displayed a failure to select the frequent winner over the frequent loss avoider, consistent with a deficit in representing reward value rather than prediction error signaling (Gold, Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012). Furthermore, greater negative symptoms were associated with reduced learning rate, impaired learning from gains, better loss avoidance learning, and impaired selection of the frequent winner over the frequent loss avoider in the transfer phase (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011). Based on these past studies, it was hypothesized that relative to healthy controls (CN), CHR participants would display: (1) reduced learning in the acquisition phase, consistent with impaired rapid learning; (2) impaired learning from gains, but intact loss avoidance learning; (3) deficits in value representation, as reflected in transfer phase behavior and the failure to select frequent winners over frequent loss avoiders. Additionally, the patterns in hypotheses 1–3 were expected to significantly correlate with motivational negative symptoms.

Methods

Participants

Participants included 46 CHR participants and 51 CN. CHR participants were recruited from the Georgia Psychiatric Risk Evaluation Program (G-PREP) and the New York Psychosis Risk Evaluation Program (NY-PREP). These programs perform diagnostic evaluations for youth displaying prodromal syndromes. Participants were also recruited through online and printed advertisements and through presentations or meetings with members of community mental health centers and local school systems. CHR participants met the criteria for a prodromal syndrome on the Structured Interview for Prodromal Syndromes (SIPS) (Miller et al., Reference Miller, McGlashan, Rosen, Cadenhead, Ventura, McFarlane, Perkins, Pearlson and Woods2003) (see Supplementary Material for details).

CN were recruited through online and printed advertisements. CN were eligible if they had no current major psychiatric disorder diagnosis or schizophrenia-spectrum personality disorder as established on the SCID-5 (First, Williams, Karg, et al., 2015) and SCID-5-PD (First, Williams, Benjamin, & Spitzer, 2015). CN also had no family history of psychosis and were not taking psychotropic medications. Groups did not significantly differ in age, sex, parental education, or race (See Table 1). CHR had lower education than CN.

Table 1. Sample characteristics

Note: One healthy control was missing race/ethnicity. CHR, clinical high-risk group; CN, healthy control group; M, mean; SD, standard deviation.

Procedures

After providing informed consent, participants completed clinical interviews. All participants were rated on the SCID-5. CN also completed the SCID-5-PD, and CHR was rated on the SIPS and the Brief Negative Symptom Scale (Kirkpatrick et al., Reference Kirkpatrick, Strauss, Nguyen, Fischer, Daniel, Cienfuegos and Marder2011), Clinical High-Risk Adaptation Version (Strauss & Chapman, Reference Strauss and Chapman2018). The positive symptoms SIPS subscale was used to measure positive symptoms, the SIPS dysphoric mood item was used as an index of depression, and the Global Assessment of Functioning (Hall, Reference Hall1995) modified scale from the SIPS was the measure of functioning. After completion of the rating scales, all participants completed the PRLT. Participants were compensated for study procedures at a rate of $30 per hour and received a PRLT task bonus reflecting a proportion of their monetary task winnings of $5.00. Participants provided written informed consent for a protocol approved by the University of Georgia or Binghamton University Institutional Review Boards.

For CHR participants, Zhang et al. (Reference Zhang, Xu, Tang, Li, Tang, Cui, Wei, Wang, Hu, Liu, Li, Lu, McCarley, Seidman and Wang2018)’s Shanghai at Risk for Psychosis (SHARP) cross-sectional conversion risk score was calculated. The SHARP risk calculator is a tool that gives a participant’s cross-sectional probability of converting to a formal psychotic disorder at any point in time. It is derived based on clinical variables from the SIPS that are most predictive of conversion to full psychosis, with an algorithm that weighs functional decline, positive symptoms, negative symptoms, and dysphoric mood. Higher scores on this risk calculator significantly predicted conversion probability in the original Shanghai development and validation samples (Zhang et al., 2018) as well as in independent replications (Osborne & Mittal, Reference Osborne and Mittal2019).

Measures

The Probabilistic Reinforcement Learning Task (PRLT) is based on Frank, Seeberger, & O’Reilly, Reference Frank, Seeberger and O’Reilly2004; Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012. In the acquisition phase, participants were asked to learn how four pairs of stimuli were associated with monetary gains or loss avoidance. Stimuli in this task were landscape images (see Supplementary Material), and each pair was presented separately. Two of the stimuli were associated with potential gains. Choosing the correct stimulus was probabilistically reinforced with the monetary outcome (winning money is indicated by a coin along with the word “win”). If the incorrect stimulus was chosen, participants received no reward, which was indicated by seeing the phrase “Not a winner, try again.” One of the stimulus pairs reinforced the correct response on 90% of trials, while the other pair reinforced the correct response on 80% of trials. The other two stimuli pairs required participants to learn to avoid losses. The correct response was probabilistically associated with not losing money (the feedback of “Keep your money” is shown), while the incorrect response was probabilistically associated with losing money (feedback of “Lose” was presented along with a coin with an X through it). The correct response was associated with avoiding losses on 90% of trials in one pair and 80% of trials in the other pair. After a practice session, participants completed 160 learning trials with the four stimulus pairs presented in a randomized order. Each pair was shown 40 times, and the 160 trials were divided into 4 blocks of 40 trials for analyses.

The transfer phase consisted of 64 trials where the original four stimulus pairs from the learning phase were shown 4 times. In addition, 24 novel pairings of the stimuli used in the learning phase were presented twice. Novel pairings consisted of a trained item from the acquisition phase presented with every other trained item (e.g., a stimulus from the 90% gain learning was paired with items from the 80 and 90% loss avoidance and 80% gain learning pairs). Participants were asked to choose the stimulus in each pair that they believed was “best” based on the first task phase. No feedback was provided during this phase. Following (Barch et al., Reference Barch, Carter, Gold, Johnson, Kring, MacDonald, Pizzagalli, Ragland, Silverstein and Strauss2017; Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012), pairings that allowed for differentiation of prediction error signaling and value representation were of greatest interest (see Supplementary Table 1).

The Brief Negative Symptom Scale (BNSS; Kirkpatrick et al., Reference Kirkpatrick, Strauss, Nguyen, Fischer, Daniel, Cienfuegos and Marder2011) Clinical High-Risk Adaptation Version (Strauss & Chapman, Reference Strauss and Chapman2018) was used to assess negative symptoms. The BNSS is a 13-item clinical interview-based rating scale of negative symptoms that covers the 5 NIMH consensus conference domains of anhedonia, avolition, asociality, blunted affect, and alogia. Based on confirmatory factor analysis (Strauss et al., Reference Strauss, Ruiz, Visser, Crespo and Dickinson2018), the standard 5 domain scores were computed using averages of the items comprising each domain. Internal consistency for the BNSS total score in this sample was .89 and .90 for anhedonia, .87 for avolition, .85 for asociality, .94 for blunted affect, and .97 for alogia. Interviews for clinical rating scales were administered by a licensed clinical psychologist (GPS) or examiners trained to reliability standards (>0.80) using gold-standard training videos created by the PI.

Data analysis

To test hypotheses about learning from gains and losses, a linear mixed effect model was used to compare the learning phase trials on the PRLT between all CHR and CN. Group (CHR, CN), valence (gain, loss), probability (80/20%, 90/10%), and learning block (1–4) were fixed factors, and subject-specific intercepts were random factors. Significant interactions were followed up by pairwise Tukey posthoc contrasts using the emmeans package in R. To test hypotheses about rapid learning, a learning composite score was calculated (Block 4–Block 1 accuracy; higher scores indicate greater learning of the optimal stimulus from the last block to the first block). The learning composite was compared between groups with an independent samples t-test. Hypotheses for group differences on the four novel stimulus pairings in the transfer phase were examined using one-way ANOVAs. Finally, correlations examined associations between performance in both phases and motivational negative symptom domains. Guided by our hypotheses, we focused on associations with motivational negative symptoms with the average accuracy for the gain and loss-avoidance conditions in the learning phase and the transfer performance scores for the Frequent Winner versus Frequent Loss Avoidance pairings. Supplemental exploratory associations with additional task metrics (learning composite, valence score differences; additional transfer performance scores) and clinical outcomes (cross-sectional risk of conversion to psychosis, functioning, depression, expressive negative symptoms) are presented in the Supplementary Material. The Benjamini and Hochberg correction for multiple comparisons was applied to account for the false-discovery rate. Instances where significant effects did not survive correction for multiple comparisons are noted.

Since negative symptoms are a hybrid categorical-dimensional construct (Ahmed et al., Reference Ahmed, Strauss, Buchanan, Kirkpatrick and Carpenter2015), in addition to the dimensional (correlation) analyses, negative symptom effects were also evaluated categorically across 3 groups: CN and then CHR with high or low negative symptoms using a median split with the BNSS total score (supplemental analyses were also conducted using the BNSS MAP scores).

Results

Full group analyses

Acquisition phase. The linear mixed effects model with all CHR and CN identified significant main effects of the block, probability, valence, and a trending (p = .10) main effect of group for predicting accuracy (see Table 2 and Figure 1). Post-hocs were used to interpret these effects. Post-hocs revealed that accuracy increased as probability increased (p < .01) and largely as block increased (p-values < .01, except blocks 2 and 3 did not differ, p = .99). Accuracy was significantly higher for gain than loss-avoidance conditions (p < .01).

Table 2. PRLT in CHR youth and healthy controls

Note: CHR, clinical high-risk group; DF, degrees of freedom; PRLT, probabilistic reinforcement learning task.

Figure 1. Training performance in healthy controls and clinical high-risk for psychosis samples. Mean accuracy is reported. Error bars denote the standard error of the mean (SEM).

There was also a significant interaction between probability and valence that was qualified by a significant probability X valence X group interaction. Post-hocs identified that this 3-way interaction was driven by CHR performing worse than CN in the 80% gain condition (p < .01). No significant group differences were observed for the other conditions. Thus, CHR displayed a selective deficit in learning from gains, which was driven by the 80% gain condition.

The overall learning composite did not significantly differ between CHR and CN (p = .82) (see Supplementary Table 2). In line with Gold et al. (Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012), we also examined the impact of valence on learning by calculating difference scores between the final block for the gain and loss-avoidance trials for both the 90% and 80% pairs. Positive difference scores suggest increased gain learning, and negative difference scores indicate increased loss-avoidance learning. Independent sample t-tests comparing these scores (see Supplementary Table 2) demonstrated significant group differences on the 80% pairs (t = 2.69, p < .01); CN had better learning from gains than losses, and CHR demonstrated better learning from losses than gains. Groups did not significantly differ on 90% difference scores.

Transfer phase. CHR chose stimuli associated with frequent winning over frequent losing (FW v. FL) significantly less than CN (p = .03) (see Figure 2). Groups did not significantly differ on choosing stimuli associated with frequent winning versus infrequent winning (FW v. IW) (p = .09), frequent winning versus frequent loss avoidance (p = .79) (FW v. FLA), or frequent loss avoidance versus infrequent winners (FLA v. IW) (p = .20).

Figure 2. Transfer phase performance across groups. * p < .05. FLA, frequent loss avoider; FW, frequent winner; FL, frequent loser; IW, infrequent winner. Mean accuracy is graphed. Error bars denote the standard error of the mean (SEM). *p < .05.

Correlations with Negative Symptoms. In CHR, lower average gain learning was significantly associated with greater asociality (see Table 3). Average loss avoidance learning was not significantly associated with any negative symptoms. For the test phase, a reduced likelihood of choosing stimuli associated with FW versus FLA was significantly associated with greater asociality.

Table 3. Correlations between training and transfer performance and negative symptoms in the CHR group

The Benjamini-Hochberg procedure was used to correct for multiple comparisons. All noted significant correlations were retained. BNSS, Brief Negative Symptom Scale; FW v FLA, Frequent Winner versus Frequent Loss Avoider.

Secondary Clinical Correlations. Exploratory associations with depression, functioning, and risk of conversion to psychosis are presented in Supplementary Table 3. Most notably, the FW versus FLA findings extended to functioning and risk of conversion to psychosis, with a reduced likelihood of choosing stimuli associated with FW versus FLA being associated with a higher probability of conversion to a psychotic disorder and lower functioning. No significant associations were observed between training and transfer phase performance and depression, and correlations with negative symptoms were of similar magnitude after controlling for depression (see Supplementary Table 4).

Analyses with CHR negative symptom subgroups

The results are briefly described below. See Supplementary Material, including Supplementary Table 5 and Supplementary Figure 2 for full results.

Training Phase. In the linear mixed effects model with high CHR negative symptoms, low CHR negative symptoms, and CN groups, a probability X valence X group interaction was observed. This was driven by group differences in gain learning on the 80% probability trials; the high negative symptom CHR group had poorer learning from gains than CN (p < .01).

The valence difference scores significantly differed between the three groups on the 80% pairs. The high negative symptom group learned worse from gains than losses compared to CN (p = .01).

Transfer Phase. There were significant group differences between choosing stimuli associated with FW v. FL (see Supplementary Figure 3). The high negative symptom CHR group showed the lowest preference for FW v. FL (versus CN; (p < .01)).

Discussion

Negative symptoms are one of the earliest and strongest markers of risk for developing a psychotic disorder among CHR youth (Addington & Heinssen, Reference Addington and Heinssen2012; Alderman et al., Reference Alderman, Addington, Bearden, Cannon, Cornblatt, McGlashan, Perkins, Seidman, Tsuang, Walker, Woods and Cadenhead2015; Demjaha et al., Reference Demjaha, Valmaggia, Stahl, Byrne and McGuire2012; Healey et al., Reference Healey, Penn, Perkins, Woods, Keefe and Addington2018). Therefore, the identification of mechanisms underlying negative symptoms in CHR youth could hasten the discovery of novel treatment targets for early intervention and prevention efforts. In PDs, explicit RL impairments have been identified as a reward-processing component that contributes to motivational negative symptoms (Strauss et al., Reference Strauss, Waltz and Gold2014), including impairments in rapid RL and learning from gains (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011). It is unclear whether these RL processes are also associated with negative symptoms in CHR.

Our results largely supported hypotheses. Specifically, CHR showed: (1) impaired learning from gains; (2) intact loss avoidance learning; (3) impaired learning from gains was most consistently linked to greater negative symptom severity, whether measured continuously or categorically. Thus, similar to what has been observed in PDs (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Strauss et al., Reference Strauss, Frank, Waltz, Kasanova, Herbener and Gold2011, Reference Strauss, Waltz and Gold2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011), negative symptoms were associated with a specific profile of RL impairment in CHR characterized by impaired learning from gains but not losses. However, in contrast to hypotheses and studies in PDs (Cicero et al., Reference Cicero, Martin, Becker and Kerns2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011), CHR youth had a largely intact overall learning composite compared to CN. Decrements in some but not all domains of neurocognition have been observed in CHR youth at a smaller effect size than PDs (Catalan et al., Reference Catalan, Salazar De Pablo, Aymerich, Damiani, Sordi, Radua, Oliver, McGuire, Giuliano, Stone and Fusar-Poli2021; Pratt et al., Reference Pratt, Luther, Kinney, Osborne, Corlett, Powers, Woods, Gold, Schiffman, Ellman, Strauss, Walker, Zinbarg, Waltz, Silverstein and Mittal2023). The lack of group differences in rapid learning may reflect another area of relatively preserved cognition in CHR. However, individual differences in functioning mirrored findings in the PD literature (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Reinen et al., Reference Reinen, Smith, Insel, Kribs, Shohamy, Wager and Jarskog2014; Waltz et al., Reference Waltz, Frank, Robinson and Gold2007, Reference Waltz, Frank, Wiecki and Gold2011), as the overall learning composite was associated with worse functioning in supplemental analyses. Thus, impaired rapid RL may not characterize the majority of the CHR population but does occur in the subgroup with poor functioning.

The transfer phase—where stimuli were presented in novel pairings—helped to elucidate the nature of the identified deficits. Specifically, CHR youth in the whole sample and those with elevated negative symptoms showed a lower preference than CN for stimuli associated with frequently winning over frequently losing. In addition, greater negative symptoms were associated with less frequently choosing the stimuli associated with frequently winning over those associated with frequent loss avoidance. Since the stimuli within this transfer pair are associated with positive prediction errors during acquisition (i.e., gain reward or avoid losing reward, respectively), the selection of the frequent winner over the frequent loss avoider suggests that a participant can adequately represent reward value. This pattern is similar to how individuals with PDs perform (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012). Our findings in conjunction with this past work in individuals with PDs, suggest these RL deficits begin prior to the onset of psychosis and are associated with negative symptoms transphasically (i.e., across illness phases). Specifically, negative symptoms are associated with difficulty representing reward value and using those representations to guide decision-making across phases of psychotic illness. These findings extend prior CHR studies examining value representation with other tasks (e.g., delay discounting), which also found associations with negative symptoms (Bartolomeo, Chapman, Raugh, & Strauss, Reference Bartolomeo, Chapman, Raugh and Strauss2021). Given evidence from past neuroimaging studies, these RL deficits may be driven by reduced activation of regions like the ventral striatum, anterior cingulate cortex, and ventromedial prefrontal cortex (Millman et al., Reference Millman, Gallagher, Demro, Schiffman, Reeves, Gold, Rakhshan Rouhakhtar, Fitzgerald, Andorko, Redman, Buchanan, Rowland and Waltz2020; Schmidt et al., Reference Schmidt, Antoniades, Allen, Egerton, Chaddock, Borgwardt, Fusar-Poli, Roiser, Howes and McGuire2017).

Supplemental results also indicated that no learning or transfer phase indices were linked to depression in CHR youth. Associations between task performance and negative symptoms, functioning, and cross-sectional risk of developing psychosis were also of similar magnitude after controlling for depression. This suggests our findings are not simply a result of depression or reflective of secondary processes that mask negative symptoms. The lack of an association with depression contrasts some prior CHR studies indicating aspects of reward processing (e.g., hedonic reactivity) are associated with greater depression severity (Strauss et al., Reference Strauss, Nuñez, Ahmed, Barchard, Granholm, Kirkpatrick, Gold and Allen2018; Wotruba et al., Reference Wotruba, Heekeren, Michels, Buechler, Simon, Theodoridou, Kollias, Rössler and Kaiser2014). Since secondary negative symptom processes, such as depression and anxiety, are highly prevalent in CHR (Fusar-Poli et al., Reference Fusar-Poli, Nelson, Valmaggia, Yung and McGuire2014), additional studies with large samples are needed to clarify the role of depression in the observed associations.

A final important observation was that RL deficits were associated with a greater cross-sectional risk for developing a psychotic disorder and poorer functioning in secondary analyses; however, this may not be surprising since negative symptoms themselves are critically linked to both outcomes. RL deficits may serve as an intermediate phenotype that links neural functioning abnormalities (e.g., disrupted cortico-striatal connectivity) with negative symptoms and ultimately increased risk for conversion. Current efforts are underway through the CAPER consortium (Gold et al., Reference Gold, Corlett, Strauss, Schiffman, Ellman, Walker, Powers, Woods, Waltz, Silverstein and Mittal2020) to identify if performance on this task alone or in combination with other behavioral tasks assessing mechanisms of positive symptoms, disorganized symptoms, or other negative symptom mechanisms (e.g., effort-cost computation) may best predict conversion to psychosis in a longitudinal study. Identification of computerized tasks, such as the PRLT, that can provide a more objective marker of risk and require less training and clinical time than structured clinical interviews traditionally used to assess psychosis risk, may help to increase the scalability and accessibility of identifying CHR youth at greatest likelihood of converting.

Several limitations should be considered when interpreting these findings. First, the data reported is cross-sectional, and longitudinal data can more precisely identify the role of RL impairments in conversion to a psychotic disorder. Similarly, we cannot infer that RL deficits cause negative symptoms without a prospective longitudinal design that documents the initial emergence of negative symptoms and what predicts their onset/worsening over time. Second, the CHR sample was small and several of the associations were small to medium in magnitude; although the findings are bolstered by the fact that they largely align with prior PD studies using this paradigm (Barch et al., Reference Barch, Carter, Gold, Johnson, Kring, MacDonald, Pizzagalli, Ragland, Silverstein and Strauss2017; Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012) and other RL findings in CHR youth (Gold et al., Reference Gold, Waltz, Matveeva, Kasanova, Strauss, Herbener, Collins and Frank2012; Millman et al., Reference Millman, Gallagher, Demro, Schiffman, Reeves, Gold, Rakhshan Rouhakhtar, Fitzgerald, Andorko, Redman, Buchanan, Rowland and Waltz2020; Waltz et al., Reference Waltz, Demro, Schiffman, Thompson, Kline, Reeves, Xu and Gold2015), our findings should still be considered preliminary and replication in larger samples is needed to confirm the results. Future studies should also use a combination of functional neuroimaging and computational modeling approaches to confirm the presence of mechanisms that were inferred to be present based on behavioral performance alone. Further, although the BNSS version used was adapted for CHR youth (Strauss & Chapman, Reference Strauss and Chapman2018), recently developed negative symptom measures for CHR youth, including the Negative Symptom Inventory for Psychosis Risk (NSI-PR; Strauss et al., Reference Strauss, Walker, Pelletier-Baldelli, Carter, Ellman, Schiffman, Luther, James, Berglund, Gupta, Ristanovic and Mittal2023), may be more sensitive assessments in this population. Third, only a single reward processing task was administered. Given negative symptom models have posited that RL and other aspects of reward processing interact to produce motivational symptoms (Kring & Barch, Reference Kring and Barch2014; Strauss et al., Reference Strauss, Waltz and Gold2014), future studies could investigate whether the observed impairments are primary or secondary to other reward-processing processes in CHR (e.g., value representation, and hedonic deficits). Our findings are also limited to reinforcement learning in the context of monetary rewards, and additional work is needed to see if CHR youth have differential response patterns to monetary rewards versus other types of reward (social, food, etc), as not all reward types are encoded similarly (Simon et al., Reference Simon, Skunde, Wu, Schnell, Herpertz, Bendszus, Herzog and Friederich2015; Ulrich et al., Reference Ulrich, Rüger, Durner, Grön and Graf2023; Wang, Liu, & Shi, Reference Wang, Liu and Shi2020). Finally, given that neurocognitive functioning, especially working memory, has been shown to influence RL in adults with PDs (Barch et al., 2017; Collins et al., Reference Collins, Brown, Gold, Waltz and Frank2014 & Collins et al., Reference Collins, Albrecht, Waltz, Gold and Frank2017), future work is needed to more systematically examine how working memory capacity may interact with RL processes to influence negative symptoms in CHR youth.

With replication, there are several study implications. Most critically, these results suggest that impairments in positive reward learning and specifically effectively using value representations to guide decision-making may be a trans-phasic mechanism underlying negative symptoms in the PD-spectrum. It may therefore be beneficial to develop interventions targeting these mechanisms. For example, just-in-time mobile interventions (Nahum-Shani et al., Reference Nahum-Shani, Smith, Spring, Collins, Witkiewitz, Tewari and Murphy2018) could be used to promote engagement in goal-directed activities and actively encode and consolidate reward-related outcomes that can serve as future motivation. Indeed, mobile interventions using similar techniques to target value representation in SZ have shown initial efficacy in improving motivation in SZ (Luther et al., Reference Luther, Fischer, Johnson-Kwochka, Minor, Holden, Lapish, McCormick and Salyers2020). These interventions could also be bolstered through in-person therapies focused on using such techniques as behavioral experiments and Socratic questioning to support learning associations between engagement in novel activities and rewards. The PRLT may be an important endpoint in clinical trials to identify how and whether treatments are effectively engaging RL processes to improve negative symptoms and other aspects of psychosis symptomatology.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1017/S0033291724003416.

Funding statement

This research was supported by a NARSAD Young Investigator Grant to Dr. Strauss.

Competing interest

Dr. Gregory Strauss is one of the original developers of the Brief Negative Symptom Scale (BNSS) and receives royalties and consultation fees from Medavante-ProPhase LLC in connection with commercial use of the BNSS and other professional activities; these fees are donated to the Brain and Behavior Research Foundation. Dr. Strauss has received honoraria and travel support from Medavante-ProPhase LLC for training pharmaceutical company raters on the BNSS. In the past 2 years, Dr. Strauss has consulted for and/or been on the speaker bureau for Minerva Neurosciences, Acadia, Lundbeck, Sunovion, Boehringer Ingelheim, and Otsuka pharmaceutical companies. All other authors have no relevant disclosures to report.

Ethical standard

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

References

Addington, J., Farris, M., Devoe, D., & Metzak, P. (2020). Progression from being at-risk to psychosis: Next steps. Npj Schizophrenia, 6(1), 27. https://doi.org/10.1038/s41537-020-00117-0CrossRefGoogle ScholarPubMed
Addington, J., & Heinssen, R. (2012). Prediction and prevention of psychosis in youth at clinical high risk. Annual Review of Clinical Psychology, 8(1), 269–289. https://doi.org/10.1146/annurev-clinpsy-032511-143146CrossRefGoogle ScholarPubMed
Ahmed, A. O., Strauss, G. P., Buchanan, R. W., Kirkpatrick, B., & Carpenter, W. T. (2015). Are negative symptoms dimensional or categorical? Detection and validation of deficit schizophrenia with taxometric and latent variable mixture models. Schizophrenia Bulletin, 41(4), 879–891. https://doi.org/10.1093/schbul/sbu163CrossRefGoogle ScholarPubMed
AhnAllen, C. G., Liverant, G. I., Gregor, K. L., Kamholz, B. W., Levitt, J. J., Gulliver, S. B., Pizzagalli, D. A., Koneru, V. K., & Kaplan, G. B. (2012). The relationship between reward-based learning and nicotine dependence in smokers with schizophrenia. Psychiatry Research, 196(1), 9–14. https://doi.org/10.1016/j.psychres.2011.09.011CrossRefGoogle ScholarPubMed
Alderman, T., Addington, J., Bearden, C., Cannon, T. D., Cornblatt, B. A., McGlashan, T. H., Perkins, D. O., Seidman, L. J., Tsuang, M. T., Walker, E. F., Woods, S. W., & Cadenhead, K. S. (2015). Negative symptoms and impaired social functioning predict later psychosis in L atino youth at clinical high risk in the N orth A merican prodromal longitudinal studies consortium. Early Intervention in Psychiatry, 9(6), 467–475. https://doi.org/10.1111/eip.12128CrossRefGoogle Scholar
Barch, D. M., Carter, C. S., Gold, J. M., Johnson, S. L., Kring, A. M., MacDonald, A. W., Pizzagalli, D. A., Ragland, J. D., Silverstein, S. M., & Strauss, M. E. (2017). Explicit and implicit reinforcement learning across the psychosis spectrum. Journal of Abnormal Psychology, 126(5), 694–711. https://doi.org/10.1037/abn0000259CrossRefGoogle ScholarPubMed
Bartolomeo, L. A., Chapman, H. C., Raugh, I. M., & Strauss, G. P. (2021). Delay discounting in youth at clinical high-risk for psychosis and adults with schizophrenia. Psychological Medicine, 51(11), 1898–1905. https://doi.org/10.1017/S0033291720000677CrossRefGoogle ScholarPubMed
Cannon, T. D. (2005). Clinical and genetic high-risk strategies in understanding vulnerability to psychosis. Schizophrenia Research, 79(1), 35–44. https://doi.org/10.1016/j.schres.2005.06.014CrossRefGoogle ScholarPubMed
Cannon, T. D., Bearden, C. E., Hollister, J. M., Rosso, I. M., Sanchez, L. E., & Hadley, T. (2000). Childhood cognitive functioning in schizophrenia patients and their unaffected siblings: A prospective cohort study. Schizophrenia Bulletin, 26(2), 379–393. https://doi.org/10.1093/oxfordjournals.schbul.a033460CrossRefGoogle ScholarPubMed
Catalan, A., Salazar De Pablo, G., Aymerich, C., Damiani, S., Sordi, V., Radua, J., Oliver, D., McGuire, P., Giuliano, A. J., Stone, W. S., & Fusar-Poli, P. (2021). Neurocognitive functioning in individuals at clinical high risk for psychosis: A systematic review and meta-analysis. JAMA Psychiatry, 78(8), 859. https://doi.org/10.1001/jamapsychiatry.2021.1290CrossRefGoogle ScholarPubMed
Cheng, G. L. F., Tang, J. C. Y., Li, F. W. S., Lau, E. Y. Y., & Lee, T. M. C. (2012). Schizophrenia and risk-taking: Impaired reward but preserved punishment processing. Schizophrenia Research, 136(1–3), 122–127. https://doi.org/10.1016/j.schres.2012.01.002CrossRefGoogle ScholarPubMed
Cicero, D. C., Martin, E. A., Becker, T. M., & Kerns, J. G. (2014). Reinforcement learning deficits in people with schizophrenia persist after extended trials. Psychiatry Research, 220(3), 760–764. https://doi.org/10.1016/j.psychres.2014.08.013CrossRefGoogle ScholarPubMed
Cohen, M. X., & Frank, M. J. (2009). Neurocomputational models of basal ganglia function in learning, memory and choice. Behavioural Brain Research, 199(1), 141–156. https://doi.org/10.1016/j.bbr.2008.09.029CrossRefGoogle ScholarPubMed
Collins, A. G., Albrecht, M. A., Waltz, J. A., Gold, J. M., & Frank, M. J. (2017). Interactions among working memory, reinforcement learning, and effort in value-based choice: A new paradigm and selective deficits in schizophrenia. Biological Psychiatry, 82(6), 431–439.CrossRefGoogle Scholar
Collins, A. G., Brown, J. K., Gold, J. M., Waltz, J. A., & Frank, M. J. (2014). Working memory contributions to reinforcement learning impairments in schizophrenia. Journal of Neuroscience, 34(41), 13747–13756.CrossRefGoogle ScholarPubMed
Culbreth, A. J., Moran, E. K., Kandala, S., Westbrook, A., & Barch, D. M. (2020). Effort, avolition, and motivational experience in schizophrenia: Analysis of behavioral and neuroimaging data with relationships to daily motivational experience. Clinical Psychological Science, 8(3), 555–568. https://doi.org/10.1177/2167702620901558CrossRefGoogle ScholarPubMed
Demjaha, A., Valmaggia, L., Stahl, D., Byrne, M., & McGuire, P. (2012). Disorganization/cognitive and negative symptom dimensions in the at-risk mental state predict subsequent transition to psychosis. Schizophrenia Bulletin, 38(2), 351–359. https://doi.org/10.1093/schbul/sbq088CrossRefGoogle ScholarPubMed
Devoe, D. J., Braun, A., Seredynski, T., & Addington, J. (2020). Negative symptoms and functioning in youth at risk of psychosis: A systematic review and meta-analysis. Harvard Review of Psychiatry, 28(6), 341–355. https://doi.org/10.1097/HRP.0000000000000273CrossRefGoogle ScholarPubMed
Frank, M. J., & Claus, E. D. (2006). Anatomy of a decision: Striato-orbitofrontal interactions in reinforcement learning, decision making, and reversal. Psychological Review, 113(2), 300–326. https://doi.org/10.1037/0033-295X.113.2.300CrossRefGoogle ScholarPubMed
Frank, M. J., Loughry, B., & O’Reilly, R. C. (2001). Interactions between frontal cortex and basal ganglia in working memory: A computational model. Cognitive, Affective, & Behavioral Neuroscience, 1(2), 137–160. https://doi.org/10.3758/CABN.1.2.137CrossRefGoogle ScholarPubMed
Frank, M. J., Seeberger, L. C., & O’Reilly, R. C. (2004). By carrot or by stick: Cognitive reinforcement learning in parkinsonism. Science, 306(5703), 1940–1943. https://doi.org/10.1126/science.1102941CrossRefGoogle ScholarPubMed
Fusar-Poli, P., Borgwardt, S., Bechdolf, A., Addington, J., Riecher-Rössler, A., Schultze-Lutter, F., Keshavan, M., Wood, S., Ruhrmann, S., Seidman, L. J., Valmaggia, L., Cannon, T., Velthorst, E., De Haan, L., Cornblatt, B., Bonoldi, I., Birchwood, M., McGlashan, T., Carpenter, W., … Yung, A. (2013). The psychosis high-risk state: A comprehensive state-of-the-art review. JAMA Psychiatry, 70(1), 107. https://doi.org/10.1001/jamapsychiatry.2013.269CrossRefGoogle Scholar
Fusar-Poli, P., Nelson, B., Valmaggia, L., Yung, A. R., & McGuire, P. K. (2014). Comorbid depressive and anxiety disorders in 509 individuals with an at-risk mental state: Impact on psychopathology and transition to psychosis. Schizophrenia Bulletin, 40(1), 120–131. https://doi.org/10.1093/schbul/sbs136CrossRefGoogle ScholarPubMed
Fusar-Poli, P., Salazar De Pablo, G., Correll, C. U., Meyer-Lindenberg, A., Millan, M. J., Borgwardt, S., Galderisi, S., Bechdolf, A., Pfennig, A., Kessing, L. V., Van Amelsvoort, T., Nieman, D. H., Domschke, K., Krebs, M.-O., Koutsouleris, N., McGuire, P., Do, K. Q., & Arango, C. (2020). Prevention of psychosis: Advances in detection, prognosis, and intervention. JAMA Psychiatry, 77(7), 755. https://doi.org/10.1001/jamapsychiatry.2019.4779CrossRefGoogle ScholarPubMed
Gold, J. M., Waltz, J.A., Matveeva, T.M., Kasanova, Z., Strauss, G.P., Herbener, E.S., Collins, A.G.E., & Frank, M.J. (2012). Negative symptoms and the failure to represent the expected reward value of actions: Behavioral and computational modeling evidence. Archives of General Psychiatry, 69(2), 129. https://doi.org/10.1001/archgenpsychiatry.2011.1269CrossRefGoogle ScholarPubMed
Gold, J. M., Corlett, P. R., Strauss, G. P., Schiffman, J., Ellman, L. M., Walker, E. F., Powers, A. R., Woods, S. W., Waltz, J. A., Silverstein, S. M., & Mittal, V. A. (2020). Enhancing psychosis risk prediction through computational cognitive neuroscience. Schizophrenia Bulletin, 46(6), 1346–1352. https://doi.org/10.1093/schbul/sbaa091CrossRefGoogle ScholarPubMed
Gonzalez, C., Kramar, C., Garagoli, F., Rossato, J. I., Weisstaub, N., Cammarota, M., & Medina, J. H. (2013). Medial prefrontal cortex is a crucial node of a rapid learning system that retrieves recent and remote memories. Neurobiology of Learning and Memory, 103, 19–25. https://doi.org/10.1016/j.nlm.2013.04.006CrossRefGoogle ScholarPubMed
Hall, R. (1995). Global assessment of functioning: A modified scale. Psychosomatics, 36, 267–275. https://doi.org/10.1016/S0033-3182(95)71666-8CrossRefGoogle ScholarPubMed
Harvey, P. D., & Strassing, M. (2012). Predicting the severity of everyday functional disability in people with schizophrenia: Cognitive deficits, functional capacity, symptoms, and health status. World Psychiatry, 11(2), 73–79. https://doi.org/10.1016/j.wpsyc.2012.05.004CrossRefGoogle ScholarPubMed
Healey, K. M., Penn, D. L., Perkins, D., Woods, S. W., Keefe, R. S. E., & Addington, J. (2018). Latent profile analysis and conversion to psychosis: Characterizing subgroups to enhance risk prediction. Schizophrenia Bulletin, 44(2), 286–296. https://doi.org/10.1093/schbul/sbx080CrossRefGoogle ScholarPubMed
Heerey, E. A., Bell-Warren, K. R., & Gold, J. M. (2008). Decision-making impairments in the context of intact reward sensitivity in schizophrenia. Biological Psychiatry, 64(1), 62–69. https://doi.org/10.1016/j.biopsych.2008.02.015CrossRefGoogle ScholarPubMed
Hélie, S., Ell, S. W., & Ashby, F. G. (2015). Learning robust cortico-cortical associations with the basal ganglia: An integrative review. Cortex, 64, 123–135. https://doi.org/10.1016/j.cortex.2014.10.011CrossRefGoogle ScholarPubMed
Kadakia, A., Catillon, M., Fan, Q., Williams, G. R., Marden, J. R., Anderson, A., Kirson, N., & Dembek, C. (2022). The economic burden of schizophrenia in the United States. The Journal of Clinical Psychiatry, 83(6). https://doi.org/10.4088/JCP.22m14458CrossRefGoogle ScholarPubMed
Kirkpatrick, B., Strauss, G. P., Nguyen, L., Fischer, B. A., Daniel, D. G., Cienfuegos, A., & Marder, S. R. (2011). The brief negative symptom scale: Psychometric properties. Schizophrenia Bulletin, 37(2), 300–305.CrossRefGoogle ScholarPubMed
Koch, K., Schachtzabel, C., Wagner, G., Schikora, J., Schultz, C., Reichenbach, J. R., Sauer, H., & Schlösser, R. G. M. (2010). Altered activation in association with reward-related trial-and-error learning in patients with schizophrenia. NeuroImage, 50(1), 223–232. https://doi.org/10.1016/j.neuroimage.2009.12.031CrossRefGoogle ScholarPubMed
Kring, A. M., & Barch, D. M. (2014). The motivation and pleasure dimension of negative symptoms: Neural substrates and behavioral outputs. European Neuropsychopharmacology, 24(5), 725–736.CrossRefGoogle ScholarPubMed
Luther, L., Fischer, M. W., Johnson-Kwochka, A. V., Minor, K. S., Holden, R., Lapish, C. L., McCormick, B., & Salyers, M. P. (2020). Mobile enhancement of motivation in schizophrenia: A pilot randomized controlled trial of a personalized text message intervention for motivation deficits. Journal of Consulting and Clinical Psychology, 88(10), 923–936. https://doi.org/10.1037/ccp0000599CrossRefGoogle ScholarPubMed
Marder, S. R., & Umbricht, D. (2023). Negative symptoms in schizophrenia: Newly emerging measurements, pathways, and treatments. Schizophrenia Research, 258, 71–77. https://doi.org/10.1016/j.schres.2023.07.010CrossRefGoogle ScholarPubMed
Miller, T. J., McGlashan, T. H., Rosen, J. L., Cadenhead, K., Ventura, J., McFarlane, W., Perkins, D. O., Pearlson, G. D., & Woods, S. W. (2003). Prodromal assessment with the structured interview for prodromal syndromes and the scale of prodromal symptoms: Predictive validity, interrater reliability, and training to reliability. Schizophrenia Bulletin, 29(4), 703–715. https://doi.org/10.1093/oxfordjournals.schbul.a007040CrossRefGoogle ScholarPubMed
Millman, Z. B., Gallagher, K., Demro, C., Schiffman, J., Reeves, G. M., Gold, J. M., Rakhshan Rouhakhtar, P. J., Fitzgerald, J., Andorko, N. D., Redman, S., Buchanan, R. W., Rowland, L. M., & Waltz, J. A. (2020). Evidence of reward system dysfunction in youth at clinical high-risk for psychosis from two event-related fMRI paradigms. Schizophrenia Research, 226, 111–119. https://doi.org/10.1016/j.schres.2019.03.017CrossRefGoogle ScholarPubMed
Minichino, A., Francesconi, M., Carrión, R. E., Bevilacqua, A., Parisi, M., Rullo, S., Ando’, A., Biondi, M., Delle Chiaie, R., & Cadenhead, K. (2017). Prediction of functional outcome in young patients with a recent-onset psychiatric disorder: Beyond the traditional diagnostic classification system. Schizophrenia Research, 185, 114–121. https://doi.org/10.1016/j.schres.2016.12.019CrossRefGoogle Scholar
Nahum-Shani, I., Smith, S. N., Spring, B. J., Collins, L. M., Witkiewitz, K., Tewari, A., & Murphy, S. A. (2018). Just-in-Time Adaptive Interventions (JITAIs) in mobile health: Key components and design principles for ongoing health behavior support. Annals of Behavioral Medicine, 52(6), 446–462. https://doi.org/10.1007/s12160-016-9830-8CrossRefGoogle ScholarPubMed
Osborne, K. J., & Mittal, V. A. (2019). External validation and extension of the NAPLS-2 and SIPS-RC personalized risk calculators in an independent clinical high-risk sample. Psychiatry Research, 279, 9–14. https://doi.org/10.1016/j.psychres.2019.06.034CrossRefGoogle Scholar
Packard, M. G., & Knowlton, B. J. (2002). Learning and memory functions of the basal ganglia. Annual Review of Neuroscience, 25(1), 563–593. https://doi.org/10.1146/annurev.neuro.25.112701.142937CrossRefGoogle ScholarPubMed
Pelletier-Baldelli, A., Strauss, G. P., Visser, K. H., & Mittal, V. A. (2017). Initial development and preliminary psychometric properties of the Prodromal Inventory of Negative Symptoms (PINS). Schizophrenia Research, 189, 43–49. https://doi.org/10.1016/j.schres.2017.01.055CrossRefGoogle ScholarPubMed
Piskulic, D., Addington, J., Cadenhead, K. S., Cannon, T. D., Cornblatt, B. A., Heinssen, R., Perkins, D. O., Seidman, L. J., Tsuang, M. T., Walker, E. F., Woods, S. W., & McGlashan, T. H. (2012). Negative symptoms in individuals at clinical high risk of psychosis. Psychiatry Research, 196(2–3), 220–224. https://doi.org/10.1016/j.psychres.2012.02.018CrossRefGoogle ScholarPubMed
Pratt, D. N., Barch, D. M., Carter, C. S., Gold, J. M., Ragland, J. D., Silverstein, S. M., & MacDonald, A. W. (2021). Reliability and replicability of implicit and explicit reinforcement learning paradigms in people with psychotic disorders. Schizophrenia Bulletin, 47(3), 731–739. https://doi.org/10.1093/schbul/sbaa165CrossRefGoogle ScholarPubMed
Pratt, D. N., Luther, L., Kinney, K. S., Osborne, K. J., Corlett, P. R., Powers, A. R., Woods, S. W., Gold, J. M., Schiffman, J., Ellman, L. M., Strauss, G. P., Walker, E. F., Zinbarg, R., Waltz, J. A., Silverstein, S. M., & Mittal, V. A. (2023). Comparing a computerized digit symbol test to a pen-and-paper classic. Schizophrenia Bulletin Open, 4(1), sgad027. https://doi.org/10.1093/schizbullopen/sgad027CrossRefGoogle ScholarPubMed
Prettyman, G. E., Kable, J. W., Didier, P., Shankar, S., Satterthwaite, T. D., Davatzikos, C., Bilker, W. B., Elliott, M. A., Ruparel, K., & Wolf, D. H. (2021). Relationship of ventral striatum activation during effort discounting to clinical amotivation severity in schizophrenia. Npj Schizophrenia, 7(1), 48. https://doi.org/10.1038/s41537-021-00178-9CrossRefGoogle ScholarPubMed
Reinen, J. M., Van Snellenberg, J. X., Horga, G., Abi-Dargham, A., Daw, N. D., & Shohamy, D. (2016). Motivational context modulates prediction error response in schizophrenia. Schizophrenia Bulletin, 42(6), 1467–1475. https://doi.org/10.1093/schbul/sbw045CrossRefGoogle ScholarPubMed
Reinen, J., Smith, E. E., Insel, C., Kribs, R., Shohamy, D., Wager, T. D., & Jarskog, L. F. (2014). Patients with schizophrenia are impaired when learning in the context of pursuing rewards. Schizophrenia Research, 152(1), 309–310. https://doi.org/10.1016/j.schres.2013.11.012CrossRefGoogle ScholarPubMed
Reiss, J. P., Campbell, D. W., Leslie, W. D., Paulus, M. P., Ryner, L. N., Polimeni, J. O., Foot, B. J., & Sareen, J. (2006). Deficit in schizophrenia to recruit the striatum in implicit learning: A functional magnetic resonance imaging investigation. Schizophrenia Research, 87(1–3), 127–137. https://doi.org/10.1016/j.schres.2006.04.027CrossRefGoogle ScholarPubMed
Sarkar, S. (2015). Conceptualization and treatment of negative symptoms in schizophrenia. World Journal of Psychiatry, 5(4), 352. https://doi.org/10.5498/wjp.v5.i4.352CrossRefGoogle ScholarPubMed
Schlosser, D. A., Campellone, T. R., Biagianti, B., Delucchi, K. L., Gard, D. E., Fulford, D., Stuart, B. K., Fisher, M., Loewy, R. L., & Vinogradov, S. (2015). Modeling the role of negative symptoms in determining social functioning in individuals at clinical high risk of psychosis. Schizophrenia Research, 169(1–3), 204–208. https://doi.org/10.1016/j.schres.2015.10.036CrossRefGoogle ScholarPubMed
Schmidt, A., Antoniades, M., Allen, P., Egerton, A., Chaddock, C. A., Borgwardt, S., Fusar-Poli, P., Roiser, J. P., Howes, O., & McGuire, P. (2017). Longitudinal alterations in motivational salience processing in ultra-high-risk subjects for psychosis. Psychological Medicine, 47(2), 243–254. https://doi.org/10.1017/S0033291716002439CrossRefGoogle ScholarPubMed
Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journal of Neurophysiology, 80(1), 1–27. https://doi.org/10.1152/jn.1998.80.1.1CrossRefGoogle ScholarPubMed
Simon, J.J., Skunde, M., Wu, M., Schnell, K., Herpertz, S.C., Bendszus, M., Herzog, W., & Friederich, H.C. (2015). Neural dissociation of food- and money-related reward processing using an abstract incentive delay task. Social Cognitive and Affective Neuroscience, 10(8), 1113–20. https://doi.org/10.1093/scan/nsu162CrossRefGoogle ScholarPubMed
Spilka, M. J., Raugh, I. M., Berglund, A. M., Visser, K. F., & Strauss, G. P. (2023). Reinforcement learning profiles and negative symptoms across chronic and clinical high-risk phases of psychotic illness. European Archives of Psychiatry and Clinical Neuroscience, 273(8), 1747–1760. https://doi.org/10.1007/s00406-022-01528-zCrossRefGoogle ScholarPubMed
Strauss, G. P., & Chapman, H. C. (2018). Preliminary psychometric properties of the brief negative symptom scale in youth at clinical high-risk for psychosis. Schizophrenia Research, 193, 435–437. https://doi.org/10.1016/j.schres.2017.07.051CrossRefGoogle ScholarPubMed
Strauss, G. P., Datta, R., Armstrong, W., Raugh, I. M., Kraguljac, N. V., & Lahti, A. C. (2021). Reinforcement learning abnormalities in the attenuated psychosis syndrome and first episode psychosis. European Neuropsychopharmacology, 47, 11–19. https://doi.org/10.1016/j.euroneuro.2021.03.014CrossRefGoogle ScholarPubMed
Strauss, G. P., Frank, M. J., Waltz, J. A., Kasanova, Z., Herbener, E. S., & Gold, J. M. (2011). Deficits in positive reinforcement learning and uncertainty-driven exploration are associated with distinct aspects of negative symptoms in schizophrenia. Biological Psychiatry, 69(5), 424–431. https://doi.org/10.1016/j.biopsych.2010.10.015CrossRefGoogle ScholarPubMed
Strauss, G.P., Nuñez, A., Ahmed, A.O., Barchard, K.A., Granholm, E., Kirkpatrick, B., Gold, J.M., & Allen, D.N. (2018). The latent structure of negative symptoms in schizophrenia. JAMA Psychiatry, 75(12):1271–1279. https://doi.org/10.1001/jamapsychiatry.2018.2475CrossRefGoogle ScholarPubMed
Strauss, G. P., Ruiz, I., Visser, K. H., Crespo, L. P., & Dickinson, E. K. (2018). Diminished Hedonic response in neuroleptic-free youth at ultra high-risk for psychosis. Schizophrenia Research: Cognition, 12, 1–7. https://doi.org/10.1016/j.scog.2017.12.001Google ScholarPubMed
Strauss, G. P., Walker, E. F., Pelletier-Baldelli, A., Carter, N. T., Ellman, L. M., Schiffman, J., Luther, L., James, S. H., Berglund, A. M., Gupta, T., Ristanovic, I., & Mittal, V. A. (2023). Development and validation of the negative symptom inventory-psychosis risk. Schizophrenia Bulletin, 49(5), 1205–1216. https://doi.org/10.1093/schbul/sbad038CrossRefGoogle ScholarPubMed
Strauss, G. P., Waltz, J. A., & Gold, J. M. (2014). A review of reward processing and motivational impairment in schizophrenia. Schizophrenia Bulletin, 40(Suppl_2), S107–S 116.CrossRefGoogle ScholarPubMed
Ulrich, M., Rüger, A., Durner, V., Grön, G., & Graf, H. (2023). Reward is not reward: Differential impacts of primary and secondary rewards on expectation, outcome, and prediction error in the human brain’s reward processing regions. NeuroImage, 283, 120440. https://doi.org/10.1016/j.neuroimage.2023.120440CrossRefGoogle Scholar
Waltz, J. A., Demro, C., Schiffman, J., Thompson, E., Kline, E., Reeves, G., Xu, Z., & Gold, J. (2015). Reinforcement learning performance and risk for psychosis in youth. Journal of Nervous & Mental Disease, 203(12), 919–926. https://doi.org/10.1097/NMD.0000000000000420CrossRefGoogle ScholarPubMed
Waltz, J. A., Frank, M. J., Robinson, B. M., & Gold, J. M. (2007). Selective Reinforcement learning deficits in schizophrenia support predictions from computational models of striatal-cortical dysfunction. Biological Psychiatry, 62(7), 756–764. https://doi.org/10.1016/j.biopsych.2006.09.042CrossRefGoogle ScholarPubMed
Waltz, J. A., Frank, M. J., Wiecki, T. V., & Gold, J. M. (2011). Altered probabilistic learning and response biases in schizophrenia: Behavioral evidence and neurocomputational modeling. Neuropsychology, 25(1), 86–97. https://doi.org/10.1037/a0020882CrossRefGoogle ScholarPubMed
Waltz, J. A., & Gold, J. M. (2007). Probabilistic reversal learning impairments in schizophrenia: Further evidence of orbitofrontal dysfunction. Schizophrenia Research, 93(1–3), 296–303. https://doi.org/10.1016/j.schres.2007.03.010CrossRefGoogle ScholarPubMed
Waltz, J. A., & Gold, J. M. (2015). Motivational deficits in schizophrenia and the representation of expected value. In Simpson, E. H. & Balsam, P. D. (Eds.), Behavioral neuroscience of motivation (Vol. 27, pp. 375–410). Springer International Publishing. https://doi.org/10.1007/7854_2015_385CrossRefGoogle Scholar
Wang, D., Liu, T., & Shi, J. (2020). Neural dynamic responses of monetary and social reward processes in adolescents. Frontiers in Human Neuroscience, 21(14), 141. https://doi.org/10.3389/fnhum.2020.00141CrossRefGoogle Scholar
Weickert, T. W., Goldberg, T. E., Callicott, J. H., Chen, Q., Apud, J. A., Das, S., Zoltick, B. J., Egan, M. F., Meeter, M., Myers, C., Gluck, M. A., Weinberger, D. R., & Mattay, V. S. (2009). Neural correlates of probabilistic category learning in patients with schizophrenia. The Journal of Neuroscience, 29(4), 1244–1254. https://doi.org/10.1523/JNEUROSCI.4341-08.2009CrossRefGoogle ScholarPubMed
Wolf, D. H., Satterthwaite, T. D., Kantrowitz, J. J., Katchmar, N., Vandekar, L., Elliott, M. A., & Ruparel, K. (2014). Amotivation in schizophrenia: Integrated assessment with behavioral, clinical, and imaging measures. Schizophrenia Bulletin, 40(6), 1328–1337. https://doi.org/10.1093/schbul/sbu026CrossRefGoogle ScholarPubMed
Wotruba, D., Heekeren, K., Michels, L., Buechler, R., Simon, J. J., Theodoridou, A., Kollias, S., Rössler, W., & Kaiser, S. (2014). Symptom dimensions are associated with reward processing in unmedicated persons at risk for psychosis. Frontiers in Behavioral Neuroscience, 8. https://doi.org/10.3389/fnbeh.2014.00382CrossRefGoogle ScholarPubMed
Yilmaz, A., Şimşek, , & Gonul, A. S. (2012). Reduced reward-related probability learning in schizophrenia patients. Neuropsychiatric Disease and Treatment, 8, 27–34. https://doi.org/10.2147/NDT.S26243CrossRefGoogle ScholarPubMed
Yung, A. R., & McGorry, P. D. (1996). The prodromal phase of first-episode psychosis: Past and current conceptualizations. Schizophrenia Bulletin, 22(2), 353–370. https://doi.org/10.1093/schbul/22.2.353CrossRefGoogle ScholarPubMed
Zhang, T, Xu, L, Tang, Y, Li, H, Tang, X, Cui, H, Wei, Y, Wang, Y, Hu, Q, Liu, X, Li, C, Lu, Z, McCarley, RW, Seidman, LJ, Wang, J; SHARP (ShangHai At Risk for Psychosis) Study Group. (2018). Prediction of psychosis in prodrome: development and validation of a simple, personalized risk calculator. Psychological Medicine, 49(12):1990–1998. doi: 10.1017/S0033291718002738.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Sample characteristics

Figure 1

Table 2. PRLT in CHR youth and healthy controls

Figure 2

Figure 1. Training performance in healthy controls and clinical high-risk for psychosis samples. Mean accuracy is reported. Error bars denote the standard error of the mean (SEM).

Figure 3

Figure 2. Transfer phase performance across groups. * p < .05. FLA, frequent loss avoider; FW, frequent winner; FL, frequent loser; IW, infrequent winner. Mean accuracy is graphed. Error bars denote the standard error of the mean (SEM). *p < .05.

Figure 4

Table 3. Correlations between training and transfer performance and negative symptoms in the CHR group

Supplementary material: File

Luther et al. supplementary material

Luther et al. supplementary material
Download Luther et al. supplementary material(File)
File 1.3 MB