STUDY PROTOCOL published: 05 November 2020 doi: 10.3389/fpsyt.2020.536112 Optimizing Behavioral Paradigms to Facilitate Development of New Treatments for Anhedonia and Reward Processing Deficits in Schizophrenia and Major Depressive Disorder: Study Protocol Amy C. Bilderbeck 1*† , Andreea Raslescu 1† , Dennis Hernaus 2 , Anja Hayen 1 , Edited by: Daniel Umbricht 3 , Darrel Pemberton 4 , Jane Tiller 5 , Birgitte Søgaard 6 , Anke Sambeth 2 , Amitai Abramovitch, Therese van Amelsvoort 2 , Andreas Reif 7 , Georgios Papazisis 8 , Victor Pérez 9,10,11,12 , Texas State University, United States Matilde Elices 9,10,11,12 , Damien Maurice 13 , Valérie Bertaina-Anglade 13 , Gerard R. Dawson 1 Reviewed by: and Stephane Pollentier 14 Helen Pushkarskaya, 1 P1vital Ltd, Wallingford, United Kingdom, 2 School for Mental Health and Neuroscience, Maastricht University, Maastricht, Yale Medical School, United States Netherlands, 3 F. Hoffmann-La Roche Ltd, Basel, Switzerland, 4 Janssen Pharmaceutica, Beerse, Belgium, 5 BlackThorn Dean McKay, Therapeutics, San Francisco, CA, United States, 6 H. Lundbeck, Valby, Denmark, 7 Department of Psychiatry, Psychosomatic Fordham University, United States Medicine and Psychotherapy, University Hospital Frankfurt, Frankfurt, Germany, 8 Department of Clinical Pharmacology, *Correspondence: School of Medicine, Aristotle University of Thessaloniki, Thessaloniki, Greece, 9 Institut de Neuropsiquiatria i Addiccions, Parc Amy C. Bilderbeck de Salut Mar, Barcelona, Spain, 10 Hospital del Mar Medical Research Institute (IMIM), Barcelona, Spain, 11 Departament de abilderbeck@p1vital.com Psiquitria i Medicina Legal, Universitat Autònoma de Barcelona, Barcelona, Spain, 12 Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Madrid, Spain, 13 Biotrial Neuroscience, Rennes and Didenheim, France, 14 Boehringer † These authors have contributed Ingelheim International GmbH, Ingelheim, Germany equally to this work Specialty section: Background: Behavioral tasks focusing on different subdomains of reward processing This article was submitted to may provide more objective and quantifiable measures of anhedonia and impaired Psychopathology, a section of the journal motivation compared with clinical scales. Typically, single tasks are used in relatively Frontiers in Psychiatry small studies to compare cases and controls in one indication, but they are rarely Received: 04 March 2020 included in larger multisite trials. This is due to limited systematic standardization as Accepted: 22 September 2020 well as the challenges of deployment in international studies and stringent adherence Published: 05 November 2020 to the high regulatory requirements for data integrity. The Reward Task Optimization Citation: Bilderbeck AC, Raslescu A, Consortium (RTOC) was formed to facilitate operational implementation of reward Hernaus D, Hayen A, Umbricht D, processing tasks, making them suitable for use in future large-scale, international, Pemberton D, Tiller J, Søgaard B, Sambeth A, van Amelsvoort T, Reif A, multisite drug development studies across multiple indications. The RTOC clinical study Papazisis G, Pérez V, Elices M, aims to conduct initial optimization of a set of tasks in patients with major depressive Maurice D, Bertaina-Anglade V, disorder (MDD) or schizophrenia (SZ). Dawson GR and Pollentier S (2020) Optimizing Behavioral Paradigms to Methods: We will conduct a multicenter study across four EU countries. Participants Facilitate Development of New Treatments for Anhedonia and Reward (MDD = 37, SZ = 37, with ≤80 age- and gender-matched healthy volunteers) will Processing Deficits in Schizophrenia attend a study visit comprising screening, self-report and clinically rated assessments and Major Depressive Disorder: Study of anhedonia and symptom severity, and three reward processing tasks; specifically, Protocol. Front. Psychiatry 11:536112. the Grip Strength Effort task, the Doors task, and the Reinforcement Learning doi: 10.3389/fpsyt.2020.536112 Working Memory task. The Grip Strength Effort and Doors tasks include simultaneous Frontiers in Psychiatry | www.frontiersin.org 1 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers electroencephalography/event-related potential recordings. Outcomes will be compared using a two-way group design of MDD and SZ with matched controls, respectively. Further analyses will include anhedonia assessment scores as covariates. Planned analyses will assess whether our findings replicate previously published data, and multisite deployment will be evaluated through assessments of quality and conduct. A subset of participants will complete a second visit, to assess test–retest reliability of the task battery. Discussion: This study will evaluate the operational deployment of three reward processing tasks to the regulatory standards required for use in drug development trials. We will explore the potential of these tasks to differentiate patients from controls and to provide a quantitative marker of anhedonia and/or impaired motivation, establishing their usefulness as endpoints in multisite clinical trials. This study should demonstrate where multifaceted reward deficits are similar or divergent across patient populations. Registration: ClinicalTrials.gov (NCT04024371). Keywords: anhedonia, impaired motivation, negative symptoms, reward processing, reward deficits, schizophrenia, depression INTRODUCTION prediction error (the difference between the expected and actual outcome of an action), reinforcement learning, and willingness Anhedonia, a diminished ability to experience pleasure, is to exert effort to acquire rewards. While some tasks require a symptom observed trans-diagnostically among individuals simultaneous assessment of brain activity using methods such with central nervous system disorders (1). In clinical practice, as fMRI [e.g., in the Monetary Incentive Delay task (12)] or anhedonia is closely linked to impaired motivation, loss of EEG, others (including the Effort Expenditure for Rewards Task interest, apathy, and social withdrawal, which together predict [EEfRT; Treadway et al. (13)] can be used as standalone measures, poor functional outcomes for patients, especially at the social facilitating their use in larger clinical studies. level (2–5). The phenomenology of anhedonia and motivational Combining behavioral methods with neuroimaging offers an impairment appear similar in depression and schizophrenia opportunity to identify targetable neurobiological biomarkers (SZ), but likely arise through different underlying neural associated with different reward processing abnormalities. In mechanisms. In depression, reduced reward responsivity may drug development, reward processing tasks may be used to largely account for anhedonia and motivational impairment, evaluate early pharmacodynamic activity in smaller translational whereas in SZ these behaviors may originate from deficits in studies; however, such tasks are rarely included in large, multisite neural correlates of cognitive control for effort-based decision- clinical trials across different diagnoses and/or countries. It making (6). Despite the considerable prevalence of anhedonia is still debatable whether they are sufficiently standardized, and its related constructs in many psychiatric indications, few sensitive, and reliable for use in either go/no-go decisions or studies have evaluated these symptoms as a primary focus of as endpoints in late-stage drug development studies. While the drug development [e.g., see Lally et al. (7)], as their resolution reliability of these measures has been extensively studied, there was assumed to occur with successful treatment of the core has been no systematic validation of the behavioral, fMRI, and indication (8). EEG paradigms used to evaluate aspects of reward processing In an effort to improve drug development and discovery across the many tasks available (14). Reward processing tasks in this important area of unmet need, modern behavioral are typically implemented in single-site academic studies with neuroscience has sought to understand anhedonia and negative small sample sizes, and it is common for task parameters to symptoms as trans-diagnostic deficits in reward processing vary either subtly or substantially between research studies. These in the context of the Research Domain Criteria (RDoC) differences make comparisons of results or pooling of data framework (9–11). Brain imaging and electrophysiology methods across studies problematic. In addition, robust data on test– such as functional magnetic resonance imaging (fMRI) and retest reliability is scarce, and often specific to one task in the electroencephalography (EEG) have been used to develop context of a single indication, further limiting generalizability behavioral reward processing tests. These tests provide objective to other experimental settings. Standardized and systematically and quantifiable measures of aspects of anhedonia and impaired optimized tasks are required to leverage their full potential in motivation, which may represent drug-sensitive functional drug development. biomarkers for cross-disorder domains of dysfunction. These Despite the trans-diagnostic manifestation of reward tests include paradigms probing anticipation and consumption processing deficits, prior studies have typically focused on of positive stimuli, delay discounting, reward response bias, differences between one particular diagnostic group and healthy Frontiers in Psychiatry | www.frontiersin.org 2 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers controls, and rarely include multiple populations with different Reward Task Optimization Consortium (RTOC), supported and central nervous system disorders. Exploring transdiagnostic fostered by the European College of Neuropsychopharmacology samples should help to determine whether seemingly similar Experimental Medicine Framework, was formed to facilitate this deficits arise from disparate mechanisms in different disorders development process. (6). This type of approach, which has been successfully The first consortium study aims to conduct initial work implemented in early-phase drug development programs to toward the optimization of three experimental tasks, to treat anhedonia across diagnoses [e.g., the NIMH FAST-FAIL support operational deployment of these tasks as endpoints initiative, Krystal et al. (15)], could help to optimize reward in interventional clinical trials. The task battery comprises the processing paradigms for use in clinical trials. Furthermore, the Grip Strength Effort task (24), the Doors task (27), and the relationship between behavioral endpoints and the severity of Reinforcement Learning/Working Memory (RLWM) task (28), anhedonia and impaired motivation is poorly understood for evaluating reward responsiveness, reward valuation and reward any given indication. learning, respectively. These tasks were selected by consensus Few studies have evaluated the sensitivity of these tasks based on a comprehensive review of publications demonstrating to pharmacological interventions. Amphetamine-induced specificity for subdomains, maturity of the tasks for use in trials, improvements in EEfRT performance have been demonstrated accessibility, construct validity, and tolerability (14, 23, 24). in healthy individuals (16), and the 5-HT reuptake inhibitor Our study objectives comprise: escitalopram can increase motivation to exert effort for reward in 1. Develop software and hardware to implement the task the Grip Strength Effort task (17). In individuals with Parkinson’s battery in clinical trials compliant with regulatory standards, disease, dopamine depletion reduces motivation as measured by while evaluating and maximizing suitability for multisite the Grip Strength Effort task (18), and dopamine replacement implementation in different languages. can restore cognitive motivation (19). Other studies have shown 2. Define a standardized set of task parameters. that pharmacological manipulation of dopaminergic signaling 3. Confirm operational deployment in multisite studies by in healthy individuals can influence choice performance in ensuring that task outcome measures are broadly comparable instrumental learning tasks (20, 21). to previously published data. Currently available research rarely includes multiple tasks 4. Replicate previously reported differences in performance and assessing different subdomains of reward processing. Assessing neural activity of participants with MDD or SZ compared with different subdomains of reward processing in the same healthy controls, where available. participants would help to understand the contribution of the 5. Evaluate test–retest reliability for key task outcome measures different components and subdomains (6). Multiple paradigms in a subset of participants. Good test–retest reliability and are required to assess reward valuation, reward responsiveness utility as repeated measures are necessary characteristics and reward learning separately (14, 22), but these have been for tasks used as endpoints in clinical trials (29, 30), and largely studied in isolation rather than in a complementary would further confirm the suitability of our task battery for manner. Studies comparing multiple paradigms that address the deployment in clinical trials. same reward processing sub-construct are also valuable, but remain scarce. One such study (23, 24) compared five effort- In addition, we also plan to compare and contrast the pattern based decision-making paradigms related to reward valuation of reward processing deficits (assessed with the three proposed for use in clinical trials of individuals with SZ. Each paradigm tasks) observed in the MDD and SZ groups, and to assess was evaluated on its ability to differentiate patients and healthy the relationships between behavioral/EEG data and clinical controls, test–retest reliability, utility as a repeated measure interview/self-report measures of anhedonia. (practice effects), patient tolerability, and correlations with self- reported and clinically rated motivational disturbances. The five paradigms showed varying psychometric strengths and weaknesses, with the EEfRT (13) and Grip Strength Effort task METHODS (25, 26) deemed most promising for use in clinical trials subject to some modifications and refinements (24). Design However, these studies do not offer a complete trans- All participants will complete a study visit comprising screening, diagnostic picture of reward processing tasks and their administration of three reward processing tasks, and completion potential usefulness as endpoints in clinical trials. Rather than of a short set of clinical assessments and questionnaires repeating the same experiments across different positive valence measuring anhedonia/motivational deficits/negative symptoms. subdomains and indications, we aim to integrate findings from Initially, 8 healthy control participants will be recruited to available literature and optimize the most promising of these address any clearly evident implementation issues regarding tasks for implementation in future large-scale, multisite clinical hardware or software use, and to ensure that mean performance trials. The culmination of this process would be a first step on behavioral tasks and EEG read-outs are broadly in toward a neuroscience-informed reward processing test battery, alignment with expectations based on the published literature. similar to those used to assess cognitive impairments, which Subsequently, individuals with MDD or SZ will be recruited until could objectively quantify anhedonia and differentiate between 37 participants per group have completed the study. Further impaired subdomains in different psychiatric disorders. The healthy control participants will be recruited, for a maximum Frontiers in Psychiatry | www.frontiersin.org 3 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers TABLE 1 | Eligibility criteria. Group Inclusion criteria Exclusion criteria Retest visit exclusion criteria All participants Age of 20–55 years Diagnosis of schizoaffective disorder, bipolar disorder, Use of opioids, psychostimulants or eating disorder, obsessive-compulsive disorder, benzodiazepines within 2 weeks, or post-traumatic stress disorder, attention deficit diazepam within 4 weeks prior to hyperactivity disorder or autism spectrum disorder retest (healthy controls and individuals with MDD are also excluded if diagnosed with any psychotic disorder) History of substance use disorder according to DSM-5 In the opinion of the investigator, too criteria, except nicotine or caffeine, within 6 months prior ill to participate or likely to present a to screening danger to themselves or others Diagnosis of any clinically significant neurological disease (e.g., multiple sclerosis, epilepsy), mental retardation, or developmental disability Use of opioids, psychostimulants or benzodiazepines within 2 weeks, or diazepam within 4 weeks prior to screening In the opinion of the investigator, too ill to participate or likely to present a danger to themselves or others Individuals with MDD Primary DSM-5 diagnosis of MDD, confirmed by the result of the MINI conducted by the site at screening Current major depressive episode according to DSM-5 criteria, which has not lasted longer than 6 months If undergoing treatment, treatment with an antidepressant for at least 4 continuous weeks Individuals with SZ Primary DSM-5 diagnosis of SZ, Treatment with clozapine in the last 6 months before Treatment with clozapine between confirmed by the MINI at screening screening screening and retest Acute exacerbation requiring hospitalization within the last 3 months Score >3 (mild) in Clinical Global Impression of Parkinsonism as measured by ESRS-A at screening DSM-5, Diagnostic and Statistical Manual of Mental Disorders 5th edition; ESRS-A, Extrapyramidal Symptom Rating Scale–Abbreviated; MDD, major depressive disorder; MINI, Mini-International Neuropsychiatric Interview; SZ, schizophrenia. of 80 healthy control participants, to ensure acceptable age and patients and controls (10 individuals with SZ, 10 individuals with gender matching between healthy control and patient groups. MDD, and 20 healthy controls) and similar numbers of retest To investigate test–retest reliability, 16 participants from participants (4 individuals with SZ, 4 individuals with MDD, and each group will be recruited to attend a further retest visit 3– 4 healthy controls). Participants will be 20–55 years of age. 5 weeks after the initial visit. The retest visit will comprise Patients will partly be identified through clinical programs re-administration of the task battery as well as a subset of affiliated with the participating research sites, during clinical the clinical assessments of anhedonia and negative symptoms. assessments. Healthy control participants and individuals with Patients and control groups participating in the retest visit will MDD or SZ will also be recruited through social media and not be matched, but we will ensure each retest group is broadly advertising, and from cohort registers of patients who have representative of its own initial group in terms of age and gender. agreed to be informed about future studies. Participants will The study commenced in September 2019 and is expected be contacted using their preferred mode of communication. to run until the end of Q4 2020, accounting for 3 months of Eligibility of potential participants based on the major study downtime due to the COVID-19 pandemic and resulting Europe- criteria will be assessed via a short telephone screening. Eligible wide lockdowns. participants will receive the participant information sheet and will be offered adequate time to consider whether or not to Participants and Procedure participate. Where eligible participants indicate they would Subject recruitment and study visits will take place at four like to participate, the study team will help them plan their European research centers: Maastricht University, University study visit. Hospital Frankfurt, Institute of Neuropsychiatry and Addictions Eligibility criteria are listed in Table 1. At the retest visit, Parc de Salut Mar of Barcelona, and Aristotle University of participants will be asked about their recent medical history and Thessaloniki. Each center will aim to recruit similar numbers of medication use to assess their continued eligibility. Baseline data Frontiers in Psychiatry | www.frontiersin.org 4 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers for participants excluded from retest will be included in analyses and is reduced in individuals experiencing depressive symptoms as long as the eligibility criteria were met at baseline. (27, 37), suggesting good construct and discriminatory validity. At the initial study visit, participants who have provided The RLWM task (28) was selected to isolate the contribution informed consent will be screened by trained researchers using of working memory processes to reinforcement learning, which the MINI for axis I disorders (31), the Quick Inventory of are demonstrated by the effects of load and delay (39, 40), and Depressive Symptomatology – Self-Report (QIDS-SR-16) (32), thus enables a more specific assessment of reward learning in the Positive and Negative Symptom Scale (PANSS; individuals this study. with SZ only) (33), the Extrapyramidal Symptom Rating Scale– The present study aims to optimize these three tasks for Abbreviated (ESRS-A; individuals with SZ only) (34), and a computerized, in-clinic administration with minimal participant comprehensive questionnaire capturing demographics, relevant burden across research centers in four European countries and medical history, and current disease management. Eligible their respective languages, in agreement with the stringent data participants will complete the selected anhedonia scales and integrity and standardization requirements for use in drug the three reward processing tasks, in the following order: Grip development studies. Strength Effort task, Doors task, RLWM task. The Grip Strength We will employ several measures to ensure standardized Effort and Doors tasks will be administered in conjunction with operational deployment of the task battery. We will conduct EEG recording. The session, including EEG setup, instructions standardized face-to-face training of key staff and supply between individual tasks, and refreshment breaks, will last standardized manuals to help maintain consistent high standards approximately 165 min. At the end of the study visit, participants of data collection across all study sites, and hold regular all- will be reimbursed for their participation; participants will be site meetings to discuss operational performance. Test runs of informed that the amount of reimbursement will depend on task behavioral and EEG data collection will also be conducted prior performance, although in reality all participants will receive a to the study start, to identify potential issues and check the data fixed amount. Participants also attending the retest visit will be for abnormalities. The task battery will be delivered to all study reimbursed separately after completion of each visit. participants at the initial and retest visits across all sites using Participants who withdraw or fail to complete all study the P1vital R ePRO system (P1vital Ltd, UK), a readily scalable procedures will be replaced until 37 participants per group have online platform. A standardized script agreed by staff at all completed the initial visit. If, in the opinion of the investigator, sites and translated into the relevant languages (Dutch, Spanish, the EEG data obtained for a participant is not suitable for German, and Greek) will be used to provide instructions to the analysis, the participant will be replaced. participants, who will then complete standard practice trials for each task. However, while the key elements of the task delivery will be standardized, the exact presentation of the tasks may be Reward Processing Paradigms adapted to fit site-specific conventions, such as the way in which Task Selection monetary amounts are displayed (e.g., e0,40 or 0.40e). The three tasks selected for optimization in this study were chosen based on a review of the published literature, which The Grip Strength Effort Task demonstrates that these tasks are advanced in terms of The Grip Strength Effort task (Figure 1) measures participants’ methodological aspects and/or are suited to measure a specific willingness to exert effort for variable monetary rewards, a proxy RDoC domain/subdomain of reward processing. The evidence to measure for motivation. On each trial, participants choose to support the selection each of these tasks is discussed below. perform either an easy or hard grip. The hard grip requires the The Grip Strength Effort task measures reward valuation participant to squeeze a grip force response device at a high based on participants’ willingness to exert effort for variable levels percentage of their maximum possible grip strength (e.g., 90%), of reward (24, 35) with good reliability (intra-class correlation while the easy grip requires a lesser force (e.g., 50% of their coefficient: 0.63), tolerability, and patient/control discriminant maximum). The easy grip is always worth a specific lower amount validity (23, 24). Moreover, the Grip Strength Effort task has (e.g., e0.10), while the reward associated with the hard trial a straightforward experimental design that easily lends itself to varies (e.g., between three levels of reward value, e0.10, e0.20, concurrent measurement of brain activity using EEG, a novel or e0.40). The primary dependent variable is the percentage of combination that will be used in the present study to investigate hard grip choices at the different reward levels. the spatiotemporal neural dynamics of reward anticipation and During a trial, participants receive visual feedback of the force effort cost in the context of cost-benefit decision-making. produced via a thermometer scale marked with a target line, The Doors task is the RDoC paradigm of choice for measuring representing the target grip strength. If the participant reaches initial responsiveness to positive (gain) and negative (loss) the target grip strength, they receive the reward associated with outcomes, in conjunction with neurophysiological assessments their grip strength choice on that particular trial. The task is to measure mesocorticolimbic activity during reward processing preceded by a calibration step, which allows for the targets and (36–38). Reward-related responses to this task have often been visual feedback to be tailored to the participant’s performance. quantified in terms of feedback negativity, an event-related Participants complete a series of four practice trials prior to potential (ERP) index of reward evaluation (36). The feedback the assessment task. Two of these practice trials present a forced negativity ERP is positively correlated with behavioral and self- choice to allow the participant to experience both trial types; on report measures of the participant’s sensitivity to reward (38) the first the participant can only choose the easy task, and on Frontiers in Psychiatry | www.frontiersin.org 5 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers FIGURE 1 | The Grip Strength Effort task. In the Grip Strength Effort task, participants will complete 4 practice trials, 4 calibration trials, and 54 assessment trials, for a total task duration of approximately 20 min. EEG, electroencephalography; FRN, feedback-related negativity. the second they can only choose the hard task. The subsequent co-workers (28), which measures participants’ ability to two practice trials are identical to the main assessment, in that learn reward-based associations, and separately identifies the participants can freely choose either the easy or hard task. The contributions of reinforcement learning and working memory practice trials can be repeated as many times as needed to learn to performance. the task, and no performance criterion is required to proceed to Firstly, participants learn to select one of three actions (e.g., the main assessment. key presses) for each stimulus in a block using reward feedback We are not aware of any published studies combining the Grip (learning phase). Incorrect choices provide no reward (0 points), Strength Effort task with EEG recording, but the straightforward while correct choices lead to reward (either +1 or +2 points, task design readily lends itself to capturing brain activity in the probabilistically). The probability of obtaining 2 points vs. 1 point present study. is fixed for each stimulus and is drawn from a set of 0.2, 0.5, or 0.8 (reinforcement learning manipulation). Stimuli are presented The Doors Task individually, randomly intermixed, and the number of stimuli in The Doors task (Figure 2) measures initial responsiveness to a block varies from 2 to 5 (working memory manipulation). unpredictable reward receipt and loss and must be paired with Following the learning blocks, participants are presented brain imaging measures that can quantify this construct, such with an unexpected test of RL-based contributions to learning as fMRI or EEG. On each trial, participants choose between (transfer phase). On each test trial, participants choose which two identical stimuli (often two doors) presented on the left of two stimuli previously encountered in the learning blocks is and right side of the screen. Wins and losses are presented in more rewarding. No feedback is provided during this phase, so a pre-determined, pseudorandom order, such that left or right performance is entirely based on the participant having learned door choices do not influence results and participants achieve the probabilistic reward magnitude associations for all stimuli, exactly 50% wins associated with a positive reward (e.g., +e0.50) which might depend on the size of the block the stimulus was and 50% losses (e.g., −e0.25). The Doors task will be used drawn from. in conjunction with EEG recording in the current study, as Prior to the main assessment, participants will complete 43 described in previous studies. practice trials across two blocks related to the learning phase. Participants will complete four practice trials prior to the main In the first block comprising four trials, participants learn to assessment, in which the participant can choose freely between associate shapes with rewards without any time restrictions, the two doors, and feedback is presented as described above. The allowing the researcher to coach the participant through the practice trials can be repeated as many times as needed to learn completion of the task. In the second block, responses must the task, and no performance criterion is required to proceed to be submitted within a 3 s time limit, and a missed response is the main assessment. recorded if the time limit is exceeded. This time limit is also enforced in the main assessment. In the second half of this The RLWM Task block, participants must achieve 65% correct responses; if this The RLWM task (Figure 3) is a new experimental criterion is not met, the practice trials must be repeated. If protocol developed over multiple iterations by Collins and three unsuccessful attempts are made, the researcher is instructed Frontiers in Psychiatry | www.frontiersin.org 6 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers FIGURE 2 | The Doors task. In the Doors task, participants will complete 4 practice trials and 60 assessment trials, for a total task duration of approximately 15 min. EEG, electroencephalography; FRN, feedback-related negativity. FIGURE 3 | The Reinforcement Learning/Working Memory task. In the Reinforcement Learning/Working Memory task, participants will complete 43 practice trials, 507 learning trials, and 156 test trials, for a total task duration of approximately 40 min. not to proceed any further with the task. No practice trials are of administration, good construct validity and good reliability. provided for the transfer phase, to preserve the surprise element. All patient-reported scales will be delivered at the initial and retest visits using the P1vital R ePRO system. Self-Report and Clinical Assessments of Anhedonia and Related Constructs SHAPS Participants will complete several self-report and clinical The self-report, 14-item SHAPS was developed to measure assessments in addition to those included in screening. These hedonic capacity (41). Participants report levels of include two self-report measures of anhedonia/motivation, the consummatory anhedonia over the last few days on a 4- Snaith-Hamilton Pleasure Scale (SHAPS) (41) and the Behavioral point scale from strongly agree to strongly disagree. The SHAPS Inhibition System/Behavioral Approach System scale (BIS/BAS) will be administered to all study participants. (42), as well as the Brief Negative Symptom Scale (BNSS) (43). The SHAPS is commonly used to assess anhedonia in These measures were selected based on short length and/or ease depressed and suicidal populations (44–46), and has excellent Frontiers in Psychiatry | www.frontiersin.org 7 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers reliability (Cronbach’s alpha coefficient >0.9), satisfactory STATISTICAL ANALYSES construct validity, and particularly promising convergent validity (44, 47). The primary objective of the study is to evaluate operational deployment of the behavioral tasks by replication of published results, as determined by statistically significant separation of the BIS/BAS respective indication groups from the healthy controls on the The self-report BIS/BAS is comprised of 24 items in four primary task variable for each of the three tasks. subscales: BIS, BAS drive, BAS reward responsiveness, and The sample size was calculated using G∗ Power 3.1.9.2 (58) BAS fun-seeking. The BAS subscales are designed to measure and powered to detect a statistically significant main effect of motivation toward a desired goal or rewarding stimulus, whereas group on the percentage hard trials chosen in the Grip Strength the BIS subscale assesses the participant’s desire to avoid an Effort task, with a significance threshold of p < 0.05, power of unpleasant situation or outcome (42, 48). Each item is scored 0.95, Cohen’s f = 0.3516, and three measurements (low, medium, from 1 (strongly agree) to 4 (strongly disagree). The BIS/BAS will high levels of reward in the Grip Strength Effort task). Cohen’s be administered to all study participants. f was calculated based on the effect size estimate (partial eta2 ) The BIS/BAS has been used in studies of mood disorders reported by a previous study that employed a probabilistic reward including bipolar disorder and depression, and in SZ (49–52). learning task (59) to measure similar reward-related underlying The reliability of the scale has been found to be favorable in a large psychological constructs to the tasks employed in the current sample of individuals with anxiety and mood disorders (Raykov’s study. The required sample was calculated as N = 37 per group; rho: 0.88) (49). the study will therefore aim to recruit 40 participants per patient group, to accommodate a modest degree of drop-out or non- BNSS compliance. The BNSS is a researcher- or clinician-administered 13-item After the first 8 healthy control participants have completed scale, which assesses five domains of negative symptomatology their study visit, we will conduct an initial analysis to address any in SZ: alogia, anhedonia, avolition, asociality, and blunted evident implementation issues regarding hardware or software affect. Each item is rated from 0 (non-existent) to 6 (highest use. These analyses will comprise evaluations of: participants’ severity) (43). The BNSS is almost exclusively administered to understanding of the tasks, especially where there is a choice to participants with SZ and SZ-type disorders (53–55) and will only be made; biological plausibility of response times; task duration; be administered to individuals with SZ in the present study. and amount of reward earned. Where possible, we will also The BNSS has been found by two separate studies to assess whether primary behavioral and EEG task endpoints have excellent reliability, with Cronbach’s alpha coefficients are in alignment with the published literature. The 8 subjects >0.9 (43, 56), and also has good convergent and discriminant included in the initial analysis will not be included in the validity (57). main analyses. Primary and secondary task endpoints will be compared between patient groups and their age and gender-matched EEG Data Acquisition controls (two-way group design of MDD and matched controls To control for variability in EEG recording conditions, an EEG and also SZ and matched controls). No additional correction for kit, including EEG caps and a specific EEG acquisition system multiple comparisons is planned. Efficacy outcomes for each task (amplifier and dedicated laptop) will be provided by Biotrial will be described by age, gender, site and level of education. These Inc. to each clinical site. At each site, the same EEG kit will variables will also be adjusted for the case-control comparison, be used for the whole set of tests performed during this study. if possible. We will also control for the effects of medication EEG will be recorded on 64 channels, with a sampling frequency class, given our inclusion criteria permit the participation of of 2,048 Hz and resolution of 16 bits. Raw EEG data will be medicated patients with SZ and MDD. Prior to database lock and recorded with a 0.1–70 Hz filter. A digital narrow 50-Hz centered analysis, data on participant medication will be classed according notch filter will be applied to the data to reduce electromagnetic to compound and dose, in consultation with a study psychiatrist. noise due to main power supply. Self-calibration and impedance Typical classes for this sample are expected to include atypical measurement modules are integrated in the acquisition system. antipsychotics, sedatives, SSRIs and other mood stabilizers. A EEG signals will be collected using EEG caps with electrodes three-way group comparison between the SZ, MDD, and control located according to the International 10–20 Jasper system, and groups may also be performed. conductive gel will be used to ensure good contact between The study efficacy outcomes will be descriptively summarized electrodes and the scalp, and to reduce impedance. In order to by site to explore potential cultural differences that are known to improve the artifact rejection process, two electrooculography be associated with response. Sites will be considered surrogates to (EOG) channels will be recorded using electrocardiogram patch cultural differences since they represent differences in location, electrodes. Vertical and horizontal EOG are bipolar channels that language, and potentially in ethnicity. detect movements of the eyes and eyelids. Artifacts including Further analyses will include anhedonia/motivational scores, high frequency noise due to muscular activity, environmental as measured by the SHAPS, BNSS, and BIS/BAS, as covariates. noise (i.e., electric peak or wire oscillation), and eye or eyelid This approach will also allow investigation of the influence movements (i.e., blinks) will be rejected by visual inspection. of anhedonia (and subdomains) on task performance across Frontiers in Psychiatry | www.frontiersin.org 8 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers all participants, regardless of their indication. Furthermore, Feedback-locked FRN will be averaged separately for each interactions between clinical assessment scores and group level of reward value (e0.10, e0.20, and e0.40) and will be will indicate where effects of anhedonia may be specific to scored as the most negative peak occurring 200–400 ms after patient group(s). feedback onset. A 200 ms period prior to feedback onset will Test–retest reliability for the three reward processing tasks be used as baseline. The level of difficulty of the task (easy will be evaluated separately for each participant group using vs. difficult) will also be considered, to control for cases where intraclass correlations. Variability of selected task endpoint scores the same level of reward (e0.10) could correspond to either a for the same participant across the two time points will be low-effort/low-reward or high-effort/low-reward option. While compared with the total variation across all endpoint scores for the Grip Strength Effort task is well-established and has the all participants within each group. properties required for generating ERPs at the time of feedback, Participants with missing and/or unusable data for any of the no electrophysiologically coupled studies that could guide our three tasks will be excluded from that particular analysis. Where analyses have been conducted so far. However, based on existing self-report data is missing and/or incomplete, the participant’s fMRI studies, we assume a certain degree of similarity in behavioral data endpoints will still be included in the analysis electrophysiological responses between the Doors and the Grip where possible. Strength Effort tasks. Computational modeling approaches may also be applied to yield further insights. This approach has been exemplified by an RLWM Task exploratory re-analysis of a study assessing effortful goal-directed The RLWM task will be analyzed as described previously (28). behavior (60). This approach includes investigation of differences in overall percentage of correct responses in the learning phase and percentage of correct responses in each block size between the Individual Paradigm Analyses patient and healthy control groups. Doors Task In order to quantify the effects of working memory and Performance in the Doors task will be analyzed as described reinforcement learning on task performance, we will also conduct previously (61) in terms of modeling the reward feedback stage trial-by-trial logistic regression analyses on data from both the of the task. learning and transfer phases. For the learning phase, we will Stimulus-locked ERPs will be averaged separately for losses model individual participants’ choice of response on each trial as and gains, using the 200 ms period before stimulus onset as a function of the following predictor variables: set size (number baseline. The primary task endpoint will be the feedback-related of stimulus images in the block), iteration (how many times negativity (FRN) component of the ERP, defined as the activity the stimulus has been encountered), pcor (number of previous on losses minus the activity on gains occurring at the FCz correct choices for the current stimulus), and delay (number of electrode site approximately 300 ms following feedback delivery. trials since the last correct choice for the current trial’s stimulus). The difference score measures neural sensitivity to outcome We will then compare the effects of each of these predictor valence regardless of the source, and this scoring method has variables on learning across participant groups. been used in a number of prior studies (27, 62–69). Feedback- For the transfer phase, we will define the following locked FRN will be averaged separately for each feedback type characteristics for each stimulus: value (reward history; the (i.e., gain, loss) and will be scored as the most negative peak average of all feedback received for this image), set size, and occurring around 300 ms after feedback onset, using a 200 ms block (number of the block in which the stimulus image period prior to feedback onset as baseline. was encountered). Logistic regression modeling of individual The relative effect sizes of other secondary biomarkers, participants’ choices will be used to quantify the effect of these comprising reward positivity (RewP), frontal alpha (8–13 Hz) characteristics on performance in the transfer phase, and these asymmetry, spectral analysis, and P300 will also be compared. effects will be compared between groups. RewP will be quantified as the mean activity in a specified time period after gain feedback onset at FCz. The P300 peak will be defined by its delay following feedback onset, amplitude (µV), DISCUSSION latency (msec) and area under the curve (µV∗ msec). Spectral analysis of absolute and/or relative amplitude will be performed This pilot study will evaluate and optimize the operational in delta, theta, alpha, beta and gamma bands, as defined by the deployment of three reward processing tasks for use in future guidelines for the recording and evaluation of pharmaco-EEG large-scale, international clinical trials. While these tasks are data in human participants (70). extensively used in neuropsychiatric research, they have not been developed to meet the regulatory standards required for Grip Strength Effort Task use in pharmaceutical industry-sponsored trials, and the task The primary endpoint for this task will be the percentage of parameters can vary substantially between studies. By replicating hard grip choices at each reward level (e0.10, e0.20, and e0.40). previous findings in two patient populations and healthy controls FRN will be the main reward feedback EEG biomarker of in a tightly controlled experimental setting, we will provide interest, and RewP and frontal asymmetry be investigated in a initial evidence of the trans-diagnostic reliability and test–retest secondary analysis. consistency of our chosen tasks to support their use in future Frontiers in Psychiatry | www.frontiersin.org 9 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers validation studies. This trial began recruiting in November 2019, restrictions on medication use among participants, although all and completion is anticipated in Q4 2020. participants receiving treatment are required to be on a stable The present study will also add to the growing body of medication regimen, raising the possibility of varying treatment literature supporting the identification of reliable biomarkers for effects between participants receiving different medications. cross-disorder domains of dysfunction, and explore differences Future research exploring the sensitivity of these paradigms to in reward processing deficits, as measured by the test battery, in medication use is needed to support validation of their use in individuals with MDD and SZ compared with matched controls. interventional clinical trials; however, the present study may Previous research has demonstrated that both MDD and SZ are permit exploration of residual symptoms of anhedonia and its associated with decreased willingness to expend effort, blunted behavioral correlates in a cross-sectional sample of patients, striatal responses during reward anticipation, and reduced many of whom will be well-managed and receiving effective prediction error signaling to rewards relative to healthy controls medication. Finally, this study did not incorporate cognitive (71). Although it has been argued that the mechanisms of testing to identify any baseline differences in cognitive ability reward circuit dysregulation underlying anhedonia and impaired between cohorts, although the RLWM task will be used to assess motivation may differ between the two diagnoses (72), these how working memory impacts reward learning performance conclusions have often been based on literature reviews and across the patient and control groups. meta-analyses of studies using disparate methodologies. The In conclusion, the proposed study takes the first step toward interpretation of results across studies using conceptually similar optimizing a neuroscience-informed battery of experimental paradigms can be challenging due to differences in the specific tasks assessing different facets of reward processing for use in methodology, such as differing task reward values or the use future large clinical trials across multiple disease indications. of probabilistic vs. deterministic reward reinforcement (6). Few studies have employed the same tasks to investigate reward ETHICS STATEMENT processing in a trans-diagnostic manner [see Arrondo et al. (73) for an example of a study across multiple indications], This study has been granted ethical approval by a local Ethics and, to our knowledge, no previous study has employed a Review Board at each of the four study sites, including the battery of tasks targeting different reward processing constructs Medical Ethical Committee of Maastricht University, who will in multiple patient groups compared with healthy controls. act as study sponsor. The trial will be conducted in compliance As reward processing deficits are multifaceted rather than the with the principles of the 1996 Declaration of Helsinki, the result of a disturbance to a singular process, our study design Medical Research Involving Human Subjects Act (WMO), and might be better placed to provide initial evidence for where Good Clinical Practice and in accordance with all applicable deficits are similar and where they diverge across different regulatory requirements including, but not limited to, the patient populations. Research Governance Framework. Written informed consent will Notably, the current study is the result of a pre-competitive be obtained from all study participants. collaboration between academic researchers and representatives from the pharmaceutical industry. This approach to achieving AUTHOR CONTRIBUTIONS common aims and objectives has become increasingly popular, due in part to the economy of scale that pre-competitive GRD and SP conceived of the research study and provided consortia can provide (74). Moreover, bringing together industry key input for the devising the protocol, the writing of which experts who have extensive experience with experimental was coordinated by AB. ARa prepared the first draft of the medicine models in psychiatry and academics at the forefront manuscript. DH, AH, DU, DP, JT, AS, TA, ARe, GP, VP, ME, of neuropsychiatric research ensures that the projects conducted DM, VB-A, and BS contributed to the protocol design and review within the pre-competitive consortium framework remain of the final manuscript. DM and VB-A provided expertise for relevant to drug development while also exploring innovative EEG data collection and analysis. DH provided expertise for methodologies to fulfill ambitious objectives. As such, the pre- the Reinforcement Learning/Working Memory task design and competitive consortium platform established for this study analysis. All authors contributed to the article and approved the provides a promising framework for the future development, submitted version. optimization, and operationalization of the selected reward processing tasks and additional anhedonia biomarkers. ACKNOWLEDGMENTS Certain limitations in the design of this study should be noted, which present research questions to be addressed in future ME holds a Juan de la Cierva research contract awarded studies. Potential order effects due to the consistent task order in by El Instituto de Salud Carlos III (ISCIII) (FJCI-2017- all patients (Grip Strength Effort task, Doors task, RLWM task) 31738). Editorial support in the form of copyediting could vary across clinical populations. However, the consistent of the first draft, collation and incorporation of author task order in the present study may be appropriate given feedback to develop subsequent drafts, assembling our focus on exploring operational feasibility for interventional tables and figures, and referencing was provided by studies, where assessments typically follow a fixed task schedule Mark Condon, DPhil, of Fishawack Communications to avoid confounding of the primary comparison between Ltd, UK, and was funded by Boehringer Ingelheim intervention and control. The present study also places limited International GmbH. Frontiers in Psychiatry | www.frontiersin.org 10 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers REFERENCES 19. Mcguigan S, Zhou SH, Brosnan MB, Thyagarajan D, Bellgrove MA, Chong TT. Dopamine restores cognitive motivation in Parkinson’s disease. Brain. 1. Rømer Thomsen K, Whybrow PC, Kringelbach ML. Reconceptualizing (2019) 142:719–32. doi: 10.1093/brain/awy341 anhedonia: novel perspectives on balancing the pleasure networks 20. Pessiglione M, Seymour B, Flandin G, Dolan RJ, Frith CD. Dopamine- in the human brain. Front Behav Neurosci. (2015) 9:49–49. dependent prediction errors underpin reward-seeking behaviour in humans. doi: 10.3389/fnbeh.2015.00049 Nature. (2006) 442:1042–5. doi: 10.1038/nature05051 2. Mcmakin DL, Olino TM, Porta G, Dietz LJ, Emslie G, Clarke G, 21. Guitart-Masip M, Economides M, Huys QJM, Frank MJ, Chowdhury R, Duzel et al. Anhedonia predicts poorer recovery among youth with selective E, et al. Differential, but not opponent, effects of L -DOPA and citalopram serotonin reuptake inhibitor treatment-resistant depression. J Am Acad on action learning with reward and punishment. Psychopharmacology. (2014) Child Adolesc Psychiatr. (2012) 51:404–11. doi: 10.1016/j.jaac.2012. 231:955–66. doi: 10.1007/s00213-013-3313-4 01.011 22. Dillon DG, Rosso IM, Pechtel P, Killgore WDS, Rauch SL, Pizzagalli DA. Peril 3. Kiwanuka JN, Strauss GP, Mcmahon RP, Gold JM. Psychological predictors and pleasure: an rdoc-inspired examination of threat responses and reward of functional outcome in people with schizophrenia. Schizophr Res. (2014) processing in anxiety and depression. Depression Anxiety. (2014) 31:233–49. 157:299–304. doi: 10.1016/j.schres.2014.04.030 doi: 10.1002/da.22202 4. Grove TB, Tso IF, Chun J, Mueller SA, Taylor SF, Ellingrod VL, et al. Negative 23. Horan WP, Reddy LF, Barch DM, Buchanan RW, Dunayevich E, Gold affect predicts social functioning across schizophrenia and bipolar disorder: JM, et al. Effort-based decision-making paradigms for clinical trials in findings from an integrated data analysis. Psychiatr Res. (2016) 243:198–206. schizophrenia: part 2—external validity and correlates. Schizophr. Bull. (2015) doi: 10.1016/j.psychres.2016.06.031 41:1055–65. doi: 10.1093/schbul/sbv090 5. Kupferberg A, Bicks L, Hasler G. Social functioning in major 24. Reddy LF, Horan WP, Barch DM, Buchanan RW, Dunayevich E, Gold depressive disorder. Neurosci Biobehav Rev. (2016) 69:313–32. JM, et al. Effort-based decision-making paradigms for clinical trials in doi: 10.1016/j.neubiorev.2016.07.002 schizophrenia: part 1—psychometric characteristics of 5 paradigms. Schizophr 6. Culbreth AJ, Moran EK, Barch DM. Effort-cost decision-making in psychosis Bull. (2015) 41:1045–54. doi: 10.1093/schbul/sbv089 and depression: could a similar behavioral deficit arise from disparate 25. Kurniawan IT, Seymour B, Talmi D, Yoshida W, Chater N, Dolan RJ. psychological and neural mechanisms? Psychol Med. (2018) 48:889–904. Choosing to make an effort: the role of striatum in signaling physical doi: 10.1017/S0033291717002525 effort of a chosen action. J Neurophysiol. (2010) 104:313–21. doi: 10.1152/jn. 7. Lally N, Nugent AC, Luckenbaugh DA, Ameli R, Roiser JP, Zarate CA. Anti- 00027.2010 anhedonic effect of ketamine and its neural correlates in treatment-resistant 26. Cléry-Melin M-L, Schmidt L, Lafargue G, Baup N, Fossati P, Pessiglione bipolar depression. Transl Psychiatr. (2014) 4:e469. doi: 10.1038/tp.20 M. Why don’t you try harder? an investigation of effort production in 14.105 major depression. PLoS ONE. (2011) 6:e23178. doi: 10.1371/journal.pone.00 8. Remington G, Foussias G, Fervaha G, Agid O, Takeuchi H, Lee J, et al. Treating 23178 negative symptoms in schizophrenia: an update. Curr Treat Options Psychiatr. 27. Foti D, Hajcak G. Depression and reduced sensitivity to non-rewards versus (2016) 3:133–50. doi: 10.1007/s40501-016-0075-8 rewards: Evidence from event-related potentials. Biol Psychol. (2009) 81:1–8. 9. Gorwood P. Neurobiological mechanisms of anhedonia. Dialogues Clin doi: 10.1016/j.biopsycho.2008.12.004 Neurosci. (2008) 10:291–9. 28. Collins AGE, Albrecht MA, Waltz JA, Gold JM, Frank MJ. Interactions among 10. Insel T, Cuthbert B, Garvey M, Heinssen R, Pine DS, Quinn K, et al. working memory, reinforcement learning, and effort in value-based choice: Research Domain Criteria (RDoC): toward a new classification framework a new paradigm and selective deficits in schizophrenia. Biol Psychiatr. (2017) for research on mental disorders. Am J Psychiatr. (2010) 167:748–51. 82:431–9. doi: 10.1016/j.biopsych.2017.05.017 doi: 10.1176/appi.ajp.2010.09091379 29. Green MF, Nuechterlein KH, Gold JM, Barch DM, Cohen J, Essock S, 11. Treadway MT, Zald DH. Parsing anhedonia: translational models of reward- et al. Approaching a consensus cognitive battery for clinical trials in processing deficits in psychopathology. Curr Direct Psychol Sci. (2013) 22:244– schizophrenia: the NIMH-MATRICS conference to select cognitive domains 9. doi: 10.1177/0963721412474460 and test criteria. Biol Psychiatr. (2004) 56:301–7. doi: 10.1016/j.biopsych.2004. 12. Knutson B, Westdorp A, Kaiser E, Hommer D. FMRI visualization of brain 06.023 activity during a monetary incentive delay task. Neuroimage. (2000) 12:20–7. 30. Carter CS, Barch DM, Buchanan RW, Bullmore E, Krystal JH, Cohen J, doi: 10.1006/nimg.2000.0593 et al. Identifying cognitive mechanisms targeted for treatment development in 13. Treadway MT, Buckholtz JW, Schwartzman AN, Lambert WE, Zald schizophrenia: an overview of the first meeting of the Cognitive neuroscience DH. Worth the ’EEfRT’? the effort expenditure for rewards task as an treatment research to improve cognition in schizophrenia initiative. Biol objective measure of motivation and anhedonia. PLoS ONE. (2009) 4:e6598. Psychiatr. (2008) 64:4–10. doi: 10.1016/j.biopsych.2008.03.020 doi: 10.1371/journal.pone.0006598 31. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller 14. Delgado MR, Glimcher PW, Hajcak G, Pizzagalli DA, Treadway MT, Yerys E, et al. The Mini-International Neuropsychiatric Interview (MINI): BE. Behavioral Assessment Methods for RDoC Constructs: Positive Valence The development and validation of a structured diagnostic psychiatric Systems Final Report. (2016). Available online at: https://www.nimh.nih.gov/ interview for DSM-IV and ICD-10. J Clin Psychiatr. (1998) 59:22–33. about/advisory-boards-and-groups/namhc/reports/behavioral-assessment- doi: 10.1037/t18597-000 methods-for-rdoc-constructs.shtml#section-ii-b (accessed February 2, 2020). 32. Rush AJ, Trivedi MH, Ibrahim HM, Carmody TJ, Arnow B, Klein DN, et al. 15. Krystal AD, Pizzagalli DA, Mathew SJ, Sanacora G, Keefe R, Song A, et al. The 16-Item quick inventory of depressive symptomatology (QIDS), clinician The first implementation of the NIMH FAST-FAIL approach to psychiatric rating (QIDS-C), and self-report (QIDS-SR): a psychometric evaluation in drug development. Nat Rev Drug Discov. (2018) 18:82. doi: 10.1038/nrd. patients with chronic major depression. Biol Psychiatr. (2003) 54:573–83. 2018.222 doi: 10.1016/S0006-3223(02)01866-8 16. Wardle MC, Treadway MT, Mayo LM, Zald DH, De Wit H. Amping up 33. Kay SR, Fiszbein A, Opler LA. The Positive and Negative Syndrome effort: effects of d-amphetamine on human effort-based decision-making. Scale (PANSS) for schizophrenia. Schizophr Bull. (1987) 13:261–76. J Neurosci. (2011) 31:16597–602. doi: 10.1523/JNEUROSCI.4387- doi: 10.1093/schbul/13.2.261 11.2011 34. Chouinard G, Margolese HC. Manual for the Extrapyramidal 17. Meyniel F, Goodwin GM, Deakin JW, Klinge C, Macfadyen C, Milligan H, Symptom Rating Scale (ESRS). Schizophr Res. (2005) 76:247–65. et al. A specific role for serotonin in overcoming effort cost. Elife. (2016) doi: 10.1016/j.schres.2005.02.013 5:e17282. doi: 10.7554/eLife.17282 35. Hartmann MN, Hager OM, Reimann AV, Chumbley JR, Kirschner M, Seifritz 18. Le Heron C, Plant O, Manohar S, Ang YS, Jackson M, Lennox G, et al. E, et al. Apathy but not diminished expression in schizophrenia is associated Distinct effects of apathy and dopamine on effort-based decision-making in with discounting of monetary rewards by physical effort. Schizophr Bull. Parkinson’s disease. Brain. (2018) 141:1455–69. doi: 10.1093/brain/awy110 (2015) 41:503–12. doi: 10.1093/schbul/sbu102 Frontiers in Psychiatry | www.frontiersin.org 11 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers 36. Carlson JM, Foti D, Mujica-Parodi LR, Harmon-Jones E, Hajcak G. Ventral 55. Eisenstein SA, Bogdan R, Chen L, Moerlein SM, Black KJ, Perlmutter JS, et al. striatal and medial prefrontal BOLD activation is correlated with reward- Preliminary evidence that negative symptom severity relates to multilocus related electrocortical activity: a combined ERP and fMRI study. NeuroImage. genetic profile for dopamine signaling capacity and D2 receptor binding (2011) 57:1608–16. doi: 10.1016/j.neuroimage.2011.05.037 in healthy controls and in schizophrenia. J Psychiatr Res. (2017) 86:9–17. 37. Bress JN, Smith E, Foti D, Klein DN, Hajcak G. Neural response to reward doi: 10.1016/j.jpsychires.2016.11.007 and depressive symptoms in late childhood to early adolescence. Biol Psychol. 56. Strauss GP, Gold JM. A psychometric comparison of the clinical assessment (2012) 89:156–62. doi: 10.1016/j.biopsycho.2011.10.004 interview for negative symptoms and the brief negative symptom scale. 38. Bress JN, Hajcak G. Self-report and behavioral measures of reward Schizophr Bull. (2016) 42:1384–94. doi: 10.1093/schbul/sbw046 sensitivity predict the feedback negativity. Psychophysiology. (2013) 50:610–6. 57. Strauss GP, Keller WR, Buchanan RW, Gold JM, Fischer BA, Mcmahon doi: 10.1111/psyp.12053 RP, et al. Next-generation negative symptom assessment for clinical trials: 39. Collins AGE, Frank MJ. How much of reinforcement learning is working validation of the Brief negative symptom scale. Schizophr Res. (2012) 142:88– memory, not reinforcement learning? A behavioral, computational, 92. doi: 10.1016/j.schres.2012.10.012 and neurogenetic analysis. Eur J Neurosci. (2012) 35:1024–35. 58. Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using doi: 10.1111/j.1460-9568.2011.07980.x G∗ power 3.1: Tests for correlation and regression analyses. Behav Res 40. Collins AGE, Brown JK, Gold JM, Waltz JA, Frank MJ. Working memory Methods. (2009) 41:1149–60. doi: 10.3758/BRM.41.4.1149 contributions to reinforcement learning impairments in schizophrenia. J 59. Pizzagalli DA, Iosifescu D, Hallett LA, Ratner KG, Fava M. Reduced Neurosci. (2014) 34:13747–56. doi: 10.1523/JNEUROSCI.0989-14.2014 hedonic capacity in major depressive disorder: evidence from a probabilistic 41. Snaith RP, Hamilton M, Morley S, Humayan A, Hargreaves D, Trigwell P. A reward task. J Psychiatr Res. (2008) 43:76–87. doi: 10.1016/j.jpsychires.2008. Scale for the assessment of hedonic tone the snaith–hamilton pleasure scale. 03.001 Br J Psychiatr. (1995) 167:99–103. doi: 10.1192/bjp.167.1.99 60. Cooper JA, Barch DM, Reddy LF, Horan WP, Green MF, Treadway MT. 42. Carver CS, White TL. Behavioral inhibition, behavioral activation, and Effortful goal-directed behavior in schizophrenia: computational subtypes affective responses to impending reward and punishment: the BIS/BAS and associations with cognition. J Abnorm Psychol. (2019) 128:710–22. scales. J Personal Soc Psychol. (1994) 67:319–33. doi: 10.1037/0022-3514. doi: 10.1037/abn0000443 67.2.319 61. Levinson AR, Speed BC, Infantolino ZP, Hajcak G. Reliability of the 43. Kirkpatrick B, Strauss GP, Nguyen L, Fischer BA, Daniel DG, Cienfuegos A, electrocortical response to gains and losses in the doors task. Psychophysiology. et al. The brief negative symptom scale: psychometric properties. Schizophr (2017) 54:601–7. doi: 10.1111/psyp.12813 Bull. (2011) 37:300–5. doi: 10.1093/schbul/sbq059 62. Nieuwenhuis S, Slagter HA, Von Geusau NJA, Heslenfeld DJ, Holroyd 44. Nakonezny PA, Carmody TJ, Morris DW, Kurian BT, Trivedi CB. Knowing good from bad: differential activation of human cortical MH. Psychometric evaluation of the Snaith-Hamilton pleasure areas by positive and negative outcomes. Eur J Neurosci. (2005) 21:3161–8. scale in adult outpatients with major depressive disorder. Int Clin doi: 10.1111/j.1460-9568.2005.04152.x Psychopharmacol. (2010) 25:328–33. doi: 10.1097/YIC.0b013e32833 63. Dunning JP, Hajcak G. Error-related negativities elicited by monetary eb5ee loss and cues that predict loss. NeuroReport. (2007) 18:1875–8. 45. Ballard ED, Wills K, Lally N, Richards EM, Luckenbaugh DA, Walls T, doi: 10.1097/WNR.0b013e3282f0d50b et al. Anhedonia as a clinical correlate of suicidal thoughts in clinical 64. Hajcak G, Moser JS, Holroyd CB, Simons RF. It’s worse than you thought: ketamine trials. J Affect Disord. (2017) 218:195–200. doi: 10.1016/j.jad.2017. The feedback negativity and violations of reward prediction in gambling 04.057 tasks. Psychophysiology. (2007) 44:905–12. doi: 10.1111/j.1469-8986.2007. 46. Toups M, Carmody T, Greer T, Rethorst C, Grannemann B, Trivedi MH. 00567.x Exercise is an effective treatment for positive valence symptoms in major 65. Christie GJ, Tata MS. Right frontal cortex generates reward-related depression. J Affect Disord. (2017) 209:188–94. doi: 10.1016/j.jad.2016.08.058 theta-band oscillatory activity. NeuroImage. (2009) 48:415–22. 47. Franken IHA, Rassin E, Muris P. The assessment of anhedonia in clinical and doi: 10.1016/j.neuroimage.2009.06.076 non-clinical populations: further validation of the Snaith–Hamilton Pleasure 66. Moser JS, Simons RF. The neural consequences of flip-flopping: The feedback- Scale (SHAPS). J Affect Disord. (2007) 99:83–9. doi: 10.1016/j.jad.2006. related negativity and salience of reward prediction. Psychophysiology. (2009) 08.020 46:313–20. doi: 10.1111/j.1469-8986.2008.00760.x 48. Rizvi SJ, Pizzagalli DA, Sproule BA, Kennedy SH. Assessing anhedonia in 67. Gu R, Huang Y-X, Luo Y-J. Anxiety and feedback negativity. Psychophysiology. depression: Potentials and pitfalls. Neurosci Biobehav Rev. (2016) 65:21–35. (2010) 47:961–7. doi: 10.1111/j.1469-8986.2010.00997.x doi: 10.1016/j.neubiorev.2016.03.004 68. Hämmerer D, Li S-C, Müller V, Lindenberger U. Life span differences 49. Campbell-Sills L, Liverant GI, Brown TA. Psychometric evaluation of in electrophysiological correlates of monitoring gains and losses during the behavioral inhibition/behavioral activation scales in a large sample of probabilistic reinforcement learning. J Cognit Neurosci. (2010) 23:579–92. outpatients with anxiety and mood disorders. Psychol Assess. (2004) 16:244– doi: 10.1162/jocn.2010.21475 54. doi: 10.1037/1040-3590.16.3.244 69. Foti D, Kotov R, Klein DN, Hajcak G. Abnormal neural sensitivity to monetary 50. Schlosser DA, Fisher M, Gard D, Fulford D, Loewy RL, Vinogradov gains versus losses among adolescents at risk for depression. J Abnorm Child S. Motivational deficits in individuals at-risk for psychosis and Psychol. (2011) 39:913–24. doi: 10.1007/s10802-011-9503-9 across the course of schizophrenia. Schizophr Res. (2014) 158:52–7. 70. Jobert M, Wilson FJ, Ruigt GSF, Brunovsky M, Prichep LS, Drinkenburg doi: 10.1016/j.schres.2014.06.024 WHIM. Guidelines for the recording and evaluation of pharmaco-EEG data in 51. Dempsey RC, Gooding PA, Jones SH. A prospective study of bipolar disorder man: the International Pharmaco-EEG Society (IPEG). Neuropsychobiology. vulnerability in relation to behavioural activation, behavioural inhibition and (2012) 66:201–20. doi: 10.1159/000343478 dysregulation of the behavioural activation system. Eur Psychiatr. (2017) 71. Whitton AE, Treadway MT, Pizzagalli DA. Reward processing 44:24–9. doi: 10.1016/j.eurpsy.2017.03.005 dysfunction in major depression, bipolar disorder and schizophrenia. 52. Wang X, Zhou X, Dai Q, Ji B, Feng Z. The role of motivation in cognitive Curr Opin Psychiatr. (2015) 28:7–12. doi: 10.1097/YCO.00000000000 reappraisal for depressed patients. Front Hum Neurosci. (2017) 11:516. 00122 doi: 10.3389/fnhum.2017.00516 72. Lambert C, Da Silva S, Ceniti AK, Rizvi SJ, Foussias G, Kennedy 53. Choi K-H, Jaekal E, Lee G-Y. Motivational and behavioral activation as an SH. Anhedonia in depression and schizophrenia: a transdiagnostic adjunct to psychiatric rehabilitation for mild to moderate negative symptoms challenge. CNS Neurosci Ther. (2018) 24:615–23. doi: 10.1111/cns. in individuals with schizophrenia: a proof-of-concept pilot study. Front 12854 Psychol. (2016) 7:1759. doi: 10.3389/fpsyg.2016.01759 73. Arrondo G, Segarra N, Metastasio A, Ziauddeen H, Spencer J, Reinders NR, 54. Chung YS, Barch DM. Frontal-striatum dysfunction during reward et al. Reduction in ventral striatal activity when anticipating a reward in processing: Relationships to amotivation in schizophrenia. J Abnormal depression and schizophrenia: a replicated cross-diagnostic finding. Front Psychol. (2016) 125:453–69. doi: 10.1037/abn0000137 Psychol. (2015) 6:1280–1280. doi: 10.3389/fpsyg.2015.01280 Frontiers in Psychiatry | www.frontiersin.org 12 November 2020 | Volume 11 | Article 536112 Bilderbeck et al. Optimizing Reward-Related Biomarkers 74. Dawson GR, Craig KJ, Dourish CT. Validation of experimental has been a consultant to or has received honoraria or grants from AstraZeneca, medicine methods in psychiatry: the P1vital approach and experience. Bristol-Myers Squibb, Janssen Cilag, Lundbeck, Otsuka, Servier, Medtronic, and Biochem Pharmacol. (2011) 81:1435–41. doi: 10.1016/j.bcp.2011. Exeltis. SP is a full-time employee of Boehringer Ingelheim. 03.013 The remaining authors declare that the research was conducted in the absence of Conflict of Interest: The authors declare that this study received funding from any commercial or financial relationships that could be construed as a potential Boehringer Ingelheim International GmbH, H. Lundbeck, Janssen Pharmaceutica, conflict of interest. BlackThorn Therapeutics, and F. Hoffmann-La Roche Ltd. All members of the consortium, including representatives from the funders, provided their expertise Copyright © 2020 Bilderbeck, Raslescu, Hernaus, Hayen, Umbricht, Pemberton, to ensure the optimal study design, contributed to preparation of the manuscript, Tiller, Søgaard, Sambeth, van Amelsvoort, Reif, Papazisis, Pérez, Elices, Maurice, and were involved in the decision to submit for publication. Bertaina-Anglade, Dawson and Pollentier. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, AB, ARa, AH, and GRD are employees of P1vital Ltd. DU is a full-time employee distribution or reproduction in other forums is permitted, provided the original of F. Hoffmann-La Roche. DP is a full-time employee of Janssen Pharmaceutica. author(s) and the copyright owner(s) are credited and that the original publication JT is a full-time employee of Blackthorn Therapeutics. BS is a full-time employee in this journal is cited, in accordance with accepted academic practice. No use, of Lundbeck. DM and VB-A are full-time employees of Biotrial Neuroscience. VP distribution or reproduction is permitted which does not comply with these terms. Frontiers in Psychiatry | www.frontiersin.org 13 November 2020 | Volume 11 | Article 536112