Negative symptoms in schizophrenia: Newly emerging measurements, pathways, and treatments

.


The role of negative symptoms
Schizophrenia is a serious psychiatric disorder that presents as a combination of positive symptoms, such as hallucinations and delusions; cognitive dysfunction, such as deficits in memory and executive function; and negative symptoms, such as avolition, alogia, and expressive deficits (Bowie and Harvey, 2006;Correll and Schooler, 2020). Patients with schizophrenia present with substantial heterogeneity in terms of clinical characteristics, treatment response, and long-term prognosis, and attempts at mapping this heterogeneity have been challenging (Alnaes et al., 2019;Huber, 1997;Malhotra, 2015). This likely manifests an underlying biological and genetic heterogeneity as functional imaging studies have suggested disruptions in circuitry between various regions expressed differentially in various patients, while structural studies have shown variations in white and grey matter volume across patients with schizophrenia versus healthy controls (Gopal et al., 2016;Honea et al., 2005;Olabi et al., 2011).
As the traditional subtypes of schizophrenia have little anchoring in neurobiology, they have been replaced by a dimensional approach to understanding variations among patients in terms of positive, negative, and cognitive symptoms and their underlying neural circuitry and neurobiology. Among these symptom domains, negative symptoms are a core feature of schizophrenia that are predominant, enduring and clinically relevant in up to 60 % of patients (Bobes et al., 2010). They are often the earliest symptoms to appear, and they may significantly diminish the ability of patients to function in daily life (Bobes et al., 2010;Galderisi et al., 2018;Sicras-Mainar et al., 2014;Villalta-Gil et al., 2006). Negative symptoms can be categorized as primary, which are intrinsic to the underlying pathophysiology of the disorder, and secondary, which are symptoms that occur due to comorbidities, adverse effects of treatment, or environmental factors. Detecting the underlying cause of negative symptoms can be particularly troublesome when patients are being treated with antipsychotics that can lead to secondary negative symptoms (e.g., dysphoria, diminished expression). When symptoms are secondary to antipsychotic treatment, it is plausible that decreases in dosing may improve these symptoms; however, if negative symptoms are primary to the underlying disorder, reducing the antipsychotic dose may instead be detrimental and should always be done with great caution (Galderisi et al., 2021a). Regardless of the root of the issue, patients with more prominent negative symptoms have worse functional outcomes, with observed correlations existing between negative symptoms and impaired occupational, household, and recreational functioning, as well as relationship difficulties (Foussias and Remington, 2010;Galderisi et al., 2018). Currently, no pharmacological therapies are approved specifically to treat negative symptoms, making this an unmet need (Fusar-Poli et al., 2015;Galderisi et al., 2021a;. Although negative symptoms are one of five core dimensions of the psychopathology underlying schizophrenia within the Diagnostic and Statistical Manual of Mental Disorders 5 (American Psychiatric Association, 2013), they can also be observed in schizophrenia-spectrum disorders, disorders with comorbid psychosis, paranoid personality disorder, and neurological disorders such as Parkinson's and Alzheimer's disease (Strauss and Cohen, 2017). Within these disorders, negative symptoms may differ in whether they are persistent vs transient or primary vs secondary (e.g., asociality in paranoid personality disorder due to mistrust of others) (Strauss and Cohen, 2017). Because identical or similar aspects of negative symptoms in these disorders are thought to be driven by comparable abnormalities in neurophysiology and circuitry, transdiagnostic investigations of negative symptoms may be useful in advancing our understanding of this symptom dimension.
In recent years, progress has occurred in the assessment, characterization, conceptualization, and insights into the biological underpinnings of negative symptoms. This article will review these aspects in more detail. To advance research on the treatment of negative symptoms, this article will focus on describing possible improvements for measurement tools, developing a better understanding of the neural circuitry underlying negative symptoms, and identifying promising targets that may facilitate the development of new treatments.

Clinical rating instruments for negative symptoms
Research into the clinical manifestations of negative symptoms has revealed at least 5 symptoms that include blunted affect, alogia, avolition, asociality, and anhedonia. Factor analyses demonstrate that these characteristics can be grouped into 2 subdimensions: diminished expression and avolition/apathy (Blanchard and Cohen, 2006;Correll and Schooler, 2020;Foussias et al., 2014;Kirkpatrick and Fischer, 2006;Kirkpatrick et al., 2011;Stiekema et al., 2016). An assessment of measuring negative symptoms in clinical trials should assess at least these 2 subdimensions and ideally these 5 symptoms.
The Positive and Negative Syndrome Scale (PANSS) was among the first to incorporate an assessment of negative symptoms (Kay et al., 1987). Although it contains a negative symptom subscale, subsequent factor analyses have demonstrated that the PANSS negative symptoms scale is not ideal to capture negative symptoms (Marder et al., 1997;van der Gaag et al., 2006a;van der Gaag et al., 2006b). These principle component analyses demonstrated that some symptoms subsumed under the negative symptoms subscale do not load onto the same factor with other negative symptoms; hence, the variance captured is lower. The factor structure most widely used in trials are the "Marder" factors (Marder et al., 1997). The Marder factor of negative symptoms, also referred to as negative symptoms factor, comprises 7 items of the PANSS, including blunted affect, emotional withdrawal, poor rapport, passive/ apathetic withdrawal, motor retardation, active social avoidance, and lack of spontaneity in conversation (Hopkins et al., 2017;Kay et al., 1987;Marder et al., 1997). This negative symptom factor and similarly derived factor solutions have been shown to capture negative symptoms more accurately. Importantly, the US Food and Drug Administration has accepted the Marder negative symptom factor as a valid endpoint for phase 3 trials of negative symptoms (US Food and Drug Administration, 2020;Younis et al., 2020). Although when the PANSS was conceived the differentiation of negative symptoms into dimensions of expressive deficits and apathy/anhedonia was not known, more recent analyses have demonstrated that the PANSS is able to capture these dimensions as well (Harvey et al., 2017;Umbricht et al., 2012).
Because cross-factor correlations have been observed with the PANSS, improvements in the severity of symptoms within one domain may mistakenly be observed due to actual improvements in another, correlated domain (Hopkins et al., 2018). A recent statistical improvement to enhance the ability of the PANSS to capture independent symptom dimensions of schizophrenia is the uncorrelated PANSS score matrix (UPSM) (Hopkins et al., 2018). The UPSM was developed to remove the intercorrelations within PANSS factors. Briefly, by adjusting the weight of each factor using pooled randomized controlled trial data, a score matrix of coefficients for each factor was created; this was then used to generate transformed PANSS factors that have minimal intercorrelations (Hopkins et al., 2018). If the PANSS is used in trials of novel treatments for negative symptoms, factor solutions like the negative symptom factor or the UPSM are preferred as endpoints.
Although conceived around the same time, the Scale for the Assessment of Negative Symptoms (SANS) has not been used in clinical trials as much as the PANSS (Andreasen, 1989). This may relate to the challenges it poses in its administration and to the fact that some clinical symptoms it assesses were later found to not be part of negative symptoms (Correll and Schooler, 2020). In addition, the PANSS has been preferred as it provides a comprehensive and reliable assessment of psychopathology, which is ideal for clinical trials (Kumari et al., 2017).
Newer scales, such as the Brief Negative Symptom Scale (BNSS) and the Clinical Assessment Interview of Negative Symptoms (CAINS), were developed to address the limitations that existing instruments have in accurately measuring anhedonia and expressive deficits, and they were designed to be used in both inpatient and outpatient settings (Kirkpatrick et al., 2011;Kring et al., 2013). The BNSS and CAINS attempted to incorporate insights on the finer distinctions of anhedonia, such as the differences between anticipatory and consummatory anhedonia (Gard et al., 2006;Strauss and Gold, 2016). This concept is based on evidence that patients with schizophrenia do not show deficits in the enjoyment of actually pleasurable stimuli (consummatory hedonia) but are deficient in imagining the pleasure of events in the future (anticipatory hedonia) (Strauss and Gold, 2016). Not surprisingly, it is challenging to get reliable assessment of these concepts in patients who also show deficits in recollection (Strauss and Gold, 2016). Nonetheless, these newer scales have been implemented in phase 2 trials in negative symptoms of schizophrenia and are recommended by the European Psychiatric Association to be used as a complement to first-generation scales (Galderisi et al., 2021b). Whether they are superior in signal detection than the PANSS Marder factor, however, remains to be proven.
Although the newer rating scales are a welcome addition to the assessment instruments available to clinicians and for clinical trials, they suffer nonetheless from the well-known limitations of any rating scale, namely, incorrect and biased recollection of symptoms by the patients, a desire to provide socially acceptable answers, expectation bias by patients and clinicians, as well as disordered clinician interviews (Chang et al., 2019;Jongs et al., 2022;Savander et al., 2021). Novel measures that do not rely on patient reporting and subjective assessments by clinicians may be able to address these deficiencies. Indeed, data collection such as digital phenotyping, a moment-by-moment quantification of individual-level data from digital devices, such as smartphones, has the potential to collect both active and passive data from patients that are both clinically relevant and actionable (Torous et al., 2016). Items such as GPS and accelerometers that assess daily movement and adverse effects from treatment (e.g., tremors), cellular devices that track trends in patient's communication frequency, and digital surveys that assess patient well-being can collect highly valuable, easily attainable pieces of data that can provide further insight into an individual patient's symptoms (Henson et al., 2020;Torous et al., 2016). In a recent proof-ofmechanism study, measure of gestures derived from actigraphy correlated with measures of anhedonia in the BNSS, indicating a potential for use of activity-and gesture-based digital measures as outcomes for clinical trials in patients with schizophrenia (Umbricht et al., 2020a). A correlation heat map of how active (e.g., surveys) and passive (e.g., mobility/sociability via GPS and texts) data correspond with features of schizophrenia vs that of healthy controls indicates that significant associations can be detected in numerous constructs, including negative symptoms, using digital phenotyping via smartphones (Henson et al., 2020).
While these measures hold great promise in collecting objective data about the behavior of patients, they have not been shown to fully capture the nuances of negative symptoms. This may be because they are still evaluated against traditionally assessed negative symptoms and because of a lack of large-scale studies that use analytical approaches appropriate for revealing potentially novel dimensions of negative symptoms. Nonetheless, it is recommended that these methods be incorporated into studies of negative symptoms in order to collect data that would allow for a critical analysis of their ability to capture clinically relevant changes and potentially provide new insights into the structure of negative symptoms.

Neural circuitry of negative symptoms
Research into the underlying neurobiology and neural circuitry of negative symptoms has revealed that deficits in positive reward learning play a critical role in negative symptoms and that a distributed network mediates negative symptoms, with the ventral striatum as a central component. Recent data have demonstrated that among the 2 distinct dimensions of negative symptoms, i.e., anhedonia-amotivation and expressive deficits, the former drives functional impairments (Bègue et al., 2020;Hu et al., 2022). Accordingly, targeting the mechanisms underlying anhedonia-amotivation may lead to significant improvements in functional outcomes. Patients with a high degree of negative symptoms have been shown to exhibit impaired positive reward learning, while also having intact loss avoidance learning (Gold et al., 2012). Instead of learning through predicting positive reward, these patients learn to avoid punishments via prediction errors, effectively learning through negative rather than positive reward (Abohamza et al., 2020;Gold et al., 2012). In everyday language this means, these patients learn what not to do to avoid a bad outcome, but do not learn what to do to achieve a good outcome. This deficit in learning from positive outcomes leads to a deficit in establishing internal representations of positive goals, which then results in a deficit in motivational "pull" in given situations. This conceptualization posits that the lack of positive reward learning leads to deficient representation of behavioral goals, resulting in decreased motivation and engagement with the world. Recent updates to this model also emphasize impaired computation of cost/benefit ratios in the context of available rewards and the effort used to obtain them (Bègue et al., 2020).
Functional neuroimaging studies support this hypothesis and have identified dysfunction in anticipatory/reward pathways as part of the neural mechanisms underlying motivational impairments. These studies have demonstrated that the activity of the ventral striatum in response to negative prediction errors is intact (Strauss et al., 2014), while positive prediction errors correlate with reduced neural response in the ventral striatum, insula, frontal cortex, amygdala, putamen, hippocampus, and cingulate (Mitra et al., 2016;Strauss et al., 2014). White matter volume loss is also seen in the cingulate, insular cortex, and temporal cortex of patients with severe negative symptoms (Mitra et al., 2016;Ohtani et al., 2014). Current evidence supports the view that positive reward learning is primarily mediated by the direct pathway in basal ganglia/subthalamic circuitry, with the D1 receptor playing a critical role, while the indirect pathway, mediated by D2-expressing neurons in the striatum, is more critical for negative reward learning (Macpherson and Hikida, 2019;Waltz et al., 2011). Using behavioral assays that can differentiate between these 2 types of reward learning, it should be possible to phenotype the particular deficit in a given patient and, thus, arrive at a more physiologically based characterization rather than a rating scale-based assessment of a given patient. This approach has not been utilized yet in any pharmacological phase 2 or phase 3 study; however, it has been used as a readout in a recent proof-ofmechanism study of a PDE10 inhibitor in schizophrenia, which previously had shown positive effects on reward functioning in healthy volunteers (Umbricht et al., 2021).
In addition, both the ventral striatum and orbital frontal cortex are thought to be involved in processing the expected value of potential outcomes, and it is hypothesized that the orbital frontal cortex may also modulate the reinforcement/learning processes in the basal ganglia (Hernaus et al., 2019). Thus, dysfunction in the reward system associated with negative symptoms, such as avolition and anhedonia, is most likely a mix of dysfunction within the cortical and basal ganglia components of this circuitry. It will be important to further elucidate the pharmacology of this circuitry to advance treatments of negative symptoms. The interested reader is referred to Bègue et al. who have published a detailed review on this topic, including a discussion of the current understanding of functional changes in neural circuitry associated with the pathophysiology of negative symptoms (Bègue et al., 2020).
While imaging approaches are critical to identify the neural circuity involved in negative symptoms, they cannot be deployed in large phase 2 and phase 3 trials. However, behavioral assessments that probe positive and negative reward learning, as well as effort-based decisionmaking can be implemented relatively easily in such trials (e.g., probabilistic reward learning tasks, effort-based decision-making tasks, gripstrength tasks). It is thus possible to characterize patients by the key deficits underlying their negative symptoms. It is plausible that despite having a similar clinical profile of negative symptoms, some patients may exhibit significant differences in response to pharmacological interventions. These differences may occur as a function of the constellation of deficits in reward learning and motivational factors that are present in a given patient. Consequently, by showing an effect in specific subgroups of patients, defined by their profile in behavioral assessments of reward functioning and motivation, it may be possible to enhance the chance of success for a given compound while also advancing precision psychiatry.

Novel treatments
An unmet need still exists for schizophrenia treatments with greater efficacy in negative symptoms, despite substantial efforts. Although dopamine receptor blockade offers efficacy in reducing the positive symptoms of schizophrenia, it is not effective in improving primary negative symptoms (Howes et al., 2015;Potkin et al., 2020).
Investigation of glutamate in the pathophysiology of schizophrenia was initially triggered by the psychotomimetic properties of N-methyl-Daspartate (NMDA) receptor antagonists (Javitt and Zukin, 1991;Javitt et al., 2012). Support for a role of glutamate in negative symptoms emerged from the finding that NMDA antagonists, like ketamine, elicit negative symptoms in healthy volunteers (Beck et al., 2020). Subsequently, genetic findings have supported the hypothesis that glutamatergic abnormalities play a larger role in schizophrenia . Further research suggested that changes in corticolimbic NMDA receptor-mediated neuronal transmission modulate downstream dopaminergic transmission via excitation of the striatal structures associated with positive psychosis symptoms Uno and Coyle, 2019). The two categories of glutamate-modulating medications for schizophrenia that have been explored are treatments aimed at augmenting NMDA signaling and treatments that reduce synaptic glutamate, which rises in response to glutamate hypofunction due to upstream deficits in NMDA signaling .
The current focus in developing treatments has been on augmenting NMDA-dependent neurotransmission. The first generation of compounds focused on the obligatory co-agonist glycine. As NMDA receptors require both glutamate and glycine to be activated, glycine reuptake inhibitors can enhance glutamatergic transmission by increasing glycine concentrations within the synapse (Umbricht et al., 2014). During the development of these compounds, most treatments showed some evidence of efficacy in phase 2 trials, but then either failed in phase 3 or were abandoned (Buchanan et al., 2007;Correll et al., 2017;Gomes and Grace, 2021;Heresco-Levy et al., 2004). Factors that may have contributed to the lack of success in these trials include the absence of any stratification by behavioral reward learning or genetic biomarkers (GRIN2A) and a lack of tight quality control (Heresco-Levy et al., 2004;Javitt et al., 2012;Umbricht et al., 2014;Umbricht et al., 2020b). In addition, the NMDA modulator memantine has been reported as a potential treatment against negative symptoms (Di Iorio et al., 2017); however, the only controlled trial found a small effect in patients treated with clozapine (Veerman et al., 2016).
Other compounds focusing on enhancing NMDA functioning are in development, including d-amino acid oxidase (DAAO) inhibitors (Pei et al., 2021); however, a phase 2 study of TAK-831 (luvadaxistat), a DAAO inhibitor did not meet its primary endpoint of improvement over placebo in change from baseline in negative symptoms (Neurocrine Biosciences, 2021). Interestingly, another glycine reuptake inhibitor, BI 425809, has demonstrated effects on cognitive deficits but not on negative symptoms, indicating that such compounds may also affect symptom domains other than negative symptoms (Fleischhacker et al., 2021). While negative symptoms and cognitive deficits are considered separate domains of the psychopathology of schizophrenia, they both contribute to real-world functioning, and improvement in one domain may have a positive effect on another (Foussias et al., 2014;Millan et al., 2014).
Novel targets for treatment also include both nicotinic acetylcholine receptors (nAChRs) and muscarinic acetylcholine receptors (mAChRs), as early studies showed that activation of these resulted in antipsychotic/pro-cognitive effects in animal models as well as improvements in positive and cognitive symptoms in patients with schizophrenia (Jones et al., 2012). Unfortunately, results with α7-nAChR and agonists have been inconclusive, with generally no significant effects on cognitive function or negative symptoms (Jin et al., 2017). While studies with mAChR agonists, such as xanomeline, have been slightly more successful in terms of showing significant PANSS total score improvements and trends toward improvements in negative symptoms, significant cholinergic side effects also occurred (Shekhar et al., 2008). A recent efficacy and safety study of xanomeline combined with trospium, a peripheral anticholinergic that was added to reduce cholinergic adverse effects, showed significant improvements in both PANSS negative symptom subscore and the PANSS Marder negative symptom score compared with placebo over 5 weeks of treatment (Brannan et al., 2021). Anticholinergic adverse events were higher in the active treatment group versus placebo; however, discontinuations were similar across groups. Of note, this was a study in hospitalized patients with an acute exacerbation of positive symptoms, and in this scenario, assessments of negative symptoms are highly problematic and confounded by the presence of positive symptoms. Hence, the validity of this finding is yet to be proven in longer and larger trials using an appropriate patient population. Recently, positive phase 3 study results using the xanomeline/trospium combination were released; thus, it may soon be possible to assess the effects of xanomeline on negative symptoms in more detail (Karuna Therapeutics, 2022).
Emraclidine (CVL-231) is a muscarinic M 4 -selective positive allosteric modulator in development for schizophrenia. A small phase 1b study in acutely exacerbated patients with schizophrenia demonstrated statistically significant and clinically meaningful antipsychotic effects in PANSS total, positive, and negative scores (Cerevel Therapeutics, 2021); however, this compound is in very early stages of development. Further research is necessary to fully determine the role of acetylcholine receptor activation in treating the negative symptoms of schizophrenia.
Ulotaront and ralmitaront are novel compounds that activate trace amine-associated receptor 1 (TAAR1). TAAR1 is a G protein-coupled receptor that is normally activated by endogenous trace amines and that interacts with dopamine and serotonin receptors (Borowsky et al., 2001;Bunzow et al., 2001;Gomes and Grace, 2021;Revel et al., 2013). TAAR1 receptors have been shown to be present in mammalian brain areas that have been implicated in schizophrenia (i.e., prefrontal cortex, striatum, amygdala, nucleus accumbens, ventral tegmental area, dorsal raphe) (Dedic et al., 2021;Gainetdinov et al., 2018;Revel et al., 2013). TAAR1 has also been shown to be activated by many endogenous neurotransmitters, including dopamine, serotonin, norepinephrine, and some of their metabolites (Dedic et al., 2021;Espinoza et al., 2011;Gainetdinov et al., 2018). Activation of TAAR1 is thought to modulate presynaptic dopamine synthesis and postsynaptic dopamine-dependent signaling (Espinoza et al., 2011;Lindemann et al., 2008). In addition, TAAR1 activation has been shown to counteract hypofunctionality of glutamate NMDA receptors and regulate glutamate transmission in the prefrontal cortex (Espinoza et al., 2015;Gainetdinov et al., 2018;Revel et al., 2011). Thus, TAAR1 appears to modulate key neurotransmitter systems implicated in various aspects of the pathophysiology of schizophrenia. Compounds targeting TAAR1 may thus offer a highly novel approach in the treatment of schizophrenia.
Preclinical studies indicated that ulotaront may be effective in treating negative symptoms and depression using rodent models of social withdrawal (subchronic phencyclidine [PCP] treatment) and forced swim immobility time (Dedic et al., 2019). Ulotaront was also able to reverse PCP-induced social interaction deficits in rats and modest improvements were observed in the forced swim test (Dedic et al., 2021;Dedic et al., 2019). These data suggested the potential benefits of ulotaront, and possibly ralmitaront, in treating negative symptoms of schizophrenia.
In a phase 2 clinical study, 4 weeks of treatment with ulotaront demonstrated significant improvements over placebo in PANSS total score in patients with acute schizophrenia (Koblan et al., 2020). Ulotaront also exhibited improvements across a range of schizophrenia symptoms, including those measured by PANSS positive, negative, and general psychopathology subscales the BNSS; and the UPSM, although the patients were not selected for negative symptoms (Koblan et al., 2020). A long-term extension of this study supported these results, and although efficacy on negative symptoms showed improvements from baseline, these data were examined using descriptive statistics only (Correll et al., 2021). The same caveats mentioned above regarding negative symptoms vis-à-vis xanomeline apply here as well. Future studies confirming the long-term efficacy and safety of ulotaront and, more specifically, further examining negative symptoms in appropriately selected patients will be of great interest. Ralmitaront (RO6889450) is a TAAR1 partial agonist that is currently in phase 2 development. It was being evaluated in two doubleblind, placebo-controlled, randomized trials: one in patients with an acute exacerbation of positive symptoms (NCT04512066) and one in stable patients with negative symptoms (NCT03669640). However, it was recently announced that the trial in patients with an acute exacerbation did not meet the primary endpoint of reduction in PANSS total scores versus placebo (Ghost, 2022). The trial in patients with negative symptoms is ongoing. More information regarding the utility of targeting TAAR1 in the treatment of negative symptoms should, thus, be available in the near future.
Pimavanserin, an inverse agonist at the 5-HT 2A receptor with less inverse agonist effects at the 5-HT 1A receptor, is currently approved for the treatment of psychosis in patients with Parkinson's disease (Acadia Pharmaceuticals, 2016;Cummings et al., 2014). A large 26-week phase 2 trial of pimavanserin as adjunctive treatment for negative symptoms in patients with schizophrenia who were on a stable antipsychotic regimen showed a significant, albeit small, effect (effect size = 0.211) on the primary outcome measure (NSA) but failed to differentiate from placebo on secondary measures, including functional assessments (Bugarski-Kirola et al., 2022). It is currently being evaluated in an identically designed phase 3 trial (NCT04531982). Another compound, roluperidone, also targets the 5-HT 2A receptor, with additional antagonist activity at the σ 2 receptor and α 1 -adrenergic receptors. It showed efficacy in negative symptoms in a phase 2 trial but failed to meet the primary outcome in the subsequent phase 3 trial (Davidson et al., 2017;Davidson et al., 2022). It is questionable, however, if negative 5-HT 2A receptor modulation, either through an antagonist or inverse agonist, should be expected to affect negative symptoms because prior attempts with pure 5-HT 2 antagonists (e.g., M100907) have not shown clinical utility and because atypical antipsychotics with strong 5-HT 2A antagonism do not significantly improve negative symptoms (Kim, 2021).

Conclusions and future directions
This review has focused on both the challenges and new opportunities for developing medications that are effective for the treatment of negative symptoms of schizophrenia. Important areas in which there has been progress include an improved understanding of the neurobiology that underlies negative symptoms, improved methods for assessing negative symptoms in clinical trials, the development of new drugs that are potentially broadly effective for both positive and negative symptoms, and novel medications that may be utilized as adjunctive treatments to currently administered antipsychotics.
In the area of assessment, new scales that are more sensitive for measuring the anhedonia-amotivation component of negative symptoms are a welcome advance, even though these instruments use clinician ratings that are based on an individual's recollection of their past interest in rewards. Methods using digital phenotyping have the potential for measuring an individual's in-the-moment interest rather than relying on their recollections, which may prove to be quite valuable in capturing real-life effects in patients. Newer behavioral assessments allow for finer characterization of specific deficits in positive and negative reward systems and motivational factors capturing key drivers of anhedonia and amotivation -potentially meaningful for determining responses to pharmacological interventions. It will be important to include these newer measures in clinical trials in order to move the development of novel medications in the direction of precision medicine.

Role of funding source
Sumitomo Pharma America, Inc. and Otsuka Pharmaceutical Development & Commercialization Inc., provided funding to Pharma-Write, LLC, for medical writing and editorial support for this manuscript.

CRediT authorship contribution statement
All authors participated equally in the conceptualization and data curation for the article; writing, reviewing, and editing of the draft; and have provided approval of the final submitted manuscript.