Reward Processing in Mood Disorders and Schizophrenia: A Neurodevelopmental Framework

Abstract

Major depressive disorder (MDD), bipolar disorder, and schizophrenia involve disruptions in processing rewarding stimuli. In this paper, we propose that distinct mechanistic pathways underlie these disruptions in mood disorders versus schizophrenia, highlighting the importance of understanding these differences for developing personalized treatments. We summarize evidence suggesting that reward processing abnormalities in mood disorders are driven by dysregulated motivational systems, with MDD characterized by blunted responses to reward cues and bipolar disorder by heightened responses. In contrast, we argue that reward processing disruptions in schizophrenia do not reflect abnormalities in motivation or hedonic experience but rather impairments in the cognitive representation of past and future rewards, along with misdirected attention to irrelevant stimuli. To integrate these findings, we present a neurodevelopmental framework for the onset of mood and psychotic disorders, exploring how disruptions in normative brain development contribute to their pathophysiology, timing, and onset. Additionally, we move beyond viewing these conditions as homogeneous disorders, discussing how reward processing profiles may align with specific symptom dimensions.

Introduction

Mood and psychotic disorders, such as schizophrenia, are among the leading causes of ill-health and disability. Unipolar depression, without a history of hypomania or mania, affects over 300 million people globally each year and is the leading cause of disability worldwide, according to the World Health Organization (WHO), and is a major contributor to the global burden of disease (World Health Organization 2017). Unipolar depression frequently is classified as Major Depressive Disorder, which the Diagnostic and Statistical Manual of Mental Disorders(DSM-5^th ed; American Psychiatric Association 2013) defines as at least five symptoms of depression during the same two-week period that cause clinically significant distress or impairment. Depression frequently reoccurs, and approximately 70% of depressed individuals will experience another episode within 5 years (Birmaher et al. 2002).

Bipolar disorder is characterized by extreme swings of mood (euphoria or irritability versus sadness), behavior (excessive goal striving, energy, and talkativeness versus anhedonia and lethargy), and cognition (grandiosity and racing thoughts versus worthlessness) (Nusslock and Frank 2011). Bipolar disorder occurs on a continuum and encompasses three diagnoses: cyclothymia, bipolar II disorder, and bipolar I disorder (American Psychiatric Association 2013). All three diagnoses are characterized by hypomanic ⁄manic and depressive symptoms (except for instances of pure mania), but differ in severity, with bipolar I disorder being the most severe and cyclothymia the least severe. These bipolar spectrum disorders are relatively common, occurring in 4.4% of the US population (Merikangas et al. 2007), and are associated with significant work impairment, high rates of divorce, substance abuse, and heightened suicide spectrum behaviors (Miklowitz and Johnson 2006).

Schizophrenia is a severe mental illness characterized by significant alterations in perception, thoughts, mood, and behavior and reoccurring episodes of psychosis. Positive symptoms, such as hallucinations, delusions, and disorganized speech are often the reason a person presents for treatment. However, schizophrenia also involves negative symptoms, including apathy or lack of motivation, diminished emotional expression, and social withdrawal. Although many descriptions of schizophrenia emphasize positive symptoms, earlier conceptualizations viewed negative symptoms as core features of the disorder (Bleuler 1911), and negative symptoms substantially contribute to the illness’s long-term impact (Provencher and Mueser 1997). Schizophrenia affects 24 million individuals globally (James et al. 2018), and is associated with considerable social and occupational impairment and a reduced life expectancy (Charlson et al. 2018).

The causes of mood disorders and schizophrenia are complex, involving multiple biopsychosocial processes. However, growing evidence implicates corticostriatal circuits in the brain involved in reward processing, salience processing, approach motivation, and goal-directed behaviors (e.g., Johnson et al. 2012b; Pizzagalli 2014; Treadway 2016; Whitton et al. 2015). Reward processing relates to the value an individual places on potential rewards, the perceived probability of reward receipt, and learning about environmental cues predicting rewards (i.e., reinforcement or associative learning). These cues can be external (presence of a desired reward) or internal (expectancies of a reward). Salience processing assesses the novelty and intensity of a stimulus that makes it stand out, often independent of its value. Approach motivation and goal-directed behaviors regulate the pursuit of rewards and goals in the environment.

Unipolar depression (without a history of hypomania or mania, hereafter referred to as hypomania/mania) is associated with reduced positive emotion, reward processing, approach motivation, and reward-related brain activity in corticostriatal circuitry (Pizzagalli 2014; Nusslock and Alloy 2017; Treadway 2016). Bipolar disorder, in contrast, is associated with elevated reward processing (Alloy et al. 2015; Johnson et al. 2005, 2012b). It is suggested that a heightened sensitivity to rewards cause individuals with bipolar disorder to experience an excessive increase in approach motivation (i.e., hypomania/mania) during life events involving the pursuit of rewards, and an excessive decrease in approach motivation during life events involving the loss of rewards (e.g., definite failure). This suggests that risks for unipolar depression versus bipolar disorder are characterized by distinct and opposite profiles of reward processing and approach motivation, which has important implications for understanding differential risk for these conditions (Nusslock and Alloy 2017). Finally, the aberrant salience or dopamine hypothesis of schizophrenia proposes that positive and negative symptoms result from irregular (as opposed to chronically enhanced or reduced) reward processing that ascribes value or salience to irrelevant stimuli (resulting in positive symptoms) while failing to appropriately respond to or engage with relevant and meaningful reward stimuli (resulting in negative symptoms) (Howes and Kapur 2009; Whitton et al. 2015).

Given that reward processing is implicated in both mood disorders and schizophrenia, it’s reasonable to conclude that both mood and non-affective psychosis disorders involve shared abnormalities in reward-related brain function, and that reward processing is a common or transdiagnostic pathway to these diverse conditions. We disagree with this conclusion and instead agree with Whitton and colleagues (2015) that an equifinality perspective of reward processing abnormalities in mood disorders and schizophrenia may be preferable. Equifinality is the principle that a given end state (i.e., reward processing abnormalities) in different people or groups of people can be reached through different means or mechanisms. The central thesis of this paper is that potentially different mechanistic pathways subserve abnormalities in reward processing in mood disorders versus schizophrenia. In brief, we argue that the primary reward processing abnormalities in mood disorders are driven by dysregulated hedonic and motivational systems that have either blunted (in the case of unipolar depression) or heightened (in the case of bipolar disorder) responses to contextually appropriate reward cues. By contrast, we argue that the primary reward processing abnormalities in schizophrenia do not reflect an aberration in motivation or hedonic experience, but in the cognitive representation of past and future rewards and the misallocation of attention and salience to inappropriate or task-irrelevant stimuli. Thus, although individuals with mood disorders and schizophrenia may display similar deficits in reward processing on self-report, behavioral, and neural measures, we suggest that the underlying pathophysiology of reward processing abnormalities in mood disorders versus schizophrenia may be different. Examining these differences is critical for understanding the causes of mood disorders and schizophrenia and developing prevention and intervention strategies.

We have three goals for this paper. First, we provide an overview of abnormalities in reward processing in unipolar depression, bipolar disorder, and schizophrenia. We then extend this overview and apply an equifinality perspective first proposed by Whitton and colleagues (2015). We argue that the primary reward processing abnormality in mood disorders involves hedonic and motivational mechanisms, whereas the primary abnormality in schizophrenia involves disruptions in the cognitive representation of past and future rewards (Figure 1). We use “primary abnormality” intentionally, because we acknowledge that mood and psychotic symptoms clearly involve multiple and distributed psychological, behavioral, and biological processes. However, here we’re interested in considering the initial disruptions in brain development that set these processes in motion.

Figure 1: Reward Pathways to Mood Disorders Versus Schizophrenia.

We propose that the primary abnormality in reward processing in mood disorders involves disrupted hedonic and motivational mechanisms, whereas the primary abnormality in schizophrenia involves disruptions in the cognitive representation of past and future rewards.

Second, in medicine, disorders once considered unitary based on clinical presentation often turn out to be heterogeneous with meaningful subtypes. For example, under the DSM definition of a Major Depressive Episode, which requires the presence of 5 out of 9 possible symptoms, two individuals may be diagnosed with major depression while only sharing a single symptom in common (American Psychiatric Association 2013). This heterogeneity may mask associations that are related to specific symptoms, rather than the diagnostic category (Insel and Cuthbert 2015). Accordingly, where appropriate, we move beyond considering unipolar depression, bipolar disorder, and schizophrenia as unitary disorders and discuss relationships between specific profiles of reward processing and specific symptoms.

Third, adolescence is a developmental period considered an ‘age of risk’ for the onset of unipolar depression and bipolar disorder, with the steepest increase occurring between ages 15 and 18 (Avenevoli et al. 2015; Nusslock and Frank 2011). This rise in mood disorder symptoms has been linked partly to normative increases in reward-related brain activity occurring during adolescence (Forbes and Dahl 2012). Adolescence also involves increases in social rewards and stressors, and interpersonal stressors during adolescence are particularly likely to precipitate depression, whereas goal-striving can precipitate hypomanic/manic symptoms and episodes (Alloy et al. 2015; Hammen 2005). Thus, adolescence involves a confluence of maturing reward-related brain activity and both social and achievement-related life events that can generate risk for the onset of mood disorders. The onset of schizophrenia is later and typically occurs between the late teens and early 30s, with peak incidence occurring in the mid-twenties (McCutcheon et al. 2020b). This period is critical for the maturation of the prefrontal cortex, which is driven, in part, by microglial cells (Schalbetter et al. 2022). Microglia are the immune cells of the brain, where they devour (i.e., phagocytize) pathogens such as bacteria and viruses and clean up debris. They also play a critical role in the normative development of the prefrontal cortex during the transition to adulthood through trimming or pruning extraneous synapses and dendrites on cortical pyramidal neurons (Lenz and Nelson 2018). Growing evidence suggests that excessive synaptic pruning and dendritic retraction of prefrontal neurons by microglia may be an initial trigger of schizophrenia (Cannon 2015; McGlashan and Hoffman 2000). This excessive pruning can disrupt the balance between excitation and inhibition in the prefrontal cortex, which, in turn, can affect the top-down regulation of dopamine activity and reward processing in the striatum (Cannon 2015). Thus, from this perspective, altered dopamine activity in the subcortex in schizophrenia may be a downstream consequence of disruptions of the maturation of the prefrontal cortex. Collectively, this suggests that both risk for mood disorders and schizophrenia involve disruptions of the normative maturation of corticostriatal circuitry, albeit in different ways. Accordingly, the present paper proposes a neurodevelopmental framework for the onset of mood disorders and schizophrenia, exploring how disruptions in normative brain development can inform our understanding of the pathophysiology and timing of these mental health conditions.

Corticostriatal Reward Circuitry

Corticostriatal neural circuitry is central to the brain’s reward, motivational, and salience systems (Berridge and Kringelbach 2015; Haber and Knutson 2010). Primate research suggests that corticostriatal connections involve three pathways that parcellate the striatum into limbic, associative, and sensorimotor functional subdivisions based on their inputs and outputs (Figure 2) (Alexander and Crutcher 1990; McCutcheon et al. 2019). The corticostriatal limbic pathway is anchored to the ventral striatum, which is in the sub-cortex, and which receives dopaminergic input from a midbrain structure called the ventral tegmental area. The ventral striatum is centrally involved in assessing the value of rewarding stimuli, learning what cues predict rewards, and driving goal-directed motivation (Haber and Knutson 2010). The ventral striatum responds to both primary (e.g. food) and secondary (e.g. money) rewards and both social and non-social rewards, suggesting that different types of rewards share a ‘common neural currency’ in the brain (Berridge and Kringelbach 2015; Haber and Knutson 2010). Reducing activity in ventral striatum decreases one’s sensitivity to rewards and lowers motivation to pursue rewards in the environment (Treadway and Zald 2011).

Figure 2: Striatal Anatomy.

Summary of primate tracing studies mapping connections between the cortex (top row), striatum (middle), and midbrain (bottom) (Goldman-Rakic and Selemon 1986; Kunishio and Haber 1994; Künzle 1978). These studies show that corticostriatal connections run in three parallel pathways that parcellate the striatum into limbic, associative, and sensorimotor subdivisions. The corticostriatal limbic pathway projects from the orbitofrontal cortex to the ventral striatum, the corticostriatal associative pathway projects from the dorsolateral prefrontal cortex to the caudate and rostral putamen (i.e., dorsal striatum); and the corticostriatal sensorimotor pathway projects from the motor areas to the caudal putamen (i.e., dorsal striatum). For simplicity, portions of the striatum are not depicted in the figure. This figure is adapted from McCutcheon et al. 2019.

Both animal and human research highlight the central role of dopamine neurotransmission in the ventral striatum (Haber and Knutson 2010). Relative to placebo injection, ligand-based positron emission tomography (PET) research indicates that amphetamine injection increases ventral striatal dopamine, and these increases correlate with positive affective experiences (Volkow et al. 1999). Alcohol, cocaine, and secondary rewards such as gambling all increase dopamine release in the ventral striatum (Cox et al. 2009). As discussed below, however, dopamine activity in the ventral striatum is more strongly involved in reinforcement learning and the motivational pursuit of rewards (i.e., wanting), and less involved in savoring or enjoying rewards (i.e., ‘liking’; see Berridge and Kringelbach 2015 for review).

The corticostriatal limbic pathway also involves the orbitofrontal cortex (OFC), which is the portion of the cortex most directly involved in reward processing (Haber and Knutson 2010). Different portions of the OFC subserve different reward functions. The mid- and medial OFC code the subjective experience of pleasure, assess the probability of attaining a particular reward, and guide reward-based learning and decision making. The lateral OFC is sensitive to both rewards and punishments and is implicated in more general arousal and salience processing (Berridge and Kringelbach 2015). Growing evidence highlights the role that both glutamate and GABA, the brain’s primary excitatory and inhibitory neurotransmitters, play in providing ‘top-down’ regulation of reward processing and goal-directed behavior in the ventral striatum (Sesack et al. 2003). Glutamatergic and GABAergic neurons descend from the OFC to the ventral striatum where they modulate dopamine transmission to facilitate motivation and goal directed behaviors to pursue rewards in the environment (Sesack and Grace 2010).

The corticostriatal associative pathway involves dopaminergic projections from the substantia nigra in the midbrain to more dorsal portions of the striatum, including the dorsal caudate and dorsal putamen (Haber 2014; McCutcheon et al. 2019). Whereas dopamine signaling in the limbic or ventral striatum is sensitive to the potential value of stimuli, the associative or dorsal striatum tracks the novelty and intensity of stimuli and moderates communication between limbic and motor regions (Haber 2014; Seiler et al. 2022) . Thus, the associative striatum acts as an integrative hub for information processing that is sensitive to the salience of stimuli through converging cortical inputs, including afferent inputs from the dorsolateral prefrontal cortex (Averbeck et al. 2014). Finally, sensorimotor regions of the striatum receive afferent projections from motor and premotor cortical areas and influence voluntary movement through basal ganglia motor loops (McCutcheon et al. 2019). As this paper focuses on reward processing, sensorimotor pathways are beyond its scope; see Mittal et al. 2008 for research on motor abnormalities in schizophrenia.

Reward Processing in Unipolar Depression

Diminished positive emotions and a lowered sensitivity to rewarding stimuli are fundamental to depression (Lewinsohn and Graf 1973). Indeed, anhedonia, which involves diminished interest or pleasure in rewarding stimuli (American Psychiatric Association 2013), is a cardinal symptom of depression. According to reward hyposensitivity models of depression, individuals who have low sensitivity to rewards are vulnerable to depression because they have lower hedonic drive and approach motivation, are less likely to pursue rewarding and fulfilling life experiences, and experience significant drops in motivation and goal-directed behaviors following loss or failure (Nusslock and Alloy 2017; Pizzagalli 2014; Treadway 2016). Consistent with this perspective, both adults and youth in a depressive episode self-report lower sensitivity to rewards (Kasch et al. 2002), reduced extraversion and pleasure sensitivity (Kotov et al. 2010), and engage less frequently in reward pursuit and goal directed behaviors (Forbes 2009). Studies using behavioral paradigms report that individuals with depression display less sensitivity to rewards (Cella et al. 2010), expend less effort for rewards (Treadway et al. 2012), and fail to exhibit a response bias toward rewarding stimuli (Pizzagalli et al. 2008). At the neurophysiological level, unipolar depression is characterized by blunted reward responsiveness, as indexed by the feedback negativity (Foti and Hajcak 2009) to monetary gains versus losses, and lower approach-related brain activity (Nusslock and Alloy 2017). Finally, consistent with the vulnerability-stress perspective of the reward hyposensitivity model, life events involving failures and loss predict the onset and recurrence of depression (Nusslock and Alloy 2017) (blue pathway in Figure 3).

Figure 3: Reward Sensitivity Vulnerability-Stress Model of Mood Disorder Symptoms.

Reward hyposensitivity models of major depressive disorder (MDD) suggest that risk for MDD, and particularly motivational anhedonia, is associated with a chronically reduced sensitivity to rewards and a propensity to experience an excessive decrease in approach motivation during life events involving failure and loss (blue pathway, e.g., Nusslock and Alloy 2017; Pizzagalli 2014; Treadway 2016). Reward hypersensitivity models of bipolar disorder, by contrast, propose that risk for bipolar disorders is characterized by a chronically elevated sensitivity to rewards and a propensity to experience excessive approach motivation during events involving the pursuit or attainment of rewards, reflected in hypomanic/manic symptoms (red pathway), and an excessive decrease in approach motivation during failure or loss, reflected in bipolar depressive motivational anhedonia (green pathway, e.g., Johnson et al. 2012b; Nusslock and Alloy 2017).

Reduced sensitivity to rewarding stimuli appears to be a mood-independent, pre-existing vulnerability for depression. Even individuals in remission from depression report lower reward sensitivity (Pinto-Meza et al. 2006, although see Kasch et al. 2002 for conflicting findings). Studies examining reward-related differences in offspring of parents with and without depression suggest that offspring of depressed parents exhibit deficits in goal-directed behaviors (Mannie et al. 2015; although see Morris et al. 2015 for conflicting results). In prospective studies, reduced positive affect at age 3 predicted depressive cognitive styles at age 7 (Hayden et al. 2006) and was associated with a maternal history of depressive disorders (Durbin et al. 2005). Additionally, behavioral and neurophysiological measures of reduced reward sensitivity predict depressive symptoms and the first onset of MDD, suggesting that a blunted sensitivity to rewards may predate the onset of the illness (Nusslock & Alloy 2017; Nusslock et al. 2024).

Human neuroimaging studies report that individuals with MDD display lower activation in limbic portions of the striatum (i.e., the ventral striatum) during reward anticipation (Forbes 2009), reward receipt (Pizzagalli et al. 2009; Wacker et al. 2009), reward prediction errors (the difference between experienced versus predicted rewards; Kumar et al. 2008), and to other positive stimuli (although see Knutson et al. 2008 for contrary results). Low ventral striatal activation is present among individuals with MDD during remission (Schiller et al. 2013), suggesting that blunted reward responsiveness is independent of mood state. Low ventral striatal activation also is present in offspring of depressed individuals who have yet to develop a depressive episode (Gotlib et al. 2010). This implies that low reward-related brain function in the ventral striatum may represent a pre-existent vulnerability for depression, rather than a consequence of the illness, although future research is needed to confirm this claim.

The ventral striatum undergoes a normative, but rapid maturation during adolescence, resulting in a developmental period characterized by a heightened sensitivity to and motivation for rewards (Somerville and Casey 2010). For example, Steinberg and colleagues (2009) showed that sensitivity to monetary rewards peaks in adolescence, with a steady increase from late childhood to adolescence and a subsequent decline from late adolescence to adulthood. This maturation of the ventral striatum coincides with the ‘age of risk’ for the onset of depression (Avenevoli et al. 2015), and individuals who fail to show this normative increase in ventral striatal activation may face elevated risk for depression (Nusslock et al. 2024). Adolescence also is a period when the brain’s reward systems are particularly sensitive to stress, which can precipitate depression in adolescence. In line with this view, those exposed to adversity during adolescence often fail to show the normative increase in reward sensitivity during this time period and are at increased risk for depression (Casement et al. 2015). The ventral striatum is especially responsive to social rewards and stressors during adolescence (Crone and Dahl 2012). In fact, neural responses to social rewards during adolescence may distinguish depressed from non-depressed individuals as well as or better than neural responses to other classes of rewards (He et al. 2019). Future research should explore how the normative development of the ventral striatum interacts with social stressors during adolescence to heighten risk for depression.

Meta-analytic research reports that although individuals with MDD exhibit lower activation in the ventral striatum to rewards, they show enhanced activation in portions of the OFC (Ng et al. 2019). The logic here is that individuals with, and at risk for, depression tend to dampen positive and rewarding emotions, and engage the prefrontal cortex in a manner that decreases ventral striatal activity to rewarding emotions (Nusslock et al. 2024). Supporting this perspective are several studies reporting disrupted functional connectivity between the OFC and ventral striatum among individuals with depression, which is thought to reflect maladaptive regulation of striatal systems involved in reward processing and motivation by the OFC (Anderson et al. 2023; Quevedo et al. 2017). Growing evidence highlights the role that glutamate, the brain’s primary excitatory neurotransmitter, plays in providing this ‘top-down’ regulation of reward processing and goal-directed behavior (Sesack et al. 2003). Glutamatergic neurons descend from the OFC to the ventral striatum where they modulate dopamine transmission to facilitate motivation and goal directed behaviors to pursue rewards in the environment (Sesack et al. 2003; Sesack and Grace 2010). Animal studies report that disrupted glutamate signaling between the OFC and ventral striatum impairs motivation for rewards and reward-based decision making (Stanton et al. 2019). In humans, multiple meta-analyses of studies using magnetic resonance spectroscopy report that individuals with MDD have altered glutamate in the medial prefrontal cortex, including the OFC (Belleau et al. 2019), which specifically has been linked to anhedonia and motivational deficits (Stanton et al. 2019). Like the ventral striatum, the OFC goes through an important developmental inflection point during adolescence and early adulthood and increases its projections to the striatum (Somerville and Casey 2010). Collectively, this highlights how adolescence and early adulthood is a critical period for the maturation of the corticostriatal limbic circuit, including both subcortical reward processing and top-down regulation by the OFC.

Reward Hyposensitivity and Motivational Anhedonia.

Most of the research on reward hyposensitivity in unipolar depression has focused on individuals with a DSM diagnosis. As referenced above, however, there is a growing recognition of the need to move beyond considering psychiatric disorders as unitary constructs and to instead examine the relationship between brain systems and specific symptoms (Insel and Cuthbert 2015). It’s argued that this approach can unpack the heterogeneity that characterizes many psychiatric diagnoses to identify meaningful subtypes, facilitate a more precise understanding of the pathophysiology of psychiatric symptoms, and help translate this precision into more personalized prevention and intervention strategies (Nusslock and Alloy 2017). On this topic, reward hyposensitivity in the context of unipolar depression is strongly related to the symptom of anhedonia (Nusslock and Alloy 2017; (Nusslock et al. 2024). Anhedonia is one of two required symptoms for the DSM diagnosis of MDD and involves diminished interest or pleasure in all, or almost all, activities (American Psychiatric Association 2013). Anhedonia is common, although not pervasive in depression, and close to 40% of individuals with MDD experience clinically significant anhedonia (Nusslock and Alloy 2017). Self-report (Treadway and Zald 2011), behavioral (Pizzagalli et al. 2005; Treadway et al. 2012), and neurophysiological studies (Liu et al. 2014) suggest that anhedonia in the context of depression is associated with reward hyposensitivity. Human brain imaging studies indicate that, among individuals with MDD, anhedonia is associated with reduced ventral striatal activation (Pizzagalli 2014; Whitton et al. 2015) and deficits in functional connectivity between the ventromedial prefrontal cortex and ventral striatum during reward processing (Anderson et al. 2023).

Treadway and others (Pizzagalli 2014; Treadway 2016; Treadway and Zald 2011) argue that anhedonia in unipolar depression may reflect motivational as opposed to hedonic deficits. Supporting this perspective are preclinical studies indicating that dopaminergic activity in the corticostriatal limbic pathway implicated in depression and anhedonia is more involved in wanting and pursuing rewards (motivation) than savoring or enjoying rewards (hedonics; see Treadway 2016; Treadway and Zald 2011 for review). For instance, damaging dopamine synapses in the ventral striatum does not reduce pleasurable responses in rats (Berridge and Kringelbach 2015), whereas changing dopaminergic activity influences an organism’s motivation to seek out and work for rewards (Salamone et al. 2007) [the primary neurochemicals involved in pleasurable hedonic experiences are endogenous opioids and cannabinoids (see Treadway and Zald 2011 for review)]. Human studies show that individuals with MDD and anhedonia expend less effort for rewards compared to controls, and the longer the depressive episode, the more impaired the decision-making (Treadway et al. 2012). However, no differences were found in hedonic ratings of natural reinforcers like sucrose sweetness (Dichter et al. 2012). Taken together, this suggests that reward hyposensitivity in unipolar depression relates to a subtype of anhedonia characterized by motivational deficits (blue pathway in Figure 3). It also highlights the need for basic and clinical research on unipolar depression to distinguish motivational from hedonic components of anhedonia. Unfortunately, most of the clinical assessment and research on unipolar depression does not make this distinction, and if anything, it gives primacy to hedonic or pleasure deficits in anhedonia. For instance, the DSM-5 defines anhedonia as a “markedly diminished interest or pleasure in nearly all activities,” without specifying whether this reduction is due to a lack of motivation or a decrease in hedonic enjoyment (American Psychiatric Association 2013). Similarly, the Structured Clinical Interview for DSM Disorders (SCID; First et al. 2015) asks patients if they have “lost interest or pleasure in things they usually enjoy,” without distinguishing between motivational or pleasure-related causes. In a review of commonly used anhedonia measures, (Treadway and Zald 2011) found that these tools primarily focus on the experience of pleasure in response to positive stimuli, with minimal attention given to reduced motivation or drive. Thus, an important direction for future research in unipolar depression is to unpack heterogeneity within the symptom of anhedonia and further examine whether anhedonia in unipolar depression is primarily driven by motivational, as opposed to hedonic, deficits.

Reward Processing in Bipolar Disorder

Whereas risk for unipolar depression, particularly motivational deficits in anhedonia, is characterized by a blunted sensitivity to rewarding stimuli, risk for bipolar disorder is associated with a hypersensitivity to rewarding stimuli. These data have been summarized and conceptualized in the Reward Hypersensitivity Model of bipolar disorder (Alloy et al. 2015; Johnson 2005; Johnson et al. 2012b; Nusslock & Alloy 2017). This model proposes that a hypersensitivity to rewarding stimuli leads individuals at risk for bipolar disorder to experience an excessive increase in approach motivation and goal-directed behavior (e.g., working excessively long hours) during life events involving reward pursuit or attainment (e.g., when striving for or receiving a job promotion). In the extreme, this excessive increase in approach motivation is reflected in hypomanic/manic symptoms, such as elevated or irritable mood (when the goal-striving is blocked or frustrated), decreased need for sleep, increased psychomotor activation, extreme self-confidence, and pursuit of rewarding activities without attention to risks (red pathway in Figure 3). This model also proposes that individuals with, and at risk for, bipolar disorder experience an excessive decrease in approach motivation and goal-directed behaviors when they fail to attain goals or rewards, which, in turn, leads to depressive symptoms (green pathway in Figure 3). From this perspective, reward hypersensitivity is viewed as a risk factor for lability in approach motivation, with excessive increases in approach motivation (i.e., hypomania/mania) occurring during goal striving and reward attainment and excessive decreases in approach motivation (i.e., depression) occurring in response to irreconcilable reward loss (reward loss that is perceived to be remediable and merely a temporary thwarting of reward attainment should activate approach motivation and trigger anger and irritability symptoms of hypomania/mania – e.g., Carver and Harmon-Jones 2009).

Consistent with the Reward Hypersensitivity Model, Johnson and colleagues (2000, 2008) reported that life events involving the attainment of a reward or life goal predicted increases in manic but not depressive symptoms in patients with bipolar I disorder. Similarly, Nusslock and colleagues (2007) reported that late-adolescents with bipolar spectrum disorders were more likely to develop hypomania, but not depression, following a goal-striving event than bipolar adolescents who did not experience this event. Anger-provoking events that can activate the brain’s reward and motivational systems (e.g., goal obstacles, insults) also predict increases in hypomanic symptoms (e.g., Carver and Harmon-Jones 2009; Harmon-Jones et al. 2002). In contrast, failure or loss events frequently have been found to trigger bipolar depressive episodes (for reviews, see Alloy et al. 2015). The Reward Hypersensitivity Model also incorporates a transactional element, suggesting that individuals with heightened reward sensitivity not only react more intensely to goal- or reward-related life events, but also encounter these events more often due to “stress generation” processes (Hammen 2005; Alloy et al. 2015). It’s argued that reward hypersensitivity drives behaviors that increase exposure to the very reward-related events that provoke their reward systems. This increased exposure, in turn, triggers the onset of mood episodes through a two-hit model (see red and green pathways in Figure 3). In line with this perspective, individuals with bipolar disorder experience more reward-activating (e.g., goal-striving) and deactivating (e.g., failure) life events over a six-month follow-up than do demographically similar controls (Urošević et al. 2010).

Individuals with bipolar disorder also self-report an increased sensitivity to reward-related cues. This research has been facilitated by the development of the BAS scale within the Behavioral Inhibition System/Behavioral Activation System (BIS/BAS) scales, a self-report tool measuring reward and approach system sensitivity, created by Carver and White (1994). Individuals with bipolar I disorder (Meyer et al. 2001; Salavert et al. 2007), bipolar II disorder, cyclothymia (Alloy et al. 2008), and those prone to hypomanic symptoms (Meyer et al. 1999) self-report elevated BAS sensitivity compared to heathy controls or individuals with unipolar depression. In a study using a retrospective behavioral high-risk design, individuals with high self-reported BAS sensitivity were more likely to have a lifetime bipolar spectrum diagnosis compared to those with moderate BAS sensitivity (Alloy et al. 2006). Additionally, individuals with or at risk for bipolar disorder exhibit more ambitious goal setting, especially for goals related to fame, achievement, and financial success (Gruber and Johnson 2009; Johnson et al. 2012a), and show heightened behavioral, emotional, and cognitive responses to rewards (Johnson et al. 2005). Importantly, the relationship between bipolarity and reward sensitivity appears to be state-independent given it is not related to current levels of hypomania/mania and remains after controlling for bipolar mood symptoms (Alloy et al. 2015; Nusslock & Alloy 2017), and reward sensitivity continues to be elevated in remission relative to controls (Meyer et al. 2001).

Research also suggests that self-reported reward sensitivity predicts initial onset and progression of bipolar spectrum disorders. Heightened self-reported reward sensitivity is associated with a greater likelihood of having a lifetime bipolar spectrum diagnosis (Alloy et al. 2006), an increased risk of experiencing a first onset of the disorder (Alloy et al. 2012a), shorter intervals between recurrences of hypomanic/manic episodes (Alloy et al. 2008), a rise in manic symptoms among individuals with bipolar I disorder who have recovered (Meyer et al. 2001), and a higher chance of progressing to a more severe diagnosis from milder forms of the illness (Alloy et al. 2012b). Additionally, hypomanic/manic episodes can be triggered by both reward-seeking actions, such as applying for a job (Nusslock et al. 2007), and reward-achievement events, like receiving a job offer (Johnson et al. 2000).

Studies of human brain imaging report that individuals with bipolar disorder display heightened activation in limbic portions of the striatum (i.e., the ventral striatum) and the OFC to positive or rewarding stimuli (e.g., pictures of happy faces) compared to healthy comparison groups (Elliott et al. 2004; Hassel et al. 2008). Preliminary evidence suggests that heightened reward-related neural activation to positive stimuli may be independent of mood state, as this increased activation has been observed in both remitted (Hassel et al. 2008) and manic (Bermpohl et al. 2009; Elliott et al. 2004) individuals with bipolar disorder. However, some studies have reported decreased reward-related brain function to happy faces or other rewarding stimuli in bipolar individuals (Liu et al. 2012; Trost et al. 2014), suggesting that more research is needed on this topic.

The growing number of studies that have employed fMRI reward paradigms (e.g., card-guessing task, monetary incentive delay task) provide mixed support for the Reward Hypersensitivity Model of bipolar disorder. Nusslock et al. (2012) reported that bipolar I participants in remission displayed greater activation in the ventral striatum and both the medial and lateral OFC during reward anticipation compared to a healthy comparison group. The fact that individuals with bipolar disorder displayed elevated reward-related brain activity during remission suggests this profile of corticostriatal activity may reflect a trait-like marker of bipolar disorder. Contrary to this perspective, however, are two studies reporting that bipolar participants in remission display decreased reward-related brain activity to reward cues during the monetary incentive delay (MID) task (Johnson et al. 2019; Schreiter et al. 2016). Using an fMRI reward paradigm, Bermpohl and colleagues (2010) reported that bipolar I individuals in a manic episode displayed elevated left lateral OFC activation during reward anticipation, while healthy participants showed the inverse effect. Abler and colleagues (2008) reported that manic participants showed increased activation in the ventral striatum during reward omission compared to healthy participants, suggesting that bipolar individuals in a manic episode may have a reduced capacity to discriminate between rewards based on their actual value and relevance. Finally, in the context of bipolar depression, two fMRI studies reported reduced reward-related brain activity in both the ventral striatum (Redlich et al. 2015) and the anterior cingulate cortex (Chase et al. 2013) in individuals with bipolar I disorder during a major depressive episode compared to healthy controls. Additionally, one study found that greater depressive severity in bipolar participants was linked to lower ventral striatal activity in response to reward cues (Satterthwaite et al. 2015). These results suggest that bipolar depression has state-dependent effects on reward-related brain activity in the ventral striatum and ACC. However, Chase et al. (2013) also found that bipolar depressed participants showed increased activation in the lateral orbitofrontal cortex during anticipation, regardless of whether the trial involved reward or loss. This indicates that even in a depressive state, individuals with bipolar I disorder may retain heightened activation in certain areas of the corticostriatal limbic circuit.

The dimensional nature of bipolar disorder allows us to examine whether disruptions in reward-related brain activity predate the onset of more severe variants of the illness. Individuals with bipolar II disorder in remission exhibited increased ventral striatal and lateral OFC activation during reward anticipation compared to healthy controls (Caseras et al. 2013), although this study did not find the expected increase in ventral striatal activity in bipolar I individuals. In a PET study, participants with bipolar II depression also showed increased metabolism in the ventral striatum and OFC (Mah et al. 2007). Additionally, individuals with a hypomanic temperament, who had not yet developed bipolar disorder, displayed heightened ventral striatal and lateral OFC activation during reward processing (Harada et al. 2013), suggesting that elevated reward-related brain activity may represent a preexisting risk factor for bipolar disorder. Future studies using longitudinal high-risk designs are needed to test this claim.

Despite some inconsistencies in research using reward-based fMRI paradigms with bipolar participants, most of the research suggests that individuals with and at risk for bipolar disorder display a hypersensitivity to rewarding and goal-relevant stimuli. Researchers, including us and others (e.g., Johnson 2005; Johnson et al. 2012b), propose that this sensitivity is a risk factor for developing bipolar symptoms and episodes when exposed to reward-related life events. Taken together, this work suggests that risk for unipolar depression and bipolar disorder involve distinct and opposite profiles of reward sensitivity, approach motivation, and reward-related brain activity in the corticostriatal limbic circuit. Specifically, risk for unipolar depression, and particularly anhedonia, is associated with reward hyposensitivity and a propensity to experience an excessive decrease in approach motivation and goal-directed behavior during life events involving loss and definite failure. Bipolar disorder, by contrast, is characterized by reward hypersensitivity and a propensity to experience an excessive increase in approach motivation during events involving the pursuit or attainment of rewards, reflected in hypomanic/manic symptoms, and an excessive decrease in approach motivation during reward loss, reflected in bipolar depression.

These findings have important implications for understanding the pathophysiology of both unipolar depression and bipolar disorder, and differential risk for these two conditions. On this latter topic, both unipolar depression and bipolar disorder, as well as multiple other psychiatric conditions, are characterized by deficits in threat processing, executive control, and working memory, and these deficits may reflect transdiagnostic mechanisms (Hamilton et al. 2012; Etkin and Wager 2007). Understanding the mechanisms that drive transdiagnostic symptom clusters can provide insights into the overlapping features of psychiatric disorders and help explain the frequent comorbidity between diagnoses. However, although these transdiagnostic factors are clearly important, they may not help distinguish why certain individuals are more prone to one disorder than another. Accordingly, another goal of psychology and psychiatry is to uncover mechanisms that are specific to certain disorders, and that reflect signatures of differential risk for specific symptom profiles (Insel and Cuthbert 2015). Here, we argue that reward sensitivity and motivational systems in the brain may be particularly relevant for understanding differential risk for unipolar depression versus bipolar disorder, and that reward-related brain function may reflect biological markers that can help distinguish risk for depression versus bipolar disorder. Specifically, we propose that what differentiates risk for bipolar disorder versus unipolar depression is risk for mania, and one of the primary risk factors for mania is an individual’s tendency to experience an unusually heightened motivation to pursue rewarding stimuli. Therefore, understanding reward and approach-related processes is essential for differentiating bipolar disorder from unipolar depression.

Like unipolar depression, developmental neurobiology may help us better understand the onset and course of bipolar disorder. Although the median age of onset for bipolar disorders generally falls between 17 and 31 years, the initial surge in onsets typically occurs between ages 15 and 19 (Nusslock and Frank 2011). Indeed, admixture analyses suggest three distinct high-risk periods for developing bipolar disorder, with the earliest beginning in mid-adolescence (Bellivier et al. 2001). Thus, the transition from adolescence to young adulthood is an important developmental window for the onset of bipolar disorder. As noted, the corticostriatal limbic circuit undergoes rapid maturation during adolescence, making it sensitive to both rewards and stressors (Somerville and Casey 2010). However, unlike unipolar depression, there are not yet studies examining the arc of reward-related brain function during adolescence among individuals with, or at risk, for bipolar disorder, and this is an important topic for future research. Drawing on the Reward Hypersensitivity Model, we predict that individuals who experience a large increase in ventral striatal reactivity to rewards during adolescence should be at elevated risk for bipolar onset, and that early or recent stress exposure should amplify this profile. Importantly, measures of reward sensitivity might even one day help identify individuals with MDD who are at heightened risk for developing bipolar disorder (Alloy et al. 2012a). Nearly 40% of individuals diagnosed with MDD may be on the ‘soft’ bipolar spectrum, exhibiting subthreshold hypomanic symptoms (Nusslock and Frank 2011). These individuals have higher rates of comorbid impulse control and substance use disorders and are at greater risk of developing bipolar disorder in subsequent years (Nusslock and Frank 2011). Thus, measures of reward sensitivity might differentiate those at risk for unipolar depression from those at risk for bipolar disorder. Finally, while individuals at risk for unipolar depression may be sensitive to interpersonal stress during adolescence, those at risk for bipolar disorder may be particularly sensitive to achievement or goal-relevant life stressors. Supporting these individuals as they navigate the pressures and goal-strivings of adolescence may have therapeutic and preventative benefits.

Reward Hypersensitivity and Approach-Related Hypomanic/Manic Symptoms and Motivational Anhedonia.

Like unipolar depression, most of the research on bipolar disorder has focused on categorical diagnoses rather than specific symptom dimensions. Regarding hypomania/mania, we predict that reward hypersensitivity will be associated with a specific cluster of symptoms characterized by excessive approach motivation (red pathway in Figure 3). These symptoms include heightened energy, increased goal-directed activity, reduced need for sleep, greater confidence, and irritability when progress toward goals is obstructed. This prediction is based on the strong alignment between these clinical features and increased reward-related neural activity, which involves heightened approach motivation, reward sensitivity, and goal pursuit (Nusslock and Alloy 2017). The inclusion of decreased need for sleep in this cluster is supported by the relationship between reward processing with sleep-related factors (Murray et al. 2009), circadian influences (Alloy et al. 2015; Murray et al. 2009), and circadian genes (Forbes et al. 2012). Furthermore, the ventral striatum and the suprachiasmatic nucleus, which is the brain’s circadian clock, co-regulate each other and disrupting this communication generates manic like symptoms in both animals and humans (Alloy et al. 2015; Grippo et al. 2017). Increased confidence is part of this symptom cluster because heightened reward sensitivity, approach motivation, and bipolar spectrum disorders often are associated with a boost in confidence following the achievement of goals (Eisner et al. 2008; Johnson and Jones 2009). Irritability is included due to the neurobiological connection between anger and approach motivation (Carver and Harmon-Jones 2009) and the tendency for approach-related neural activity to increase when goal-pursuit is obstructed (Harmon-Jones et al. 2002). With respect to bipolar depression, we predict that individuals with reward hypersensitivity should be at risk for motivational anhedonia when facing certain failure or losses in the reward domain (green pathway in Figure 3). We make this prediction for two reasons. First, as noted above, we argue that reward hypersensitivity is a risk factor for a highly labile motivational system, with excessive increases or decreases in such motivation occurring during reward pursuit or reward loss, respectively (Alloy et al. 2015; Nusslock & Alloy 2017). Thus, predicting that reward hypersensitivity puts a person at risk for motivational anhedonia in the face of loss is entirely in line with the model. Second, and related, we propose that bipolar disorder is fundamentally a disorder of motivation and energy, rather than positive or negative emotion, and that these motivational disruptions are anchored to the brain’s corticostriatal limbic circuit (Alloy et al. 2015; Nusslock & Alloy 2017). Thus, any disruptions to this circuit in either an upward or downward direction should be reflected in extreme motivational states, including motivational anhedonia. Future research is needed to test these predictions.

Reward Processing in Schizophrenia

Abnormalities in dopamine transmission in corticostriatal circuitry have long been considered a primary pathology of schizophrenia (Howes and Kapur 2009; Fusar-Poli and Meyer-Lindenberg 2013). This perspective first emerged, and has been refined over the years, because of the fact that all licensed antipsychotic drugs block striatal D2 receptors (Howes and Kapur 2009; Kapur et al. 2000), and drugs that directly (e.g., amphetamine and cocaine) or indirectly (e.g., cannabis and ketamine) increase striatal dopamine transmission can induce or worsen psychotic symptoms (Howes and Kapur 2009). Interestingly, although some molecular imaging studies report increased density of D2 receptors in the striatum in schizophrenia, this finding is not consistent and the effect is modest at best (Howes et al. 2009). The research instead suggests that schizophrenia is characterized by elevated presynaptic synthesis or production of dopamine in the striatum (Howes et al. 2009). Elevated presynaptic striatal dopamine activity is associated with heightened psychotic symptoms, and blocking this heightened transmission, either by decreasing dopamine levels or blocking dopamine transmission, leads to a reduction in psychotic symptoms for most patients (see Howes and Kapur 2009 for review). There also is evidence that individuals at elevated risk for schizophrenia, but who have not yet developed the illness, display increased dopamine synthesis in presynaptic neurons in the striatum (Howes et al. 2020). Furthermore, individuals with a greater capacity to synthesize dopamine in striatal neurons during the prodromal or clinically high-risk (CHR) period are more likely to transition to psychosis and develop schizophrenia (Girgis et al. 2021; Howes et al. 2020; Van Hooijdonk et al. 2022). This latter finding suggests that increased striatal dopamine synthesis may be a pre-existent risk factor for schizophrenia, as opposed to a consequence of the illness.

Given the role that striatal dopamine plays in processing rewards and incentive motivation, one might assume from these findings that individuals with schizophrenia display a hypersensitivity to rewarding stimuli and increased approach motivation, as seen in bipolar disorder. This is not the case, and these data are instead interpreted in the context of the aberrant salience or dopamine hypothesis of schizophrenia (Howes and Nour 2016; Kapur 2003). This hypothesis proposes that negative and positive symptoms result from inappropriate (as opposed to chronically reduced or enhanced) dopamine release that fails to appropriately respond to meaningful reward cues (resulting in negative symptoms), while ascribing elevated or aberrant salience to irrelevant stimuli (resulting in positive symptoms). Evidence supporting the Aberrant Salience Hypothesis comes from research on negative symptoms in schizophrenia, which typically involve anhedonia, decreased affective expression, diminished motivation, and self-reported decreases in pleasurable experiences (see Strauss and Gold 2012 for a review). Clinically, these symptoms resemble the anhedonia and motivational deficits seen in unipolar depression. However, unlike unipolar depression, emerging evidence suggests that negative symptoms in schizophrenia do not primarily reflect deficits in the capacity for hedonic experiences or motivation. Instead, they seem to involve difficulties representing the value of rewarding experiences within cognition and working memory (Gold et al. 2008, 2013). For instance, although individuals with schizophrenia report lower positive affect and fewer pleasurable experiences in retrospective, prospective, and hypothetical self-reports (Strauss and Gold 2012), they tend to exhibit normal affective responses to positive stimuli, such as images, faces, sounds, and food, when tested in laboratory settings (Cohen and Minor 2010). Studies using naturalistic experience sampling offer a similar perspective: whereas individuals with schizophrenia experience fewer positive events in daily life (Myin-Germeys et al. 2000), they report comparable increases in positive emotion to those of healthy individuals when engaged in pleasurable activities (Oorschot et al. 2013). Furthermore, studies using the EEfRT task developed by Treadway et al. (2012) report that individuals with schizophrenia do not exhibit an overall reduction in effort expenditure for reward (as demonstrated in individuals with MDD), but instead fail to select high effort options when it is advantageous to do so (Gold et al. 2013; Barch et al. 2014). Complimenting these findings is evidence of cognitive and working memory deficits in individuals with schizophrenia during non-current reward processing [e.g., retrospective, prospective, hypothetical self-reports (Gold et al. 2008; 2013)], and compromises in brain systems that represent the value of outcomes and plans (Barch and Ceaser 2012; Barch and Dowd 2010). This suggests that negative symptoms (e.g., anhedonia) in schizophrenia may involve deficits in the ability to cognitively represent past and future rewards, as opposed to hedonic deficits in responding to and/or savoring rewards in the moment. These deficits may increase noise in the system, “drowning out” context appropriate responses to rewarding stimuli in the environment. The net result is a reduced motivational drive that over time leads to negative symptoms, such as social withdrawal and anhedonia (Howes et al. 2020; Neumann et al. 2021; Li et al. 2020).

With respect to positive symptoms of schizophrenia, the aberrant salience hypothesis argues that an abnormal synthesis and release of striatal dopamine leads to inappropriate attribution of significance or salience to neutral stimuli (Howes and Nour 2016). Over time, this process gives rise to psychotic symptoms, particularly delusions and hallucinations, as the individual attempts to explain these experiences of misassigned salience. Thus, psychosis from this perspective is viewed as a consequence of dopamine-driven aberrant salience, interpreted through the person's pre-existing cognitive and cultural frameworks. This concept explains how similar dopamine dysfunction can produce varied clinical presentations depending on the individual and their sociocultural background. In line with this perspective are studies reporting that dopamine signaling is substantially up-regulated in positive symptoms of schizophrenia (e.g., psychosis, hallucinations, delusions; see Fusar-Poli and Meyer-Lindenberg 2013 for meta-analytic review), and fMRI studies highlighting associations between aberrant striatal responses and a propensity for psychotic symptoms (see Howes and Kapur 2009 for review).

Taken together, these findings suggest that the pathophysiology of schizophrenia involves the misallocation of striatal dopamine and salience to inappropriate cues, rather than abnormally elevated or attenuated dopamine transmission to appropriate reward cues. Consistent with this perspective, striatal abnormalities in schizophrenia may be localized to associative or dorsal regions of the striatum, as opposed to limbic or ventral regions (McCutcheon et al. 2019). As noted, dopamine signaling in the dorsal striatum is particularly sensitive to the salience, novelty, and intensity of stimuli rather than their value (Haber 2014). A recent meta-analysis of PET studies reported that dopaminergic function was elevated among individuals with schizophrenia in the dorsal striatum, but that it was not elevated in limbic or ventral striatal regions (McCutcheon et al. 2018). Research on presynaptic dopamine function in individuals at clinical high risk for psychosis similarly has identified the greatest abnormalities in the dorsal striatum (Kegeles et al. 2010; Howes et al. 2009). Furthermore, conversion from clinical high risk to psychosis in these studies was associated with a progressive increase in dopamine synthesis in the dorsal striatum, while no significant changes were observed in the ventral striatum. It’s important to recognize, however, that some studies have shown alterations in ventral striatal function in individuals with psychotic disorders (e.g., Radua et al. 2015). Thus, further research is needed to investigate the topography of striatal abnormalities in schizophrenia, as this work has important implications for understanding disruptions in motivation versus salience processing in the disorder.

The aberrant salience hypothesis not only focuses on disruptions in striatal dopamine synthesis, but also on misattributions that the prefrontal cortex makes of the causes and sources of salient stimuli. As noted, individuals with schizophrenia display cognitive and working memory deficits on tasks that engage the prefrontal cortex (Barch and Ceaser 2012; Morris et al. 2013) and these deficits predate the onset of the illness, suggesting they are a pre-existent vulnerability (Fusar-Poli et al. 2012; Giuliano et al. 2012). Early research focused on linking these cognitive impairments and negative symptoms in schizophrenia to low levels of dopamine activity in the prefrontal cortex (Tamminga 2006; Goldman-Rakic et al. 2004). This facilitated an earlier variant of the aberrant salience hypothesis, which proposed that reduced dopamine activity in the prefrontal cortex mediated negative symptoms and cognitive impairments in schizophrenia, whereas elevated striatal dopamine accounted for positive symptoms and active psychosis (Davis et al. 1991). Support for this perspective came from animal studies reporting that lesions to dopamine neurons in the prefrontal cortex generated increased levels of dopamine and D2 receptor density in the striatum (Pycock et al. 1980), suggesting an integrated corticostriatal dopamine circuit that has implications for understanding the pathophysiology of schizophrenia (Davis et al. 1991).

Although research on reduced dopamine activity in the prefrontal cortex in schizophrenia has been useful, it’s clear that cortical abnormalities in this illness are more complicated and dynamic than simply blunted dopamine. One process receiving attention is the balance between excitatory and inhibitory neural activity in the prefrontal cortex (Cannon 2015; McCutcheon et al. 2020a). This balance is mediated, in part, by N-methyl-D-aspartic acid (NMDA) receptors on cortical pyramidal neurons, which are activated by the excitatory neurotransmitter glutamate. These glutamatergic neurons synapse onto interneurons expressing the inhibitory neurotransmitter g-aminobutyric acid (GABA), which then project back to glutamate neurons, completing a self-contained local circuit. Growing evidence indicates that schizophrenia involves weakened NMDA function, resulting in an imbalance between these excitatory glutamatergic and inhibitory GABAergic neurons. For instance, schizophrenia is associated with polymorphisms in genes involved in the glutamate cascade (Ripke et al. 2014), and studies using magnetic resonance spectroscopy (MRS) report increased glutamate levels in the prefrontal cortex among antipsychotic naïve schizophrenia patients in their first psychotic episode (Dempster et al. 2015). Furthermore, increased glutamine, a precursor to glutamate, has been observed in the anterior cingulate cortex in clinical high-risk cases (Stone et al. 2009), suggesting that glutamatergic dysfunction predates the onset of psychotic symptoms (although see Lutkenhoff et al. 2010) for contrary findings.

Relevant to the present paper, the balance between glutamate and GABA in the prefrontal cortex plays an important role in “top-down” or cortical regulation of dopamine activity and reward processing in the striatum. Both glutamatergic and GABAergic neurons descend from the medial prefrontal cortex to the striatum (Sesack and Grace 2010), and disrupting these pathways alters motivation and reward-based decision making in both animals and humans (John et al. 2012; Stanton et al. 2019). This raises the possibility that reward processing abnormalities in schizophrenia may not fully originate in the striatum, but instead may be driven, in part, by disrupted “top-down” regulation of striatal dopamine activity by glutamatergic and GABAergic neurons in the prefrontal cortex. Research on drugs like ketamine and phencyclidine (PCP), which alter the balance of glutamate and GABA in the prefrontal cortex through blocking NMDA receptors, support this perspective. NMDA hypofunction induced by these drugs results in changes in dopamine signaling in the striatum that resemble patterns observed in schizophrenia (Jentsch et al. 1997). Furthermore, abusers of ketamine and PCP frequently experience psychotic symptoms (Krystal et al. 1994), and low doses of ketamine worsen psychotic symptoms in individuals with schizophrenia (Lahti et al. 2001).

A final link in this chain requires a neurodevelopmental perspective (e.g., Cannon 2015). During the transition from adolescence to early adulthood, the prefrontal cortex undergoes a normative thinning that is essential for the maturation of certain cognitive functions (Tamnes et al. 2010). Microglia, which are the immune cells of the brain, play a central role in this cortical thinning. Through phagocytosis, microglia prune and retract extraneous synapses and dendrites that have been flagged for removal (Lenz and Nelson 2018). Some suggest that individuals at risk for schizophrenia experience an excessive removal of synapses and dendrites in prefrontal glutamatergic and GABAergic neurons via microglia during the transition to adulthood (Cannon 2015; McGlashan and Hoffman 2000). For example, individuals at elevated risk for schizophrenia who converted to psychosis showed a greater progressive loss of gray matter thickness compared to nonconverters and controls (Cannon et al. 2015), and postmortem studies of patients with schizophrenia report increased reductions in dendritic spines and synapses in the prefrontal cortex (Glausier and Lewis 2013). Numerous studies also support the hypothesis that an excessive loss of synapses and dendrites in schizophrenia is mediated, at least in part, by microglia, and PET studies report that individuals with schizophrenia have greater activated microglia and neuroinflammation in the prefrontal cortex (Doorduin et al. 2009; Goldsmith et al. 2023; van Berckel et al. 2008; although see Kenk et al. 2015 for contrary findings). In longitudinal research, higher levels of neuroinflammatory markers produced by microglia predicted steeper rates of gray matter reduction in the prefrontal cortex among individuals at heightened risk for schizophrenia who converted to psychosis (Cannon et al. 2015). Importantly, weakened NMDA receptor function may be an initial trigger of this excessive gray matter loss in the prefrontal cortex of individuals with schizophrenia given that inactive and weakened synapses are the first to be targeted for removal and pruning by microglia (Faust et al. 2021). By reducing and dysregulating activity at glutamate/GABA synapses, weakened NMDA receptor function could initiate overly aggressive pruning by microglia. Future research is needed to test this causal claim.

The topography of prefrontal regions showing aggressive microglia pruning in schizophrenia can shed light on early symptoms and disruptions in reward and salience processing (Cannon 2015). The first areas to be targeted appear to be dorsal and medial portions of the prefrontal cortex that are involved in source monitoring and prediction error, among other things (Cannon 2015). Source monitoring involves identifying and remembering whether an item or event is real or imagined, and prediction errors are activated when an observations or experiences are inconsistent with an internal model of reality. Weakening these processes can generate a permissive environment for false attributions and delusions to emerge. Linking this back to the aberrant salience hypothesis, this also can set the stage for ascribing elevated or aberrant salience to irrelevant stimuli and difficulties representing the value of rewarding experiences in cognition and working memory (Howes and Kapur 2009; Nusslock and Alloy 2017). Consistent with this view, individuals with schizophrenia display blunted neural prediction errors to contextually relevant cues and enhanced prediction error to contextually irrelevant stimuli (Morris et al. 2013). Subsequent areas of the brain targeted by microglia pruning include the dorsolateral prefrontal cortex (Paolicelli et al. 2011). This portion of the cortex is heavily connected to the dorsal striatum (McCutcheon et al. 2019), and disruptions in the maturation and pruning of the dorsolateral prefrontal cortex could exacerbate aberrant salience processing in the associative striatum.

Taken together, this literature on NMDA hypofunction, glutamate/GABA imbalance, and excessive microglia pruning of prefrontal neurons provides a pathway through which reward processing abnormalities in schizophrenia could primarily be driven by cognitive rather than hedonic mechanisms (Figure 4). Alternatively, early disruptions in striatal dopamine signaling could plausibly precipitate cortical and cognitive deficits in the prefrontal cortex. Under these conditions, striatal dopamine abnormalities could have upstream effects on reward-based learning in the cortex that in turn weaken NMDA receptors and glutamate/GABA balance (Kellendonk et al. 2006). Future research using multi-modal and multi-wave longitudinal designs is needed to test these perspectives. Regardless, it’s clear that disruptions in reward processing in schizophrenia involve complex cellular interactions between glutamate, GABA, neuroinflammation, and dopamine that disrupt the brain’s capacity to cognitively represent rewards.

Figure 4: Neurocognitive Model of Abnormal Reward and Salience Processing in Schizophrenia.

Building on the work of Cannon (2015), we propose a neurodevelopmental model of schizophrenia in which weakened NMDA receptor function disrupts the balance between excitatory glutamatergic and inhibitory GABAergic neurons in the prefrontal cortex (PFC). This imbalance drives excessive pruning of synapses and dendrites in the dorsomedial PFC during the transition to adulthood, impairing ‘top-down’ cortical regulation of dopamine synthesis and release in the striatum. These disruptions in turn affect reward and salience processing, increasing risk for both positive and negative symptoms of schizophrenia. Future research is needed to test this model and its causal pathways. NMDA, N-methyl-D-aspartic acid; GABA, g-aminobutyric acid.

Conclusion: An Equifinality Perspective

There is a growing interest in identifying mechanisms that are transdiagnostic or common across psychiatric disorders and symptoms (Insel and Cuthbert 2015). Given that reward processing has been implicated in anhedonia, hypomania/mania, and both positive and negative symptoms of schizophrenia, it’s reasonable to conclude that abnormalities in reward processing are a common or transdiagnostic pathway to these diverse conditions. We disagree, in part, and instead agree with Whitton et al. (2015) that an equifinality perspective of reward processing abnormalities may be preferable. As noted, equifinality is the principle that a given end state can be reached by different means or mechanisms.

Whitton et al. (2015) were the first to highlight anhedonia as an example of equifinality in the context of unipolar depression and schizophrenia. They argue that although anhedonia has a similar clinical presentation in unipolar depression and schizophrenia, it is likely driven by distinct pathophysiological mechanisms across these two disorders. Anhedonia in unipolar depression is argued to be driven by a reduced capacity for hedonic experience, motivation, or decision making, whereas anhedonia in schizophrenia is argued to be a consequence of deficits in higher-order cognitive systems involved in working memory and value representation of past and future rewards (see also, Barch et al. 2016).

We also agree with Whitton et al. (2015) that an equifinality perspective is relevant for understanding the nature of the relationship between bipolar symptoms of hypomania/ mania and positive symptoms of schizophrenia. Both these conditions are characterized by elevated dopamine signaling in striatal circuitry (Berk et al. 2007; Fusar-Poli and Meyer-Lindenberg 2013). In bipolar disorder, excessive activity in limbic or ventral portions of the striatum is typically directed towards contextually appropriate reward cues in one's environment. As discussed in the present paper, this reward hypersensitivity can then result in an excessive increase in approach and reward-related affect, which, in the extreme, is reflected in hypomanic/manic symptoms (e.g., (Alloy et al. 2015; Johnson 2005) By contrast, positive symptoms of schizophrenia appear to be associated with elevated dopamine signaling in associative or dorsal portions of the striatum to irrelevant or task inappropriate cues (e.g. Howes and Kapur 2009). Furthermore, elevated striatal dopamine signaling in hypomania/mania and schizophrenia may be driven, in part, by distinct pathophysiological mechanisms. Whereas elevated striatal signaling in risk for hypomania/mania is associated with an abnormally elevated hedonic or motivational response to reward cues (e.g., Nusslock and Alloy 2017), elevated striatal signaling in schizophrenia may be driven more by cognitive deficits in the cortex that lead to the misallocation of salience to inappropriate or irrelevant stimuli (Barch and Ceaser 2012; Gold et al. 2008, 2013). Drawing on the work of Cannon and colleagues (2015), we outline here a neurodevelopmental model of these cognitive deficits in schizophrenia and apply them to our understanding of aberrant reward processing in schizophrenia.

In summary, we agree with Whitton et al. (2015) that even though reward processing abnormalities have been observed across multiple disorders, an equifinality perspective of these abnormalities may be preferable than a transdiagnostic approach. Such a perspective does a better job of recognizing that reward processing is not a unitary construct and acknowledging that a symptom observed across different disorders may be driven by distinct pathophysiological pathways. Identifying these pathways has important implications for understanding the causes and consequences of these disorders and developing precise and personalized interventions.

Summary Points

1. Risk for major depressive disorder (MDD) and bipolar disorder are characterized by distinct and opposite profiles of reward processing and approach motivation, with MDD involving blunted responses to rewarding stimuli and bipolar disorder involving heightened responses.

2. Decreased sensitivity to rewards in MDD relates to a specific subtype of anhedonia characterized by diminished motivation. In contrast, we propose that reward hypersensitivity in bipolar disorder is associated with a cluster of hypomanic/manic symptoms characterized by excessive approach motivation, including heightened energy, increased goal-directed activity, reduced need for sleep, greater confidence, and irritability when progress toward goals is obstructed.

3. Adolescence is a period of risk for the onset of MDD and bipolar disorder, driven by disruptions in the normative development of the corticostriatal limbic circuit, which plays a key role in reward processing and approach motivation. Whereas a failure to show a normative increase in ventral striatal activation during adolescence is a risk factor for MDD, we propose that an excessively large increase in ventral striatal activation to rewards during adolescence may predispose individuals to bipolar disorder.

4. Unlike mood disorders, disruptions in reward processing in schizophrenia are not primarily driven by abnormalities in motivation or hedonic experience. Instead, they stem from impairments in the cognitive representation of past and future rewards, coupled with misdirected attention to irrelevant stimuli.

5. The literature on NMDA hypofunction, glutamate/GABA imbalance, and excessive microglia pruning of prefrontal neurons during the transition to adulthood highlights a potential neurodevelopmental pathway through which reward processing abnormalities in schizophrenia could primarily be driven by cognitive rather than hedonic mechanisms.

6. We agree with (Whitton et al. 2015) that an equifinality perspective, rather than a transdiagnostic approach, is preferable for understanding reward processing abnormalities across mood disorders and schizophrenia. We suggest that different pathophysiological pathways drive reward processing abnormalities in these two conditions. Specifically, we propose that reward processing abnormalities in mood disorders are primarily driven by disruptions in motivational mechanisms, whereas the primary abnormality in schizophrenia involves disruptions in the cognitive representation of past and future rewards.

Future Directions Points

1. Further research is needed to test the central thesis of this paper: that different mechanistic pathways underlie abnormalities in reward processing in mood disorders versus schizophrenia.

2. Research is needed to test our proposition that bipolar disorder, particularly hypomania/mania, is driven by disordered states of motivation and energy, rather than positive or negative emotion. Additionally, although evidence supports the link between decreased sensitivity to rewards in MDD and motivational anhedonia, future studies are needed to test our hypothesis that heightened reward sensitivity in bipolar disorder is associated with a cluster of hypomanic/manic symptoms characterized by excessive approach motivation.

3. Unlike unipolar depression, research has yet to examine the trajectory of reward-related brain function during adolescence in individuals with, or at risk, for bipolar disorder.

4. Further research should explore the topography of striatal abnormalities in mood disorders and schizophrenia, as this work has important implications for understanding disruptions in motivation versus salience processing in these disorders.

5. Future research employing multi-modal, multi-wave longitudinal designs is needed to test the proposed neurodevelopmental pathways through which reward processing abnormalities in schizophrenia are primarily driven by cognitive rather than motivational or hedonic mechanisms.