Delay discounting abnormalities are seen in first-episode schizophrenia but not in bipolar disorder

Abstract

Delay discounting (DD) is the phenomenon of individuals discounting future rewards as a function of time. It has been studied extensively in chronic schizophrenia (SZ) and the results of these studies have been variable. Comorbidity in chronic samples could be one reason for the mixed findings and studies in first-episode (FE) samples are surprisingly lacking. Bipolar disorder (BP) which shares some genetic and symptom features with SZ could serve as an interesting comparison group for DD but has been underexplored. Here we present the first study that combines FE SZ, FE BP with psychotic features, as well as healthy controls and study DD with two versions of the task. We found that SZ showed steeper discounting than HC and BP on the well-validated Kirby DD task. SZ showed no difference than HC on a separate DD task with smaller rewards presented with decimal places and shorter delays. As a preliminary finding, DD was found to be positively related to positive symptoms in FE SZ, while no relationship was found between negative symptoms and DD. In addition, we found comparable DD in BP compared to HC. Ultimately, our data may help elucidate the psychopathology in SZ and BP during intertemporal decision making.

Introduction

In a digital age where many things can be obtained instantly, it is becoming difficult to resist an immediate reward and wait for a larger payoff in the future. The propensity to discount future rewards as a function of time is called delay discounting (Kirby, 1997). DD is considered one dimension of impulsivity and closely linked to other dimensions such as self-report impulsivity (MacKillop et al., 2016; de Wit et al., 2007). Individual differences in DD are argued to be a stable, trait-level feature (Horan et al., 2017; Kirby, 2009) which is often associated with cognitive functions, such as working memory capacity and intelligence (Shamosh et al., 2008). These intertemporal choices are suggested to be modulated by a competition between two interacting systems, with the limbic system biased towards immediate payoffs while the lateral prefrontal cognitive control system favors more future-oriented goals achieved through self-control (Figner et al., 2010; McClure et al., 2004; Volkow and Baler, 2015).

Given the systems subserving intertemporal choices, schizophrenia patients (SZ) who show impaired cognitive functions, such as working memory maintenance (Cohen et al., 1999) as well as diminished motivation towards receiving distant reward (Barch and Dowd, 2010; Gold et al., 2008), would be expected to also show abnormalities during delay discounting, presumably through a bias towards immediate choices. Indeed, a number of studies have found that chronic SZ patients showed steeper discounting in the form of higher discounting rate (k; Ahn et al., 2011; Brown et al., 2018; Heerey et al., 2011; Heerey et al., 2007; Weller et al., 2014; Yu et al., 2017). The discounting rate is derived from a hyperbolic model in which the present value of a future reward is inversely proportional to the delay as well as the k (Kirby, 1997; but see alternative models e.g. McClure et al., 2004). Therefore higher k indicates steeper discounting and more impulsive decision-making. Elevated discounting rates in SZ have been found to be correlated with lower working memory capacity (Ahn et al., 2011; Brown et al., 2018; Heerey et al., 2007), verbal reasoning (Yu et al., 2017) and the ability to forecast future events (Heerey et al., 2011), further linking cognitive impairments to DD. Nevertheless, a number of studies reported no DD difference between chronic SZ and healthy controls, making it difficult to make a definitive statement about DD in chronic SZ (Avsar et al., 2013; MacKillop and Tidey, 2011; Wing et al., 2012). Presumably, the susceptibility of DD to contextual influences such as incidental episodes of affective state (Lempert and Phelps, 2016) may help explain some of the inconsistencies.

On the other hand, there is substantial evidence showing that schizophrenia spectrum disorders share genetic and symptom overlap with bipolar disorder (Cardno and Owen, 2014; Craddock et al., 2009; Lee et al., 2013; Lichtenstein et al., 2009), which may offer an interesting comparison group that is particularly characterized by hypersensitivity for reward as well as a diminished ability for self-control (Reddy et al., 2014; Swann et al., 2009; Whitton et al., 2015). These characteristics also tend to be associated with adverse consequences such as substance abuse (American Psychiatric Association, 2013; Swann et al., 2003) and are theoretically linked to more impulsive decision-making and steeper discounting. In particular, both SZ and BP show elevated trait impulsivity across multiple dimensions (Fortgang et al., 2016), which are correlated with brain structural differences at orbitofrontal cortex (Nanda et al., 2016). Studying DD abnormalities across SZ and BP could help further our understanding of impulsive decision making across diagnostic boundaries.

Interestingly, a limited number of studies investigating DD in BP showed mixed results. Following the seminal work of Ahn et al., 2011 showing steeper discounting in chronic BP, a recent study failed to replicate the finding (Brown et al., 2018). It was argued that this null finding was predominantly driven by a depressed sample of BP patients. Despite the suggestion that manic and depressive mood states tap into different aspects of impulsivity (Swann et al., 2008), it is not clear how they may differentially influence intertemporal decision-making (i.e., DD). Along this line, one study using individuals at risk for mania reported no DD differences relative to HC (Meyer et al., 2015). Interestingly, using an adolescent sample, Urosevic and colleagues found that BP patients did not show age-related DD improvements (older individuals are more able to wait for a larger reward) as observed in controls (Urosevic et al., 2016), highlighting potentially different developmental trajectories across the two groups.

The majority of the published literature focuses on DD in chronic SZ and BP. At earlier stages of illness, more proximal to the first episode, one hopes to minimize influences that are secondary to the illness (e.g., chronic medication effects, years of substance use, metabolic syndrome) which could serve as a first step addressing the inconsistencies among the previous DD findings. Furthermore, impulsivity is related to clinical symptoms in SZ and BP (Nanda et al., 2016), and DD specifically is associated with real-world impulsive behaviors (Reimers et al., 2009). A better understanding of DD process in SZ and BP would provide valuable information to clinical practice, especially for FE patients where early intervention is associated with better treatment outcomes (McFarlane et al., 2014). To our knowledge, no study has reported DD using a FE sample with both SZ and BP with psychotic features, as we do in the current study. Two versions of DD tasks were tested, the Kirby Monetary Choice Questionnaire following previous work (Heerey et al., 2011; Heerey et al., 2007; Kirby et al., 1999; Meyer et al., 2015) and a version evaluating the decimal effect of DD in which each monetary offer was shown either in rounded or decimal values (Fassbender et al., 2014). These two versions were selected in order to test comparability to the existing literature using the Kirby version and to evaluate whether changes in presentation style of monetary offers (e.g. decimal numbers) would reduce impulsive choices in individuals with SZ and BP, as has been previously shown in individuals with attention deficit/hyperactivity disorder (Fassbender et al., 2014). We further anticipated higher discounting rates for SZ relative to HC across both DD tasks, which were expected to be associated with negative symptoms (Heerey et al., 2007; Kring and Barch, 2014). Finally, based on the established literature on increased impulsivity in BP (Reddy et al., 2014; Fortgang et al., 2016) we hypothesized that BP would also show increased DD relative to HC.