Impaired Expected Value Computations in Schizophrenia Are Associated With a Reduced Ability to Integrate Reward Probability and Magnitude of Recent Outcomes

Abstract

Background: Motivational deficits in people with schizophrenia (PSZ) are associated with an inability to integrate the magnitude and probability of previous outcomes. The mechanisms that underlie probability-magnitude integration deficits, however, are poorly understood. We hypothesized that increased reliance on "value-less" stimulus-response associations, in lieu of expected value (EV)-based learning, could drive probability-magnitude integration deficits in PSZ with motivational deficits. Methods: Healthy volunteers (n=38) and PSZ (n=49) completed a learning paradigm consisting of four stimulus pairs. Reward magnitude (3/2/1/0 points) and probability (90%/80%/20%/10%) determined each stimulus' EV. Following a learning phase, new and familiar stimulus pairings were presented. Participants were asked to select stimuli with the highest reward value. Results: PSZ with high motivational deficits made increasingly less optimal choices as the difference in reward value (probability*magnitude) between two competing stimuli increased. Using a previously-validated computational hybrid model, PSZ relied less on EV ("Q-learning") and more on stimulus-response learning ("actor-critic"), which correlated with SANS motivational deficit severity. PSZ specifically failed to represent reward magnitude, consistent with model demonstrations showing that response tendencies in the actor-critic were preferentially driven by reward probability.

Disruption of Nrxn1α within excitatory forebrain circuits drives value-based dysfunction

“…Interestingly, the Nex-Nrxn1α mutants displayed a clear disruption of these reward variable correlations with p-dSPN activity, specifically for the fast peak immediately preceding trial initiation (Fig.8H). We suggest these data support a hypothesis wherein RPE signals are not appropriately integrated in Nex-Nrxn1α mutants, depriving striatal circuits of essential reward relevant information for subsequent action selection (Hernaus et al, 2019; Hernaus et al, 2018). Further in vivo neural recordings of corticostriatal circuits during this task together with input-specific interrogation of synaptic alterations are required to understand specifically how Nrxn1α mutations perturb communication of value-related information.…”

“…These data are reminiscent of work from schizophrenic patients in a probabilistic reinforcement learning paradigm, where modeling suggested a reduction in the learning rate in patients versus neurotypical controls (Hernaus et al, 2018). Of particular interest, these investigators interpreted alterations in learning rate not to reflect perturbations in the reward prediction error (RPE) signal itself but to changes in how those signals were integrated to update value for future actions (Hernaus et al, 2019; Hernaus et al, 2018). While we are hesitant to directly map parameters of the reinforcement model to neural circuits, this interpretation suggests potentially relevant circuit loci might be those tasked with integrating dopaminergic RPE signals, including connectivity between cortical regions (representing internal/external state) and the input nucleus of the basal ganglia.…”

“…While initial assumptions of blunted hedonic processing in schizophrenia have not been widely substantiated (Gold et al, 2008), other alterations of reward processing have been uncovered, including increased temporal decay of reward value (Ahn et al, 2011; Barch and Dowd, 2010; Heerey et al, 2007; Yu et al, 2017) and impairments in action-outcome learning (Gold et al, 2015; Morris et al, 2015). Disruptions in trial-to-trial feedback-based learning in this disorder may reflect perturbations to reinforcement learning error signals or the manner in which they are integrated to alter choice (Hernaus et al, 2019; Hernaus et al, 2018). Though typically associated with restricted patterns of interest and social reward impairments, recent studies have also revealed reinforcement learning deficits in ASD patients as well (Hill, 2004; Solomon et al, 2015).…”

Grant Report on Social Reward Learning in Schizophrenia

“…Consistent deficits in monetary reward learning are seen in this disorder [25,31,[41][42][43][44], which are related to decreased motivation and pleasure [25,[44][45][46][47][48]. Results from behavioral monetary reward learning studies may inform hypotheses about social reward learning since key reward areas such as the vmPFC and striatum respond similarly to both types of rewards in controls [13][14][15]49] and similar responses to social and non-social stimuli are seen in some behavioral studies in schizophrenia (e.g., ratings of consummatory pleasure) [50,51].…”

“…Impaired monetary reward learning in schizophrenia is driven in part by decreased value of rewards. This is seen in computational modeling studies of reward learning behavior in schizophrenia [27,31,44] as well as in decreased activation of vmPFC and other frontal cortical areas at time of choice in schizophrenia [27,41,42,56], including less frontal activation to positive feedback in patients with greater anhedonia/avolition [41]. In addition, people with schizophrenia fail to employ appropriate win/stay and lose/shift strategies, which reflect difficulty in trial to trial updating of value [27,43,46,57].…”