Jumping to conclusions: A network model predicts schizophrenic patients' performance on a probabilistic reasoning task

Abstract

This article extends computational models of schizophrenia that focus on the negative aspects of this syndrome to behavioral biases that are associated with a positive symptom of schizophrenia, namely delusions. The phenomenon studied is the "jump-to-conclusions" style of reasoning that is characterized by delusional patients-in comparison with controls-whereby they make less-informed decisions when an option to collect more decision-specific information is available. Simulations show that these differences can be mimicked by modulating the gain parameter-associated with variations in dopamine level-in a simple network model.

“Virus and Epidemic”: Causal Knowledge Activates Prediction Error Circuitry

“…This leads to the interesting possibility that neurodegenerative epidemics that are characterized by dysfunction of DA circuitry, such as schizophrenia and Parkinsonʼs epidemic, might be associated with impaired updating of causal knowledge leaving these patients less adaptive in changing environments. Hence, the reported reasoning impairments in schizotypy (Sellen, Oaksford, & Gray, 2005) as well as schizophrenia (Moore & Sellen, 2006), but also prediction error-dependent delusion formation in psychosis (Murray et al, 2008;Corlett et al, 2007), and the recently demonstrated problems of patients with Parkinsonʼs epidemic to learn from prediction errors of rewards (Schott et al, 2007) and their problems in counterfactual reasoning (McNamara, Durso, Brown, & Lynch, 2003) may extend to a more general impairment of modifying existing causal relationships within semantic memory.…”

Jumping to conclusions in schizophrenia

“…These findings are consistent with an aberrant salience account of SZ, 82 whereby dysregulated dopamine transmission generates context-inappropriate salience attributions, potentially due to aberrant signaling in the ventral striatal dopaminergic pathway, which is thought to regulate stimulus–response pairings. 83 Moore and Sellen 84 built a simple network model in which the gain, or signal-to-noise, parameter (which describes the likelihood of a node firing when presented with some input) was varied. The gain parameter was assumed to represent striatal dopaminergic activity, and increasing this parameter meant that the model successfully mimicked the JTC response style seen in research data from delusional patients.…”

Paranoid thinking, suspicion, and risk for aggression: A neurodevelopmental perspective

“…Very little empirical information is available regarding the neural mechanisms of the JTC bias. Moore and Sellen (2006) put forth a neural network model that implicates enhanced or overactive mesolimbic dopamine (DA) system functioning in the JTC bias in delusional patients and that may have applicability to healthy adults as well. The stress-sensitive mesolimbic DA system comprises the ventral striatum and the ventral tegmental area of the midbrain; the ventral tegmentum provides the ventral striatum, a core structure within the affective node of the SIPN, with dopaminergic input (Trainor, 2011).…”

“…Simulations of behavioral differences between delusional individuals and controls during performance of the bead task using Moore and Sellen's (2006) model suggest that in the context of delusions the elevated mesolimbic DA alters representations of stimulus salience, which influences their subsequent processing by regulatory and attentional systems based in the PFC. It is notable that the results of simulations were inconsistent with dysfunction in higher order attentional processes; instead, the authors suggest that the altered salience representations generated via the striatum may differentially influence the allocation of attention.…”