Abnormalities in goal-directed behavior are observed in numerous neuropsychiatric disorders. Deficits in motivation and impulsive and compulsive behaviors are just a few examples. Treating abnormalities in goal-directed behavior requires the identification of specific sub processes of learning and decision making (DM) subject to modification through interventions. Recent efforts in computational cognitive neuroscience have provided algorithmic accounts of reinforcement learning (RL) and DM that have made the identification of such sub processes possible. These algorithmic and computational accounts have the potential to produce quantifiable and sensitive biomarkers for treatment trials across a range of psychiatric disorders.
The goal of this Research Topic is to examine, evaluate, or discuss potential contributions of computational models to the understanding and treatment of neuropsychiatric illness. Computational models have helped to provide algorithmic accounts of RL and DM, while operationalizing difficult-to-observe constructs through parameter estimation. Such constructs include representations of value and certainty of value, prediction error magnitude and valence, learning rates, decision noise, the “stickiness” of choices, and the contribution of certainty of value to decision to exploit familiar contingencies vs explore unfamiliar ones. These constructs provide specific, tractable, and measurable variables that are believed to relate to clinical phenomena such as anhedonia, impulsivity, avolition, maladaptive learning about the value of stimuli and actions, and the formation of unusual beliefs. Contributors are expected to detail how the constructs captured by a given computational model of an RL and/or a DM process can contribute to the clinical symptoms/core pathology of a disorder and/or contribution to interventions for this symptom/disorder.
Contributors are welcome to use the formats of an original research article, systematic review, review, mini-review, and hypothesis and theory. Studies addressing the following themes are strongly encouraged:
• Contributions of abnormalities in RL and DM to motivational deficits in psychotic or mood disorders, or other neuropsychiatric disorders
• Models of the exploration/exploitation trade-off and its possible contribution to psychopathology
• Models of cost/benefit decision making in psychiatric illness, as well as models of learning about effort cost
• Models of the contribution of uncertainty to learning rates over time
• Models of RL mechanisms in learning about states vs. reward values in psychiatric illness
• Models of how working memory representations contribute to mechanisms of RL and/or DM, and thus associated psychopathology, in psychiatric illness
Abnormalities in goal-directed behavior are observed in numerous neuropsychiatric disorders. Deficits in motivation and impulsive and compulsive behaviors are just a few examples. Treating abnormalities in goal-directed behavior requires the identification of specific sub processes of learning and decision making (DM) subject to modification through interventions. Recent efforts in computational cognitive neuroscience have provided algorithmic accounts of reinforcement learning (RL) and DM that have made the identification of such sub processes possible. These algorithmic and computational accounts have the potential to produce quantifiable and sensitive biomarkers for treatment trials across a range of psychiatric disorders.
The goal of this Research Topic is to examine, evaluate, or discuss potential contributions of computational models to the understanding and treatment of neuropsychiatric illness. Computational models have helped to provide algorithmic accounts of RL and DM, while operationalizing difficult-to-observe constructs through parameter estimation. Such constructs include representations of value and certainty of value, prediction error magnitude and valence, learning rates, decision noise, the “stickiness” of choices, and the contribution of certainty of value to decision to exploit familiar contingencies vs explore unfamiliar ones. These constructs provide specific, tractable, and measurable variables that are believed to relate to clinical phenomena such as anhedonia, impulsivity, avolition, maladaptive learning about the value of stimuli and actions, and the formation of unusual beliefs. Contributors are expected to detail how the constructs captured by a given computational model of an RL and/or a DM process can contribute to the clinical symptoms/core pathology of a disorder and/or contribution to interventions for this symptom/disorder.
Contributors are welcome to use the formats of an original research article, systematic review, review, mini-review, and hypothesis and theory. Studies addressing the following themes are strongly encouraged:
• Contributions of abnormalities in RL and DM to motivational deficits in psychotic or mood disorders, or other neuropsychiatric disorders
• Models of the exploration/exploitation trade-off and its possible contribution to psychopathology
• Models of cost/benefit decision making in psychiatric illness, as well as models of learning about effort cost
• Models of the contribution of uncertainty to learning rates over time
• Models of RL mechanisms in learning about states vs. reward values in psychiatric illness
• Models of how working memory representations contribute to mechanisms of RL and/or DM, and thus associated psychopathology, in psychiatric illness
