NIH/NIMH – Computationally-Defined Behaviors in Psychiatry (R21 Clinical Trial Optional)

August 13, 2018 by School of Medicine Webmaster

This Funding Opportunity Announcement (FOA) solicits applications for research projects that will apply computational approaches to develop parametrically detailed behavioral assays across mental-health relevant domains of function. These projects should focus on behavior in humans and test computational models in healthy subjects. NIMH is particularly interested in the study of behavioral measures, models, and parameters that have the potential for back-translation from humans to animals, especially for pre-clinical therapeutics development, and/or in models that have the potential to be extended to clinical populations.

Background and Rationale

Understanding the regulation and dysregulation of human behavior requires the ability to investigate questions critical to mental health in both humans and non-human animals. Ideally, tools available for psychiatric and behavioral neuroscience research would include a library of behavioral assays that can be used in both humans and animals to assess mental-health relevant domains of function and to test hypotheses regarding neurobiological mechanisms. Currently, however, significant gaps exist in our ability to move research bi-directionally between humans and animals. Behavioral and cognitive assays designed for screening medications in rodents (e.g., forced swim test, tail-suspension test, sucrose preference) are rarely predictive of human outcomes. Complex paradigms or self-report measures designed for use in humans present obvious barriers for back-translating to animals.  Moreover, current paradigms for measuring human behavior often lack the computational rigor necessary to reliably model the richness and variability critical to studies of mental health and illness. These problems hinder the translational pipeline for understanding pathophysiology of mental illness and developing therapeutic intervention approaches for psychiatric disorders.

To address this gap, NIMH is interested in the development of a new set of computationally-informed behavioral paradigms and/or the deployment of novel computational models to existing paradigms to capture dimensional aspects of mental-health relevant behaviors. A quantitative understanding of the dynamic relationships and constraints of relevant parameters in the physical world has enabled the rapid accumulation of knowledge and impactful technology development in several fields of science. In basic neurobiology, for example, equations describing the dynamics of ion flow in neurons (e.g. Hodgkin-Huxley) have been fundamental to our ability to measure and understand electrophysiological data. Applications of this type of computational approach to behavioral data, however, have been much more limited. One area which demonstrates the potential power of computationally-informed behavioral paradigms is reinforcement-learning theory. This methodology has not been extensively exploited in many other behavioral domains relevant to mental health, and novel models are needed.

This funding opportunity announcement (FOA) encourages the development of theoretically motivated mathematical models able to account for quantitative, parametric behavioral measurements in humans. Human studies should focus on mental-health relevant domains of function; identify a limited set of key parameters that govern specific behavioral dynamics; define mathematical relationships between these parameters that can accurately represent an empirical dataset; and validate the model by predicting behavioral outcomes in a validation dataset. Resulting behavioral mathematical models should be parsimonious, balancing complexity with descriptive ability and predictive utility. These systems will provide high-level objectivity and consistency, enabling a better understanding of behavioral dynamics and more rigorous hypothesis generation and testing of neurobiological mechanisms.

To maximize relevance and potential generalizability, models should first be applied to human behavioral data acquired from samples encompassing a wide range of normative behavioral variance. To strengthen our translational pipeline, future applications of the same models in animals and/or clinical samples should be anticipated.

Research Objectives

The specific goal of this FOA is to support projects in human subjects that will use theory-driven computational models of parametrically detailed behaviors relevant to mental health. Projects responsive to this FOA would focus on:

  • A well-defined question in behavioral science relevant to psychiatric populations.
  • Behavioral paradigms designed to measure dimensional processes linked to a specific domain of function that is affected in psychiatric disorders.
  • Behavioral measures with rigorous psychometric properties (e.g. test-retest reliability).
  • Parametrically-detailed, mental-health relevant behavioral measures that readily lend themselves to rigorous computational analysis, predictions, and explanations.
  • Paradigms and models that have the potential for back-translation from humans to animals to advance novel therapeutic strategies or that can be extended to clinical populations in the future.

Steps to develop such a theoretical model will likely include:

  • Breaking down the human behavior into fine-grained parameters that can be mathematically described.
  • Integrating the human behavioral parameters in an experimentally-grounded mathematical formalism (e.g., a new behavioral theory). The instantiation of the model should be based on the integration of previous experimental findings, should allow the tracking and integration of all parameters, and should predict behavioral outcomes over time.
  • Experimentally validating and optimizing these theory-driven models of human behavior. The mathematical model needs to be experimentally tested, validated, and refined by rigorous determination of the relationships between all the behavioral parameters and the outcome variables. This process includes demonstration that the model can make behaviorally accurate predictions for a new experiment or dataset.

Specific Areas of Research

Examples of research projects that might be submitted under this FOA include, but are not limited to, mathematical formalisms and computational models designed to:

  • Explain and predict human behavioral constructs outlined in the Research Domain Criteria matrix, such as those in the Social Process Domain.
  • Classify and quantify behavioral variance in human subjects.
  • Describe computationally-informed behavioral assay(s) in large, “unselected” human populations encompassing a wide range of behavioral variance (e.g., mTurk).
  • Capture dynamic behavioral measurements and predict changes in behavior over a range of time scales in the same individual.
  •  Explain behaviors specific to a developmental stage.

Applications that are not responsive to this RFA and will not be reviewed include the following:

  • Studies which do not include a well-integrated theoretical and experimental plan to mathematically model behavior.
  • Studies which do not involve human subjects.
  • Studies in which the behavioral components have been extensively modeled (e.g., projects focused on value learning, reinforcement, or reward prediction error).

Mechanistic clinical trials are allowable under this FOA as a method for studying and modeling human behavior acquired from samples encompassing a wide range of normative behavioral variance. Additional information about the definition of mechanistic trials is provided in NOT-MH-18-004. Applications proposing clinical trials to develop or evaluate the efficacy of novel interventions may only be submitted under one of the NIMH clinical trials FOAs and therefore, will not be considered for this FOA.

Potential applicants are encouraged to review NIMH’s research priorities in the field of Computational Psychiatry (, the Theoretical and Computational Neuroscience program (, and Social and Affective Neuroscience Program ( .

Deadlines:  October 20, 2018; October 20, 2019; October 20, 2020 (letters of intent); November 20, 2018; November 20, 2019; November 20, 2020 (full proposals)


Filed Under: Funding Opportunities