Notice of Intent to Publish Funding Opportunity Announcement for Individually Measured Phenotypes to Advance Computational Translation in Mental Health (IMPACT-MH) Mechanism (U01 Clinical Trial Optional)
Notice Number:

Key Dates

Release Date:
March 17, 2023
Estimated Publication Date of Funding Opportunity Announcement:
March 31, 2023
First Estimated Application Due Date:
June 14, 2023
Earliest Estimated Award Date:
April 30, 2024
Earliest Estimated Start Date:
April 30, 2024
Related Announcements

NOT-MH-23-196 - Notice of Intent to Publish Funding Opportunity Announcement for Individually Measured Phenotypes to Advance Computational Translation in Mental Health (IMPACT-MH): Data Coordinating Center (U24 Clinical Trial Not Allowed)

Issued by

National Institute of Mental Health (NIMH)


The National Institute of Mental Health (NIMH) intends to promote a new initiative by publishing a Request for Applications (RFA) to stimulate and support research that will use behavioral measures and computational methods to define novel clinical signatures that can be used for individual-level prediction and clinical decision-making in mental disorders. A multi-component approach is proposed in which grantees will (1) identify and/or develop behavioral tasks (and other types of measures, as appropriate) that are optimized for measurement of individual differences in individuals with or at risk of developing mental disorders; (2) collect the data from novel clinical cohorts and/or identify existing datasets that include behavioral data and other data that are typically available in the clinical record; (3) derive novel clinical signatures that incorporate behavioral measures and information derived from the clinical record, and are informative for clinical purposes; and (4) partner with a Data Coordinating Center (DCC) as described in the companion Notice, NOT-MH-23-196 that will coordinate the harmonization of methods, aggregation of data, analysis of combined data, and creation of a data infrastructure to support data sharing with the scientific community. Applicants may propose new cohorts from one or more populations targeted to specific clinical groupings (e.g., mood/anxiety disorders, disorders of behavior regulation, psychosis) and/or care delivery settings, may leverage data from existing clinical research cohorts that have appropriate data structures or may use a combination of approaches with new and existing data.

This Notice of Intent to Publish (NOITP) is being provided to allow potential applicants sufficient time to develop meaningful collaborations and responsive projects.

The NOFO is expected to be published in Spring 2023, with an expected application due date in Summer 2023.

This NOFO will utilize the U01 activity code. Details of the planned NOFO are provided below.

Research Initiative Details


Current approaches for diagnosing mental disorders lead to a great deal of heterogeneity within diagnostic groups and do not provide a sufficiently precise characterization of individual patients to maximally inform clinical decision-making. Using machine learning and other data-driven approaches to integrate data from behavioral assessment(s) with clinically available data has the potential to generate more precise and objective clinical phenotypes. Consistent with the Research Domain Criteria framework, the National Institute of Mental Health (NIMH) seeks to support studies that will enable data-driven algorithms to generate clinical phenotypes that can optimize evaluation at the level of the individual and which are informative for clinical purposes (e.g., predicting prognosis, enhancing clinical monitoring, and selecting effective treatments). Novel clinical phenotypes, derived from individual participant-level measurements could foster more informative classification schemes, improve clinical practice, and lead to the identification of specific pathology-related mechanisms.

Research Objectives

Identifying and Deploying Measures of Individual Differences

Proposed research projects should focus on integrating data collected via behavioral tasks and measures with data available via clinical records (such as diagnostic codes, demographic data, treatment history, clinician notes, and symptom ratings). Behavioral measures are the focus of this initiative because they are relatively inexpensive to administer, may be delivered remotely and repeatedly, and allow manipulation of task and stimulus characteristics. In addition, performance on some behavioral tasks can prompt hypotheses regarding specific neural circuits and mechanisms related to phenotypes. Behavioral tasks and measures may include computerized tasks administered via web- or device-based platforms and/or passive collection of behavioral data (including speech, mobility, and sleep patterns) via mobile or wearable devices. Behavioral measurements should be selected with scalability in mind, including their potential to be implemented in routine clinical practice.

Behavioral tasks that have been shown to reliably detect differences between groups (e.g., a clinical group versus healthy controls) may not have appropriate psychometric properties to distinguish between individuals for the purposes of informing individual-level predictions. Many tasks have been developed to demonstrate a particular mechanism, and reliability is demonstrated by replication of results across multiple studies (e.g., the differences in accuracy and reaction times between go and no-go trials in an impulse-control task). Precisely because the tasks were developed to replicate consistently the findings across subjects, there may be insufficient variability among individuals or over repeated assessments to be sensitive to individual differences. In addition, there may be measures that have been primarily used in studies of group differences for which insufficient data exist to allow interpretation of individual-level performance (i.e., a lack of normative data against which to compare individual performance). Research activities focused on optimization of existing measures, development of novel measures, and/or collection of normative data for the purpose of identifying novel clinical signatures at the level of the individual would be appropriate under this NOFO. It is also highly encouraged to use or to develop behavioral tasks that yield data that are amenable to analysis using computational methods such as artificial intelligence, machine learning, or Bayesian techniques allowing iterative testing of predicted values associated with various task parameters.

Deriving Novel Clinical Signatures

For the purpose of this NOFO, a clinical signature is considered to be a clinical phenotype comprised of a combination of features derived from behavioral data and the clinical record that can be assessed at the level of the individual and is prospectively associated with a specific clinical and/or functional trajectory, differential response to specific treatment types, and/or biological or psychological mechanisms.

To derive these signatures, analyses should focus on determining the extent to which incrementally adding behavioral data to information that is typically available in the clinical record improves clinically relevant prediction at the level of the individual. For example, it would be possible to quantify the degree to which performance on a reward-related task, when added to predictors derived from the clinical record (such as patterns of clinical encounters and prescription history), improves the prediction of response to a specific treatment. Identification of these signatures could pave the way for the deployment of integrative, tailored assessment tools in treatment settings.

Data types other than behavioral data and clinical records data, such as EEG, MRI, blood-based measures, or genomics, could be used in conjunction with behavioral data to further refine or disambiguate novel clinical phenotypes. For example, early phases of classification might be based on behavioral measures and data from the clinical record; a subsequent phase in a subset of individuals might incorporate a biological measure, such that a novel clinical signature can be derived via integration of data types. These other data types may also be useful for investigating mechanisms that further inform the novel phenotypes. However, the primary goal of this initiative is to develop highly accessible tools, so the use of costly, resource-intensive methods should be considered judiciously.

Both new and existing datasets may be appropriate for the purposes of this NOFO. All proposed datasets should include:

  • data from individuals with mental disorders or at risk of developing mental disorders,
  • behavioral data from either task-based and/or naturalistic/passive assessments,
  • clinical records data,
  • sufficiently large sample size to allow for detection of robust novel phenotypes that can be identified on an individual basis.

A major goal of this NOFO is the development of data structures that are designed from the outset to enable machine learning and related algorithms. Accordingly, databases from cohorts used under this NOFO must incorporate structures that comply with the FAIR data principles (ensuring that data are Findable, Accessible, Interoperable, and Reusable). Applicants proposing the use of extant databases must provide evidence that the data structures are fully consistent with FAIR.

Clinical signatures will likely vary according to clinical population features such as age, prevalence of co-occurring illnesses, rates of health insurance, and environmental factors such as poverty and availability of healthcare services. Efforts to deploy measures of individual differences and derive clinical signatures should take into consideration the population(s) with which the eventual clinical tool(s) will be used and the setting(s) in which services are delivered. NIMH strongly encourages the participation of individuals from diverse backgrounds, including individuals from underrepresented racial and ethnic groups, those with disabilities, those from disadvantaged backgrounds, and women (see NOT-OD-20-031). The proposed research should also describe strategies to recruit and retain research participants, with clear plans to monitor for and mitigate potential biases and attrition in data acquisition and analyses.

Research Team Expertise

The project PD/PI (or multi-PDs/PIs) must be experienced in the successful establishment, coordination, and management of research projects. Projects require multidisciplinary teams with expertise in the following domains: clinical practice in mental health; behavioral assessment; phenotype development and validation; data practices; participant recruitment and retention; and ethical considerations. In line with the Notice of NIH's Interest in Diversity (NOT-OD-20-031), teams should strive for the inclusion of scientists and trainees from diverse backgrounds and life experiences to bring different perspectives, creativity, and individual enterprise to address complex scientific problems. Research teams should likewise have the experience needed to focus on advancing the likelihood that underserved or health disparity populations participate in and benefit from health research, and enhancing public trust.

Partnership with the IMPACT-MH Data Coordinating Center

As discussed in the companion NOT-MH-23-196, the IMPACT-MH Data Coordinating Center (DCC) aims to support the work of the projects mentioned in this NOFO. The DCC will be responsible for facilitating the coordination and communication among the U01 projects, leveraging opportunities for harmonization of methods and measures, building data infrastructure and pipelines, conducting analyses of aggregated datasets, and partnering with U01 sites to monitor and mitigate potential biases in collected data and analytic approaches.

By responding to this NOFO, applicants agree to cooperate with all proposed IMPACT-MH DCC activities, including:

  • Participating as a member of the IMPACT DCC Steering Committee and attending annual All Hands Meetings and quarterly Steering Committee Meetings (virtual)
  • Sharing with the DCC detailed information about planned data collection methods, measures, and procedures
  • Adopting, where appropriate, common measures identified through the IMPACT DCC harmonization process described below
  • Sharing de-identified subject-level data with the IMPACT DCC throughout the course of the project.

Data Harmonization and Monitoring. To maximize the scientific value of the data generated by projects funded under this NOFO, applicants should follow the guidelines set out in the NIH Data Management and Sharing Policy. This involves focusing on FAIR principles as well as addressing biases that may be intrinsic to the data. These biases, described by NIH’s Office of Data Science Strategy, can include bias in measurement accuracy, selection bias of the cohort, and biases introduced through missing data, among others. Applications should include plans to assess their data for these biases and a description of techniques that will be used to address them. U01 projects will coordinate with the DCC to ensure that data is collected and stored in an appropriate format so that it will be able to be used in future studies and to build clinical decision-making tools.

Best practices for data collection will be coordinated through the partnership with the DCC to ensure the longevity and usability of the data. However, each project team should have a local data expert who can monitor the data collection specific to the funded project and act as the liaison to the DCC.

Enhancing Diverse Perspectives

This NOFO requires a Plan for Enhancing Diverse Perspectives (PEDP) as part of the application (see further below). Applicants are strongly encouraged to view the available PEDP guidance material.

NIH recognizes that diverse teams working together and capitalizing on innovative ideas and distinct perspectives outperform homogeneous teams. There are many benefits that flow from a diverse scientific workforce, including fostering scientific innovation, enhancing global competitiveness, contributing to robust learning environments, improving the quality of the research, advancing the likelihood that underserved populations participate in and benefit from research, and enhancing public trust. See Notice of NIH's Interest in Diversity (NOT-OD-20-031).

To support the best science, NIH encourages inclusivity in research. Examples of structures that promote diverse perspectives include, but are not limited to:

  • Transdisciplinary research projects and collaborations among investigators from different fields.
  • Engagement from different types of institutions and organizations (e.g., research-intensive, undergraduate-focused, minority-serving, community-based).
  • Individual applications and partnerships that enhance geographic and regional heterogeneity.
  • Investigators and teams composed of researchers at different career stages.
  • Participation of individuals from diverse backgrounds, including groups historically underrepresented in the biomedical, behavioral, and clinical research workforce (see NOT-OD-20-031), such as underrepresented racial and ethnic groups, those with disabilities, those from disadvantaged backgrounds, and women.
  • Project-based opportunities to enhance the research environment to benefit early- and mid-career investigators.

Applicants must include a Plan for Enhancing Diverse Perspectives (PEDP) submitted as Other Project Information as an attachment (see Section IV). The PEDP will be assessed as part of the scientific and technical peer review evaluation.

Guidance Regarding Clinical Trials

Clinical studies that address the research objectives outlined in this NOFO are encouraged across all areas of mental health, with the exception of clinical trials focused on developing a new intervention and clinical trials to test treatment efficacy and effectiveness. Further information on support of clinical trials at NIMH can be found at:

NIH defines a clinical trial as "A research study in which one or more human subjects are prospectively assigned to one or more interventions (which may include placebo or other control) to evaluate the effects of those interventions on health-related biomedical or behavioral outcomes." (NOT-OD-15-015)

Under this NOFO, NIMH will accept applications that propose studies that meet the definition of a clinical trial if the prospective assignment to intervention and assessment of outcomes serves the purpose of deriving or validating a novel clinical signature.

Examples of Research Interest

The following are examples of the types of studies that would be considered responsive to the NOFO. These are meant to be illustrative only and are not intended to be exclusive or exhaustive.

  • Studies that deliver an intervention with established clinical efficacy and use data from the electronic health record in combination with data from behavioral tasks to develop algorithms to predict treatment response that can be applied at the level of the individual.
  • Studies that focus on tasks in order to increase within-subjects reliability and/or to optimize tasks for computational models, followed by collecting data to test clinical signatures.
  • A large observational cohort study in at-risk individuals who might not necessarily be seeking mental health care but otherwise have a clinical record that might yield data whose trajectories are predictive of eventuating in a mental disorder.
  • Studies that focus on validating a clinical signature across different settings/populations using more than one data collection site.
  • Studies which focus on using existing data to test hypotheses regarding clinical signatures and then collecting new data to validate them.
  • Studies to collect data in a new cohort to identify dimensional phenotypes via machine-learning algorithms (e.g., related but differentiable constructs of reward responsiveness, reward learning, and reward valuation) and then validate the predictive clinical signatures using multiple out-of-sample data splits.

Applicants are strongly encouraged to consult with NIMH program staff when developing plans for an application.

Funding Information
Estimated Total Funding


Expected Number of Awards
Estimated Award Ceiling

Application budgets are limited to no more than $2,500,000 direct costs per year

Primary Assistance Listing Number(s)


Anticipated Eligible Organizations
Public/State Controlled Institution of Higher Education
Private Institution of Higher Education
Nonprofit with 501(c)(3) IRS Status (Other than Institution of Higher Education)
Small Business
For-Profit Organization (Other than Small Business)
State Government
Indian/Native American Tribal Government (Federally Recognized)
County governments
Independent school districts
Public housing authorities/Indian housing authorities
Indian/Native American Tribally Designated Organization (Native American tribal organizations (other than Federally recognized tribal governments)
U.S. Territory or Possession
Indian/Native American Tribal Government (Other than Federally Recognized)
Non-domestic (non-U.S.) Entity (Foreign Organization)
Regional Organization
Eligible Agencies of the Federal Government

Applications are not being solicited at this time.


Please direct all inquiries to:

Jenni Pacheco, Ph.D.
National Institute of Mental Health (NIMH)
Telephone: 301-443-3645