NAMHC Minutes of the 251st Meeting

September 14, 2017

Department of Health and Human Services
Public Health Service
National Institutes of Health
National Institute of Mental Health

Introduction

The National Advisory Mental Health Council (NAMHC) convened its 251st meeting in open session at 9:00 a.m. on September 14, 2017, at the Neuroscience Center in Rockville, Maryland, and adjourned at approximately 12:45 p.m.  In accordance with Public Law 92-463, the policy session was open to the public.  The NAMHC then reconvened for a closed session to review grant applications at 1:45 p.m. until adjournment at approximately 5 p.m. (See Appendix A:  Review of Applications).  Joshua Gordon, M.D., Ph.D., Director, National Institute of Mental Health (NIMH) presided.

Council Members Present

(Appendix B, Council Roster)

Chairperson

Joshua Gordon, M.D., Ph.D.

Executive Secretary

Jean Noronha, Ph.D.

Council Members

• Rhonda Robinson Beale, M.D.
• David A. Brent, M.D.
• Michael F. Hogan, Ph.D.
• Richard L. Huganir, Ph.D.
• John H. Krystal, M.D.
• Marsha M. Linehan, Ph.D.
• Maria A. Oquendo, M.D., Ph.D.
• Hyong Un, M.D.
• Christopher A. Walsh, M.D.

Ad Hoc Members

• Tami D. Benton, M.D.
• Ian H. Gotlib, Ph.D.
• Alan E. Greenberg, M.D., M.P.H.
• David C. Henderson, M.D.
• Lisa H. Jaycox, Ph.D.
• Gregory A. Miller, Ph.D.

Ex Officio Members

• Steven E. Pflanz, M.D.

Liaison Representatives

• Paolo del Vecchio, M.S.W.

Department of Veterans Affairs

• Amy M. Kilbourne, Ph.D., M.P.H.

National lnstitute of Mental Health Staff

• Shelli Avenevoli, Ph.D.

Others Present at the Open Policy Session

• Lalatendu Acharya, Purdue University
• Sade Akanni, Longevity Consulting
• Bruno Averbeck, Presenter
• Emily Blair, National Alliance on Mental Health
• Sally Bond, Purdue University
• Patricia Brennan, Presenter, Director, National Library of Medicine
• Tabria Dixon, Sign Language Interpreter
• Craig Fisher, American Psychological Association
• Karen Gibson-Serrette, Longevity Consulting
• Laura Guziak, Science Writer
• Michael Knop, NIH Transcriber
• Caroline Neely, George Mason University
• Ciara Morales, Sign Language Interpreter
• Regina Piscitelli, TARA 4 Borderline Personality Disorder
• Valerie Porr, TARA 4 Borderline Personality Disorder
• Lisa Rosenberg, TARA 4 Borderline Personality Disorder
• Cordelia Running, Purdue University
• Tracy Waldeck, Association for Psychological Science
• A.J. Walker, National Assoc. of State Mental Health Program Directors

Open Policy Session Call to Order and Opening Remarks

Dr. Gordon welcomed everyone to the open policy session and extended his thoughts to colleagues in Texas and Florida, including Dr. Randy Blakely, who were unable to attend as a result of the devastating weather events.

NIH Clinical Trials Update

Mike Lauer, M.D., Deputy Director for Extramural Research, NIH

Dr. Mike Lauer began his presentation by discussing a paper that was published in the British Medical Journal in January of 2012 entitled Publication of NIH funded trials registered in ClinicalTrials.gov:  cross sectional analysis  by Yale University and National Library of Medicine (NLM) investigators. Of 635 NIH-funded trials, only 46 percent published their main results within two and a half years. The FDA Amendments Act  (FDAAA) sought to address the problem of non-reporting of trials.

While serving as Director of the Division of Prevention and Population Science at the National Heart, Lung, and Blood Institute (NHLBI), Dr. Lauer convened a team to review 244 cardiovascular trials funded and completed by NHLBI over a period of ten years. In November 2013, the research team presented their findings in the New England Journal of Medicine. Two-thirds of those clinical trials focused on clinical endpoints (e.g., premature death, myocardial infarction, stroke, an unwanted hospitalization) published their main results after a year. After two years, all of them had their results published. However, they represented a minority of the trials NHLBI funded. Most of the funded trials did not focus on clinical endpoints. Of those trials, only 12 percent published their main results within a year. By 30 months after trial completion, only 45 percent of these studies had published their results. They had essentially replicated the 2012 Yale study results.

In 2016, the Yale University researchers published a follow-up study entitled Publication and reporting of clinical trial results:  cross sectional analysis across academic medical centers  to evaluate the number of clinical trials completed between 2007 and 2010 at 51 major academic medical centers in the United States that reported their main results within two years. The academic medical center with the largest percentage of published results was the University of Minnesota. Only 55 percent of their study results were published. Many of the prestigious academic medical centers that receive significant NIH funding only published 20 to 30 percent of their study results.

The senior author and Yale Professor Harlan Krumholz posted a column on the NPR website with the headline Academic Medical Centers Get An F In Sharing Research Results . He wrote, "We have a bottleneck at our nation's bastions of research excellence. Too many times, study results are neither reported on the government website, ClinicalTrials.gov , nor published in a journal."

The Government Accountability Office (GAO) published a report in 2016 entitled Additional Data Would Enhance the Stewardship of Clinical Trials across the Agency . The GAO indicated,
“NIH’s [Office of the Director] reviews some data on clinical trial activity across NIH but has not finalized what additional data it needs or established a process for using these data to enhance its stewardship. NIH is limited in its ability to make data-driven decisions regarding the use of its roughly $3 billion annual investment in clinical trials.” 

Dr. Lauer stated that the studies highlight issues in both input and output. NIH is not collecting an adequate amount of metadata and results of clinical trials are either not reported at all or in a timely way. NIH has had many responses to these data reporting problems over the years.

In October 2014, NIH issued a revised definition of a clinical trial in parallel with the revised common rule. Drs. Francis Collins and Kathy Hudson prepared a paper for the Journal of the American Medical Association, entitled Sharing and Reporting the Results of Clinical Trials  expressing their concern about the fact that trial results were not being reported and that the NIH was proposing a rule to require NIH funded investigators to register their trials and report their results. The proposed rule was released for public comment in November of 2014. The result was the Clinical Trials Registration and Results Information Submission , released in September 2016 which stressed that scientists have an ethical obligation to report results of research involving human beings.

Some of the concerns raised in the 240 comments received from individuals, institutions, and professional societies were that the policy was overly inclusive or not inclusive enough." The NIH response noted, in part, that “The benefits of transparency and the need to fulfill ethical obligations to participants are as relevant to these types of trials as to any other type.” NIH also addressed concerns that a 12-month deadline to report results on ClinicalTrials.gov was to early: “We believe that 12 months represents an appropriate balance between investigators’ interests and the interests of the public in having access to the results of a publicly funded trial." Dr. Lauer clarified that the policy does not require publishing of results in a journal and journals have stated that they do not consider posting of main results on ClinicalTrials.gov to constitute prior publication. 

To enable systems change, Dr. Lauer noted that the clinical trial process should develop from an idea to application submission, application review, funding, IRB (Institutional Review Board) review, U.S. Food and Drug Administration (FDA) review if regulated by FDA, enrollment and data collection, and reporting of results. In order to obtain report results in a timely way, NIH must be aware that the trial exists, which requires registering. The registration must be linked to a grant via a minimal amount of metadata. Funding opportunity announcements (FOAs) will designate future registration requirements. 

Dr. Lauer noted that clinical trial requirements for grants and contracts are provided on the NIH website. Clinical trials are defined as research studies in which one or more 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. In order to build in accountability, the Department of Health and Human Services will regulate the registration and summary results reporting such that the NIH will withhold clinical trial funding if the Agency is unable to verify adequate registration and results reporting. If an institution completes a trial but does not report the results within a year of completion, this may affect the Agency's future funding decisions for that institution. NIH has several mechanisms in place to avoid this consequence.

Discussion

Marsha Linehan, Ph.D. suggested that NIH develop a new workgroup to evaluate outcomes of NIH funded research. Dr. Lauer replied that some NIH Institutes and Centers (ICs) have hired evaluation specialists and economists to evaluate research outcomes. John Krystal, M.D. expressed his concern about the loss of clinician investigators as fewer and fewer psychiatrists are going into clinical research. Current clinician investigators are already overburdened conducting clinical and administrative work on top of research and teaching. Dr. Krystal asked whether NIH had considered how to help clinician investigators manage the additional registration and reporting burdens. Dr. Krystal also noted that some clinical trials test biological hypotheses. In these cases the clinical trial is actually incidental to the biological question. If investigators are measuring some clinically-relevant outcome, the trial is no longer a research study but has to be funded through a single site investigator-initiated clinical trial grant, which imposes an additional complicated framework upon the clinical trial. If NIH is going to use the 2014 revised clinical trial definition, Dr. Krystal believes they will need a different definition for funding. 

Dr. Gordon responded that NIMH will have R01 and hopefully Exploratory/Developmental Research Grant (R21) mechanisms to address those clinical trials that fall out of the scope of the experimental therapeutics clinical trials initiative. NIMH is working to get the wording right, which is a challenge. The NIH Office of Extramural Research (OER) has developed clinical trials FOA that will include some restrictions. NIMH does not want imaging or behavioral studies to fall under the experimental therapeutics platform.  

Dr. Lauer added that NIH is developing a parent R01 that allows for clinical trials. It will require a brief synopsis with field elements that correspond to GAO requirements in addition to the standard 12-page write-up. Another feature is additional space to describe research strategy for peer reviewers. Applications to the parent R01 clinical trial FOA will continue to be distributed to study sections with appropriate expertise by the Center for Scientific Review (CSR).

Dr. Lauer stated that study registration will only require the methods section as written in the grant. This information will be pulled out of the grant by ClinicialTrials.gov to populate required registration fields. Likewise, describing main study results should not be a burden as it is part of the scientific process and should be the same data compiled for publication in a scientific journal.  Only the minority of trials that are regulated by the FDA would require compiling FDA folders.  Dr. Lauer noted that NIH is open to university grant application systems that feed into NIH grant systems, meet the needs and interests of scientists in different fields, and fulfill NIH data requirements.   

Dr. Lauer mentioned that efforts are also underway to make ClinicialTrials.gov more user friendly. Lisa Jaycox, Ph.D. agreed that it is tough to navigate the website with a psychosocial or field trial and expressed her concern regarding how the website would cover the new expanded definitions. Dr. Jaycox also noted that if study results are published on the website first, they will not be peer reviewed. Journals often reject papers with negative or incremental findings. She suggested that NIH play a role in peer review of these types of studies. 

Alan Greenberg, M.D., M.P.H. questioned what proportion of the unpublished clinical trials had been submitted in those cases in which NIH experienced trouble finding published results. He also asked whether there was a difference in publication rates in terms of statistically significant primary outcomes. Dr. Lauer replied that if they could conduct the review again, they would have systematically asked for the reasons for delay of the publishing of results. The research team did ask investigators whether they had published their papers and for copies of the publication. For those that had not published, they asked for the main result and whether it met their hypothesis. Investigators noted that in many cases their papers were rejected by journals.  However, there was also a wide range of other reasons for not publishing their results such as staff changes or lack of time. 

Dr. Lauer “graded” the papers in terms of whether they obtained a positive or negative result.  For those that were not published, he considered the progress and final reports. He discovered that positive results were published more quickly. However, among the 40 most highly cited trials, 35 of them reported out negative results. Hence, some of the most spectacular trials were negative trials. Dr. Lauer’s team was unable to find results for as many as 10 percent of the trials. For ten of these trials, the team confirmed that the investigators had not completed their analysis and did not have plans to. 

David Brent, M.D. asked whether there was some way to determine what is preventing so many investigators from reporting their study results. Dr. Lauer commented that this problem has been studied in the literature for a long time and there is no expectation among researchers that they have to report their results. Dr. Lauer stated that the new NIH policy will set up a requirement that results be disseminated or that funding would be blocked. He applauded the movement to make science more open and transparent. 

Christopher Walsh, M.D. discussed a genomic study being conducted by his colleagues on families in the Neonatal Intensive Care Unit (NICU). Their exploratory science project was classified as a clinical trial and required FDA approval. They spent the first two years of the grant doing nothing but paperwork and did not enroll a single child until well into the second year of the study. Dr. Walsh noted the new policy will require considerably more paperwork for trials that can be very burdensome. He believes NIH should use their limited resources in the most efficient possible way and determine whether the extra paperwork will be really useful or not. Dr. Walsh is also concerned that the new definitions of clinical trials may create a bias towards large institutions that are better able to manage the paperwork burden. He suggested the definition of a clinical trial be restricted to large trials to achieve an 80 percent increase in transparency. Dr. Walsh also stated that researchers are reluctant to report negative results because it tends to create a misapprehension that they have discovered something when they are not necessarily sure that they have. It is important to distinguish a negative result from the absence of an interpretable positive result. 

Dr. Lauer responded that the clinical trial policy statement discusses the decision to define a wide range of types of studies as clinical trials because the ethical obligation to the participant is the same regardless of whether results are positive or negative. He stated that the genomic study Dr. Walsh described should not have been defined as a clinical trial. NIH funds genomic studies that are not FDA regulated. Dr. Lauer reiterated that the new policy does not require that the results be published in a journal, only that they are reported on ClinicalTrials.gov within a year of completion. 

Rhonda Robinson Beale, M.D. applauded the direction NIH was taking but remarked that it will require a change in the research culture. It is important for methodologies to be recorded so that studies are repeatable, and the public and other researchers can learn from them. As there are content limits for data published in journals, a lot of study data can be lost. Autism is a good model for collaboration. 

Richard Huganir, Ph.D. agreed with Dr. Linehan that individual investigators should be penalized for failure to report data, not the universities. He added that the Society for Neuroscience has received an incredible number of emails with questions regarding the new clinical trial definition. Not all human-oriented research will fall under the scope of a clinical trial. NIH needs to clarify what is considered an intervention. He also noted that the administrative impact on young investigators with limited resources will be high.  

Dr. Gordon stated that NIH anticipates developing specific examples of what is and is not considered a clinical trial, particularly in the behavior and neuroimaging fields. He advises those conducting human research to submit applications for the February cycle up to a month early to give OER enough time to categorize studies. Dr. Gordon commented that the administrative burden should not extend beyond registering trials, specifying one expected outcome from the trial on ClinicalTrials.gov, and entering results. The results could be inconclusive. He encouraged investigators to share their stories of any unintended consequences of the new policy on their administrative work load. 

NIMH Director's Report

Joshua Gordon; M.D., Ph.D. Director, NIMH

Dr. Gordon began by thanking Drs. Brent, Huganir, Linehan, Maria Oquendo, and Hyong Un for extending their NAMHC appointments beyond their September 30th term. HHS has now approved the hiring of new special government employees so those members who have been serving unofficially for the past year should be appointed by the January NAMHC meeting. He welcomed pending U.S. Department of Defense representative Colonel Steven Pflanz to the Council. Colonel Pflanz serves as the Air Force Director of Psychological Health and the Mental Health Branch Chief in the Air Force Medical Support Agency. He is a board-certified psychiatrist and flight surgeon that previously served as Commander of Combat Operational Stress Control in Afghanistan and Chief of the Air Force Suicide Prevention Program. The Council then unanimously passed a motion to approve the Minutes of the 249th NAMHC meeting held on May 25, 2017.    

Dr. Gordon stated that his one-year anniversary as Director of NIMH will occur on September 15, 2017 and provided an update on legislation, budget, leadership, policy, research, and the portfolio activities at NIMH over his past year. On September 8, 2017, Congress approved a disaster relief package (HR 601) that included $15.25 billion in emergency spending for hurricane relief, a suspension of the debt limit through December 8, 2017, and a continuing resolution on the budget through December 8, 2017. The stop-gap funding measure includes a prohibition on reducing the indirect funding rates for NIH. Both the House and Senate HHS budget committees have passed bills that include a $2 billion increase for NIH above FY 2017 level in the Senate bill and a $1.1 billion increase for NIH above the FY 2017 level in the House bill. There are generous increases for the Brain Initiative  and increases that are slightly above inflation for the rest of NIH. If the Senate version is eventually passed, it would result in three consecutive years of $2 billion increases for the NIH.

The NIMH budget increased in 2017. Over the past year there has been a slight dip in grant funding rates but over 500 grants have been funded. The average award size increased and NIMH was able to fund several young investigators. The continuing resolution has affected NIMH’s ability to reach applications as far in 2017. If the final budget aligns with the Senate appropriation, it will make for another relatively good year.

Dr. Gordon testified at both the House and Senate appropriations hearings along with several other IC directors. He also met individually with some members of Congress, notably an NIH visit by the Senate Appropriations Committee members on June 5, 2017, an NIH visit by Senator Jerry Moran (R-KS), and a district visit with Senator Bill Cassidy (R-LA) to Louisiana State University Health Sciences Center. 

The annual Outreach Partnership Program meeting was held in July and brought together over 65 outreach partners, national partners, and federal agencies to share their education and outreach activities with a focus on suicide prevention. The Professional Coalition for Research Progress meeting was held in May. On September 15, 2017, the Alliance for Research Progress will hold their annual meeting. 

Dr. Elinore McCance-Katz, Director of the Substance Abuse and Mental Health Services Administration (SAMHSA), has been appointed to serve as the Assistant Secretary for Mental Health and Substance Abuse.

This is a new position that was created by the 21st Century Cures Act. Dr. McCance-Katz is very interested in continuing SAMHSA's collaborations with NIMH.  Paolo del Vecchio is the Director of the Center for Mental Health at SAMHSA. Dr. Gordon is working with him on several new collaborations.   

Dr. Jerome Adams was sworn in as the 20th Surgeon General of the United States. President Trump selected Dr. Francis Collins to continue as Director of the NIH. This did not require Senate approval since he had been approved previously. The President also appointed Dr. Norman "Ned" Sharpless to be the new Director of the National Cancer Institute. In October 2017, Dr. Josie Briggs will retire from her position as Director of the National Center for Complementary and Integrative Health (NCCIH).      

At the NIMH, Pamela Collins transitioned from her position as Director of the Office of Research on Disparities and Global Mental Health (ORDGMH) to return to academia at the University of Washington. Dr. Andrea Beckel-Mitchener is currently serving as the Acting Director of ORDGMH. Dr. Gordon acknowledged the efforts of Dr. Beverly Pringle, Deputy Director of ORDGMH, for leading excellent programs in global mental health. Dr. Judith Rapoport retired from the Division of Intramural Research Programs faculty in July after 41 years of service. She pioneered research in several areas of child psychiatry, most notably in-depth and longitudinal studies of childhood-onset schizophrenia and will be dearly missed.

Dr. Gordon announced the passing of colleagues Drs. Howard Eichenbaum and Larry Seidman. Dr. Eichenbaum was the Director of Boston University's Center for Memory and Brain and served on the NAMHC from 2009 to 2012. Dr. Larry Seidman was a long-time NIMH grantee and clinical psychologist at Harvard Medical School and the Massachusetts Mental Health Center/Beth Israel Deaconess Medical Center. Dr. Seidman was known for his influential work on neuropsychological, genetic, and perinatal factors underlying schizophrenia and co-occurring disorders. 

Dr. Gordon discussed changes to NIMH Exploratory/Developmental Research Grant Award (R21) applications (Notice ). NIMH has decided to stop accepting applications in response to the NIH Parent R21 FOA as NIMH was starting to be assigned R21 applications that were not in the institute’s domain because the other institutes had already exited it. NIMH will issue new NIMH-specific R21 FOAs in the coming months. They will continue to accept R21 applications under currently active R21 FOAs for which NIMH is the lead or secondary institute. Dr. Gordon remarked that the R21 program is valuable and encouraged young scientists to apply. 

The 21st Century Cures Act  instructs NIH to promote policies that encourage researchers to gain earlier independence. On June 8, 2017 NIH launched the Next Generation Researchers Initiative  to enhance stewardship of research dollars and strengthen the biomedical research workforce.    Initially, NIMH had considered capping the number of grants for well-funded investigators. This idea was abandoned in favor of specific programs to increase the number of awards for young investigators and early-career investigators with an R01 at risk of losing funding. NIMH will place more emphasis on the special council review applications. There is strong sentiment that funding for individuals identified as young investigators or early-career investigators at risk should not harm other early-career investigators in their attempts to get funded. 

A single institutional review board (IRB) policy  for multi-site clinical research goes into effect on January 25, 2018. It is intended to streamline the process by eliminating repetition of IRB reviews across sites. Dr. Gordon encouraged investigators to report on any unintended consequences of this policy. It appears to be working well where it has already been implemented.

Dr. Gordon highlighted some research findings Dr. Jubao Duan's team published Open Chromatin Profiling in hiPSC-Derived Neurons Prioritizes Functional Noncoding Psychiatric Risk Variants and Highlights Neurodevelopmental Loci  in Cell Stem Cell in September 2017.  There has been an increase in the number and complexity of studies using induced pluripotent stem cell lines to study the biology underlying psychiatric
disorders. At the same time there has been an explosion in terms of the number of genetic loci that are linked to disorders like schizophrenia. Dr. Duan’s study attempts to link the biology to the loci. The investigators examined the genome of human iPSC-derived neurons for open chromatin regions. DNA is relaxed in open chromatin regions and can be more easily transcribed into RNA to build proteins.  The technology ATAC-seq identifies these regions by sequencing sections of DNA in open chromatin regions.  Loci that are implicated from the schizophrenia genome-wide association studies (GWAS) can then be statistically associated with these open chromatin regions. The idea is that the loci associated with these open chromatin regions might be the loci that influence gene expression. The investigators also looked for sequences within the open chromatin regions that were known sequences for transcription factor binding. 

Using statistical techniques, Dr. Duan’s team discovered that the loci in the genome associated with schizophrenia are overrepresented in these open chromatin regions. Using transcription factor binding sequences, they also found 100 putatively functional single nucleotide polymorphisms (SNPs) in neuronal open chromatin regions that potentially affect transcription factor binding associated with schizophrenia. The finding suggests that one way these gene variants introduce risk for schizophrenia is by affecting the ability of these open chromatin regions to be transcribed. To test that hypothesis, the study team analyzed a particular gene locus, microRNA-137. The one locus in the genome that has been associated with schizophrenia is near microRNA-137. MicroRNA-137 has at least two different open chromatin regions that contain transcription factor binding sites and risk variants. The study team took that site and a risk site and used CRISPR, a gene editing technique, to change that variant in that human neuron into the non-risk variant to evaluate the difference in the neurobiology between human neurons that carry the risk variants in this open chromatin region and the non-risk variants. They found differences in dendritic complexity as well as differences in the expression of glutamate receptors. 

The results do not presume that this particular risk variant is very important for schizophrenia or that it works through changes in neuronal morphology or an expression of glutamate receptors. The study concludes that there is a high chance that a majority of these genes act to regulate transcription in these open chromatin regions. They were able to show that at least one of the SNPs does affect the transcription of genes and neuronal structure. The study is a good example of the use of human neuron technology and the need for future research on genes in open chromatin regions. The authors highlight the value of conducting these studies in live human neurons versus post-mortem tissues.    

Dr. Gordon then highlighted a study conducted by Dr. Amit Etkin and a large team entitled PTSD Psychotherapy Outcome Predicted by Brain Activation During Emotional Reactivity and Regulation  published in the American Journal of Psychiatry in July 2017. Etkin and colleagues used a common emotional reactivity task that identifies brain regions active when shown emotional faces. Individuals with post-traumatic stress disorder (PTSD) tended to activate the dorsal lateral prefrontal cortex and the amygdala when exposed to these faces. The investigators then treated these individuals with prolonged evidence-based psychotherapy for PTSD. 

The study found that those individuals with higher activation in the dorsal lateral prefrontal cortex had a larger clinical response to the psychotherapy than those with low levels of activity in this region. The opposite was true in those individuals with the highest amygdala reactivity, in that their clinical responses to psychotherapy were worse. Using concurrent TMS-fMRI (transcranial magnetic stimulation - functional magnetic resonance imaging) they demonstrated that by stimulating the dorsal lateral prefrontal cortex, activity in the amygdala was suppressed. Those individuals with the greatest suppression of activity when stimulating the dorsal lateral prefrontal cortex had the best response to therapy. These data suggest that people who are likely to respond to psychotherapy have strongly active dorsal lateral prefrontal cortex control over emotional responses. Dr. Gordon noted that it would be helpful to determine how to sort patients according to their potential response to therapy, such as with emotional reactivity tests. 

Dr. Gordon then briefly discussed the Emergency Department Safety Assessment and Follow-up Evaluation (ED-SAFE)  suicide prevention study conducted by Dr. Edwin Boudreaux and colleagues. The study aims to determine whether an ED-initiated intervention reduces subsequent suicidal behavior. The study team recently published results demonstrating that if those individuals who screen positive are administered a simple intervention of a series of phone calls from Master’s level clinicians over the ensuing year, suicide attempt rates are reduced. 

Dr. Gordon discussed principles of excellent science supported by NIMH. He defined good science as science with strong study designs, high standards of rigor, and most importantly, potential impact in patient treatment. He stressed that NIMH needs to consider diversity of subject matter, workforce, research participants, and timeframes. Research should help patients in both the short-term as well as in medium- and long-term timeframes. NIH has a peer review system that works well.  However, across subject areas, the peer review system cannot determine whether a study on the implementation of a suicide intervention is going to be better science than a grant studying the structure of dopamine receptors, for example. In the future, NIMH is planning to explore metrics to help measure potential impact. In the meantime, NIMH needs to actively balance its portfolio.  

The first science standard is rigor and it involves multiple complementary approaches, comprehensive assessments and outcomes. Studies need to be well-controlled and well-powered and should at least have the potential to be validated and replicable. Also important is a track record for publication or funding, information technology infrastructure, hiring appropriate personnel, and recruiting patients. When considering rigor, NIMH asks whether the study will be definitive if the hypothesis is supported. 

In terms of impact, NIMH is interested in public health significance, expanding the knowledge base, developing novel concepts, and the generation of novel tools, resources, and approaches.    NIMH is focused on moving concepts and ideas along the translational pipeline to not just study the basic science of a particular process, but rather to figure out how the science is going to impact patient care in the long run. When considering impact, NIMH asks how the study will change the field. 

Innovation is about making new connections between different areas of science. It could be high-risk and high-reward. Early stage investigators are often sources of creativity that lead to innovation and NIMH supports new or disruptive ideas that add value beyond just being different. When evaluating a grant for innovation, NIMH asks how the study will shake things up in the field. 

Investigators should present with a diversity of ideas, perspectives, and approaches. Team science and cross-pollination are really important. NIMH considers what the investigators add to the field and how they are helping the field when deciding which investigators to support.    

Dr. Gordon added that some considerations are more field-specific. Basic science should expand knowledge and novel concepts. In translational science, there should be effect sizes beyond current treatments. In addition, clinical problems that represent significant burdens to society should be addressed. The interventions should result in genuine improvements in care, be feasible on a community level, and ensure stakeholder engagement.

Other elements include the balance of short-, medium-, and long-term outcome studies in the portfolio and studies that explore gaps in knowledge. When evaluating investigators, NIMH expects early-stage investigators to be cutting edge. Middle-stage investigators should be productive with notable advances. Senior investigators should show continued creativity and generativity. They should be building the field while they are building knowledge within the field. Dr. Gordon concluded that these principles of excellent science do not supplant the review system, but rather help NIMH in its endeavor to ensure a balanced portfolio in the context of the grant session reviews. 

Discussion

Dr. Krystal commented that he applauded the values Dr. Gordon described. He suggested inviting Dr. McCance-Katz to meet with NAMHC and discuss opportunities for NIMH representation at HHS and also its synergy with SAMHSA. Maria Oquendo, M.D., Ph.D. noted that the NIMH has new definitions for low or high priority program grants. She asked if it were a consequence of policy changes at NIH. Dr. Gordon replied that changes in what is considered low program priority is partly a consequence of rebalancing the portfolio as well as the Next Generation Researchers Initiative. NIMH is adhering to the NIH definitions for program priority levels and tries to fund grants at the 10th percentile level or better unless they mark a grant as low program priority. Grants out of the purview of NIMH are designated as low program priority. 

Amy Kilbourne, Ph.D., M.P.H. mentioned that the National Academies of Medicine and the Clinical and Translational Science Award (CTSA) program are trying to redefine their definition of impact beyond papers and grants. She asked if NIMH has been involved in those conversations and if there were unique measures of impact they could borrow from other agencies. Dr. Gordon remarked that there is a lot of interest in developing metrics that include elements beyond papers, citations, and patents.  

Marsha Linehan, Ph.D. commented that she agrees with the importance of supporting the next generation of scientists as she benefitted from NIMH’s guidance in the development of her own career. She asked what advice she could give to her current graduate students to get involved with NIMH and SAMSHA. Dr. Gordon stated that NIH has an abundance of resources for young investigators and a great training team at NIMH to refer to. 

Dr. Robinson Beale hoped that NIMH arrived at some means of measuring impact, particularly as it relates to public health. Researchers are turning towards more integrated approaches. She also remarked on the importance of expanding the knowledge base and translating outcomes into concepts that can be adopted and implemented in the community in general. Dr. Gordon encouraged her to read his Director's Message when it is released for more information on these topics.   

Guest Speaker

Patricia Brennan, R.N., Ph.D. Director, National Library of Medicine

Dr. Gordon introduced Dr. Patricia Brennan, Director of the NLM. Dr. Brennan has been a pioneer in the development of information systems for patients and in data science more generally. She is also the NIMH Interim Associate Director for Data Science. 

Dr. Brennan stated that NIH believes that through the effective use and reuse of data collected in the course of clinical studies and everyday living, there is a chance to build a new path of discovery that complements traditional approaches of epidemiology, experimental studies, and clinician experiences. All of NIH is engaged in data science and developing effective data management and new methodologies. Data must be findable, accessible, interoperable, and reusable. These principles were advanced by the Future of Research Communications and e-Scholarship (FORCE11) , an internationally-accepted model for reusing study data by planning for a future that allows data to be accessed in a way that preserves protective rights that were engendered when the data were first collected, and allows data sets to be linked together into new analytical models. Some of the solutions will be IC specific. In 2015, the advisory committee to the director stated that the NLM should become, "the epicenter of data science." 

Dr. Brennan played a video about the NLM. It is the largest medical library in the world with PubMed databases available worldwide. The NLM existed prior to incorporation with NIH. It dates its origins to 1836 or 1837 and grew out of the Army Medical Library. The building was built to withstand a nuclear blast with
classic cold war structures. For its first 100 years, the NLM was focused on creating indices and enumerating the medical literature. The red Index Medicus was digitized and by the 1980s the NLM engaged with its stakeholders, including scientists, clinicians, and laypeople, through the internet. Beginning in 2006, the NLM established a data-driven model where data becomes the substrate of discovery, much as the literature has. The NLM is not abandoning the literature, but preparing for a future on the cloud. 

Dr. Brennan stated that the NLM has a fundamental commitment to new curation strategies and data science and methods development. Their goals are to preserve literature and large data sets in a purposeful manner. In addition, they will be expanding their use of standards to support interoperability and allow for effective indexing and curation of data from basic biological as well as anatomical and image-type resources.

Dr. Brennan noted that the NLM is attempting to determine what defines high-value data sets.  Protecting data requires ensuring that the technological infrastructure, including the operating systems, remains aligned with the data resources. Each time an information system is upgraded, the data storage infrastructure must be upgraded as well. 

In addition, new tools for discovery and analysis are needed. The NLM cannot store all of the data generated by NIH or the rest of the research world. Data sets will continue to grow and be stored globally by institutions, research centers, and industry. The role of the NLM is to provide the tools to discover the location of those data resources and assist with data collection and integration. Dr. Brennan remarked that the NLM is committed to global and open data science solutions. The Global Alliance for Genomics and Health  (GA4GH), as well as many of the resources in Europe such as Elixir , are making progress in developing strategies to identify and make data sets available on a wide scale. 

Dr. Brennan commented that there are three workforce targets they are addressing. The first is to develop the next generation of data scientists. Data scientists are currently in high demand in academia and at the NIH. This presents a challenge when industry salaries for data scientists are much higher. Second, there is a need to create data-savvy clinicians. Clinicians must be educated to understand when data science provides evidence for practice and guidance for care and when it does not. The third workforce target is data-savvy librarians. Librarians have to help scientists become compliant with data requirements, locate data sets, and prepare data management plans.

The NLM is launching a Data Commons Pilot to test options for storing, accessing, and sharing biomedical data and associated tools on the cloud. Through the support of the Common Fund under the direction of Vivien Bonazzi, three high-priority data sets were identified: (1) Trans-Omics for Precision Medicine (TOPMed), a cardiovascular data set; (2) Genotype-Tissue Expression (GTEx), a genome to tissue data set; and (3) model organisms data sets. The NLM is working closely with associated investigators to identify cloud storage opportunities, analytics applications, and ethics, policy, and social issues related to data integration and data sharing of these data sets.

Beginning in October 2017, it will be possible to deposit data sets attached to articles in PubMed Central, the NLM’s full-text data repository. This will allow NIH supported investigators to comply with NIH’s public access requirement. Dr. Brennan noted that the NLM has another resource repository, ClinicalTrials.gov. There is a disclaimer on the top of the webpage indicating that registry in ClinicalTrials.gov does not reflect endorsement by NIH. Only 38 percent of the trials reported in ClinicalTrials.gov are exclusively United States studies. 

The ClinicalTrials.gov webpage has two separate interfaces, one for consumers and one for investigators to deposit data. It was recently updated so it can be viewed on multiple devices. There are also new search options and options for displaying the results. Studies can be searched by age group, status, and geographic location.

Dr. Brennan defined ClinicalTrials.gov as a clinical trial reporting system. Protocols and consent forms can now be deposited into ClinicalTrials.gov so that consumers can understand the full depth of studies. 

The NLM is exploring how individual participant data could be shared. The National Academies has encouraged sharing this data if properly protected to allow for integration across studies and  improve upon the ability to learn from studies that may have been underpowered. The NIMH data archives has 30 total trials with controlled access to individual participant data.

Dr. Brennan envisions ClinicalTrials.gov to become a nexus of access to information such as journal publications, results, conference abstracts, protocols, permissions, individual participant data, and other study documents. They plan to add links to public descriptions of studies. She noted that ClinicalTrials.gov is a complement to PubMed, not a replacement. 

The Board of Regents of the NLM just approved their 2017 strategic plan. Over the past year, they have conducted four panels on four different themes: (1) Advancing biomedical discovery and translation; (2) Advancing data science, open science, and informatics; (3) Supporting the public's health; and (4) Building 21st century collections for discovery and health. Some of the key themes include research needs, research funding, infrastructure, standards, workforce development, partnerships, user communities, user engagement, educational outreach, international engagement, and health disparities. 

The panels concluded that the library that served the NIH in 1995 must change to support data-driven discovery and data-powered health in the community in 2025. The NLM collections need to go digital, but preserve history. Training is needed for librarians, clinicians, and researchers.    They also agreed that the most precious resource at the NLM is public trust. Over the next 10 years, the NLM will serve society and advance health by providing: (1) integrated information for discovery and health in the 21st century; (2) pathways for dissemination and engagement; and (3) workforce excellence for accelerated discovery and better health. Dr. Brennan encouraged the participants to read her blogs and contact her for additional information on the NLM. 

Discussion

Dr. Linehan suggested that the NLM require investigators to provide treatment use information on ClinicalTrials.gov. Dr. Brennan replied that they are exploring ways to link the NLM resources to clinical records. One potential option is to embed the formal way a treatment is delivered into the electronic health record (HER) so that it becomes the driver for care decisions and not simply a reference in someone's mind.

Dr. Robinson Beale questioned whether big data sets would have feedback loops linking to NIMH to cross-reference populations across several different studies. Dr. Brennan replied that the Common Data Elements (CDE ) repository is an NIH initiative designed to ensure cross-mapping of key data elements considered essential in individual studies. The NLM has been developing this repository with the National Institute of Nursing Research (NINR) and the National Institute of Neurological Disorders and Stroke (NINDS). 

Dr. Robinson Beale also wondered if there were any incentives to connect with electronic medical record companies to develop a patient registry. Dr. Brennan responded that the NLM manages a value set authority center. The value set authority centers are bundles of indicators of clinical phenomenon that allow EHRs to assess the impact of treatment. The NLM would like to become a driver for best practices in this area as the challenge will be reconciling across best practices and guidelines with decision logistics. She anticipates that EHR companies will release their individual institutions in the future and the system will be much more distributed. 

Dr. Robinson Beale’s final question was how to incorporate real-world evidence into data steps. Dr. Brennan stated that the NLM has been collaborating with Dr. Charles Friedman at the University of Michigan on his vision of the learning health system, what he refers to as the knowledge grid.

The NLM is planning to partner with him on this endeavor with PubMed Central. The system extracts knowledge from the full-text data and drives data collection and the clinical record back into the knowledge resources. One of the key issues is local and state sensitive information laws and the extent to which they will be able to bring the experience information back into an analytical tool. She suspects that their ability to send analytical tools out to the EHRs is going to be more useful in this system as they will not have to move data. The analysis could be run within the care system and the results would be returned. The analytic tools would aggregate the results after a range of highly distributed analytic resources. Another issue is that some institutions may not want their practice patterns to be exposed by reporting too much information about the linkage between care problems and outcomes. 

Tami Benton, M.D. commented that she supported Dr. Linehan’s suggestion for more guidance on treatments but felt that the cost should not be passed on to the public. Dr. Brennan remarked that the issue was to what extent the Federal Government should become the resource for practice planning. She noted that there has been some recent discussion of the idea of developing a payment system with a more efficient information system within it. 

Guest Speaker

Eliezer Masliah, M.D., Director, Division of Neuroscience, National Institute on Aging

Dr. Eliezer Masliah discussed the National Institute on Aging (NIA) approach towards Alzheimer's disease and related dementias (ADRD). The main objective of the Division of Neurosciences at the NIA is to understand the aging nervous system and ADRD. Alzheimer's disease is the most common neurodegenerative disorder of the aging population followed by Lewy body disease, vascular disease, and frontotemporal dementia. In 2017, 5.3 million people were living with Alzheimer’s dementia in the United States, with an expected tripling of the number of cases by the year 2050. Dr. Masliah remarked on how the mortality rates of most conditions in aging populations have decreased, while the mortality rate of neurodegenerative disorders has increased by about 70 percent. 

In response to this public health emergency, former President Obama and Congress signed the National Alzheimer's Project Act (NAPA ) in 2011. The primary research goal of NAPA is to support studies that may prevent the onset of ADRD and accelerate the development of effective treatments by the year 2025. For FY17, the NIA received an additional $400 million toward this end as part of the $2 billion package that the NIH received, for a total of $34 billion. A third of this funding is being utilized to study the basic mechanisms of ADRD. About a quarter of the funding has been appropriated for epidemiological and translational studies. Additional research is focused on studying genetics, aging, and cognition. Dr. Masliah noted that the pay lines for FY17 CSR-reviewed research applications for Alzheimer’s grants under $500K ranked at the 28th percentile. For new and early career investigators, the pay lines ranked over the 30th percentile.      

Dr. Masliah commented that NIA has received criticism for the research emphasis on β-amyloid and tau proteins, and is now increasing funding for research on synapse networks, inflammation, genetics, and apolipoprotein E. The number of individual research grants awarded has increased substantially since 2014, with 400 awards expected by the end of 2017. NIA also expects to fund between 20 and 25 new clinical trials this year. 

Dr. Masliah noted that some of the grants are in response to initiatives sponsored by ICs other than NIA, particularly NINDS. NIA would like to expand collaboration in ADRD to other ICs.  Dr. Richard Hodes, the Director of NIA, has consulted with Dr. Gordon in this respect for NIMH. 

NIA recently co-funded a few grants with NIMH that they hope to expand upon. Some involved the development of novel radiotracers for muscarinic and nicotinic receptors. One study is a novel positron emission tomography (PET) radiotracer for M1 receptor in Alzheimer’s disease. NIA is interested in funding studies in areas relevant to NIMH such as depression. Another collaboration is the Whitehall II study , a 28-year study that has shown that depressive symptoms are a prodromal feature of dementia. Dr. Masliah stated that NIA is funding several clinical trials that involve neuropsychiatric disorders both in early and late stages of Alzheimer's. They include studies on treatment of agitation and aggression with lithium, methylphenidate, escitalopram, and marijuana as well as non-pharmacological trials utilizing different cognitive modification techniques in patients with ADRD.   

Dr. Masliah mentioned that Dr. Jovier Evans of NIMH and staff from NIA collaborated on a workshop on novel approaches to understanding the mechanisms of neuropsychiatric symptoms in Alzheimer's and advancing therapy development. The Division of AIDS Research at NIMH, in collaboration with an NIA program, hosted the NeuroHIV in the ART Era meeting in October 2017. NIA is currently considering opportunities to host a symposium with NIMH to discuss how silent HIV infection after anti-retroviral therapy, might result in an increased development of disorders like Alzheimer's and dementia with Lewy bodies in older individuals. 

NIA had 27 new funding opportunities in 2017 to stimulate the field and get new applications. Many of them incorporated both basic and clinical science. Some examples include studying the role of exosomes, the impacts of sex differences, pragmatic trials for dementia care in long term services, socially assistive robots, and commonalities and interactions between neurodegenerative diseases. 

NIA organizes an Alzheimer's summit and an ADRD summit each year in collaboration with NINDS and other ICs to obtain scientific feedback from the community and to develop milestones and new funding opportunities. The 2017 ADRD milestones included continuing to support basic mechanisms of disease, greater integration and synergy among all the different NIA programs, enhancing recruitment for clinical trials and natural history studies, and increased funding for young investigators and training on ADRD. Dr. Masliah stated that NIA would like to organize their programs so that they become part of the pipeline for target discovery, drug development, and clinical trials. He hopes that by 2025, they will reach a central milestone of developing a therapeutic for Alzheimer's disease. 

Several NIA programs involve discovery of new targets through whole genome and gene typing studies. The Alzheimer's Disease Sequencing Project  (ADSP) also has an Alzheimer's Disease Genetics Consortium  (ADGC). It is mostly focused on identifying new genomic variants contributing to increased risk of developing Late-Onset Alzheimer's Disease. AGDC previously funded several genome-wide association studies (GWAS) and whole exome sequencing studies that resulted in the discovery of a number of new genes. Some of these genes were in the inflammatory pathways, such as CD33, CR1, CLU, and most notably TREM2.

Other NIA programs are using proteomics for new target discovery. The Accelerating Medicines Partnership  (AMP) is the brainchild of Dr. Collins. NIA’s AMP program is the AMP-Alzheimer’s Disease (AD) Program, a target discovery and preclinical validation project. All data are broadly shared through a central portal. It includes epigenomic, proteomic, and genomics studies utilizing 2,500 brains with pathological diagnoses. 

In 2017, NIA funded a new concept, the Alzheimer’s Disease Translational Center for Animal Model Resources (MODEL-AD). NIA felt that current rodent models of Alzheimer's disease were insufficient as they are primarily models of dominantly inherited Alzheimer's disease or familial Alzheimer's disease. Very little research was available on models of sporadic Alzheimer's disease that included the new genes being discovered by ADSP and AMP-AD. The concept for MODEL-AD is to receive input from ADSP and AMP-AD and share data through Sage Bionetworks and a portal called Synapse. The Center just developed a TREM2 mutant mouse model that is similar to the presentation of sporadic Alzheimer’s disease.   

Dr. Masliah went on to state that NIA has 31Alzheimer’s Disease centers that cover a wide area geographically across the United States. The main objective of these Alzheimer's centers is to recruit and characterize patients,
but these patients can then participate in other programs such as natural history, biomarker, or clinical trial type studies. All the centers are integrated by a National Alzheimer’s Coordinating Center (NACC) based at the University of Washington. All NACC data is available to any investigator. The samples obtained through the Alzheimer’s centers are sent to the National Central Repository for Alzheimer's Disease (NCRAD) for processing. NCRAD produces high-quality DNA, RNA, cell lines, fibroblasts, etc. The DNA is sent to ADGC for sequencing. As of 2017, all data sharing and storage is handled by the National Institute on Aging Genetics of Alzheimer's Disease Data Storage Site (NIAGADS). 

The next phase will be to link with the Alzheimer’s Disease Neuroimaging Initiative  (ADNI), a neuroimaging study that transformed into a comprehensive network of 57 sites in the United States and Canada that perform clinical measures, cognitive tests, MRIs, and PET. ADNI generated important biomarker data for Alzheimer's disease primarily through breakthrough discoveries of PET imaging. The biomarkers were positive in patients 20 to 30 years before the onset of the disease, which has inspired new ways of thinking about Alzheimer’s disease and planning future clinical trials.

Dr. Masliah commented that NIA also has a Dominantly Inherited Alzheimer Network  (DIAN) that has identified biomarkers that appear several years before the onset of Alzheimer's disease.  NIA is therefore currently funding either primary or secondary prevention clinical trials based on biomarkers and the onset of clinical symptoms in asymptomatic patients. Clinical trials for sporadic patients are also occurring that are prodromal studies of the disease. Most of these studies have not shown significant improvements in clinical symptoms but have shown some improvement in the biomarkers. 

NIA is developing an Alzheimer’s Clinical Trials Consortium (ACTC) to run trials focused on interventions that may prevent, delay, or treat the symptoms of AD and other age-related dementias. It will include multiple clinical trial sites and trial coordination and management infrastructure. An announcement will soon be made as to which of the three applications for this new consortium will be funded. Dr. Masliah reminded the Council of The National Research Summit on Care, Services, and Supports for Persons with Dementia and Their Caregivers occurring on October 16-17, 2017, as well as the 3rd Alzheimer’s Disease Research Summit to be held on March 1-2, 2018. 

NIMH Computational Psychiatry Workshop Highlights

Bruno Averbeck, Ph.D., Chief, Section on Learning and Decision Making, NIMH

Dr. Gordon stated that NIMH hosted a scientific workshop on computational psychiatry in June. Dr. Bruno Averbeck, a computational neuroscientist was one of the organizers of the workshop and provided highlights of the workshop to NAMHC. Dr. Averbeck described computational psychiatry as an emerging field defined by the application of computational and analytical techniques to psychiatry that uses data and theory-driven approaches to improve diagnosis, treatment, and prediction of psychiatric disorders. Computational psychiatry is focused on larger data sets and mechanistic theories of brain function.   

Dr. Averbeck noted that the workshop attendees were very humble and realistic, and paraphrased the sentiments of several participants as, "if we could point to any one single small problem where we made real concrete progress, then that would actually be something good and beneficial." The goal was not to start a revolution, but rather to identify some problems where progress can be made. The other organizers included Dr. Oquendo, Dr. Michele Ferrante of NIMH, Dr. Megan Kinnane of NIMH, and Dr. David Redish of the University of Minnesota. There were 50 attendees and almost everybody invited attended, which demonstrates a lot of excitement in the field. The attendees were drawn from experimental and computational neuroscience backgrounds and clinicians. The objectives of the workshop were to identify how computational psychiatry as a field can move forward, how to foster the development of a community of scholars working in fields related to computational psychiatry, and how to ensure the alignment of perspective between investigators and NIMH to foster this nascent field. Discussions were organized around four topics: (1) evaluation of the field, (2) computation, (3) psychiatry, and (4) basic/fundamental science. Groups of about 10 to 12 individuals discussed each topic and one person from each group put together a short summary of what each group had discussed during the breakout sessions to present to the entire group.

Dr. Averbeck commented that one of the challenges identified in terms of the evaluation of the field is developing a common language for people trained in diverse fields. The workshop participants also considered the need to improve upon diagnosis and prognosis as well as assign precision treatment. There is a need for implementing translation. It might be a series of behavioral tasks along with brain imaging measures collected from a subject to improve a diagnosis. It also takes time to build a multidisciplinary field. Dr. Averbeck projected that only someone with a theoretical background who has not worked in biology for five years can start to get a decent appreciation for these problems.   

Success would be measured in the clinic in terms of progress on diagnosis, treatment, and prognosis. Consortia are needed to evaluate the success of tools. From a clinical point of view, prediction could be informative without considering how the brain works. It can still be helpful if scientists use a big data approach to figure out a better treatment even though they have not developed a deeper understanding of how the brain works.

One of the psychiatrists at the workshop noted that explanatory models provide the “why” and the “why” can help in those instances where the effective treatment by itself is not always enough. For example, there is no treatment for an adult high functioning person with autism, but you can give them an explanation for their condition.  In some domains where there is no or little treatment, these metaphors can be healing. 

There were a few consistent themes across topics at the workshop. From Dr. Averbeck’s perspective, the biggest theme was the tension between using top-down theory-driven or bottom-up data-driven approaches. A theory-driven approach is a mechanistic description of how the brain gives rise to behavior. For example, in basic science it is understood that dopamine is important for reinforcement learning, and this mechanistic description has been used to generate a theory for how delusions are formed in schizophrenia. A data-driven approach would be to collect a large data set and let computational techniques “discover” the organization in the data. A computational model of a brain-behavior relationship can be used to study psychopathology.  But data-driven approaches contain theory and theory-driven approaches contain data. For example, the choice of data to collect in a big sample will be driven by theoretical models of the brain or “best-guesses” of what the relevant measurements might be. 

Workshop attendees considered whether existing big data sets such as Google, Facebook, and Twitter could be mined. They agreed that data mining on genetic and brain imaging data sets is much more straightforward. Dr. Averbeck mentioned that you cannot assume large population studies are well distributed. For example, Mechanical Turk  is an Amazon-based website being used to collect behavioral study data. The problem is that the people who self-select to complete these tasks are a subset of society. Another approach is to define new diagnostic or treatment categories. 

Dr. Averbeck concluded his presentation with an example graphing the considerable heterogeneity in Diagnostic and Statistical Manual of Mental Disorders (DSM) categories and demonstrating how psychiatric labels may not be that valuable.  He showed how computational algorithms can be used to find the structure in the data.  Data can be described using categorical or dimensional models, but it can be hard to say which is “correct.”  He used a Bayesian approach to maintain the dimensional richness of the original data and quantify classification to groups.  The problem is that scientists do not know how many groups they should classify from large data sets and are not able to instruct algorithms to identify an unknown number of groups.  There are analytical tools for determining the number of data clusters, but it is very difficult to accomplish with large, noisy, complex, high-dimensional data sets. In these instances, it is necessary for computational and clinical psychiatrists to work together and apply theory and expert knowledge to the analytical tools available.    

Discussion

Dr. Brent was surprised by the pessimism surrounding cell phone data. He was not sure if it would assist with the nosology, but he believed passive cell phone data could be promising for monitoring fluctuations in the status of illness. Dr. Averbeck noted that he did not specifically mention cell phone data, but thought that it might be valuable in activity monitoring of depression. Inferring a person’s mood from Facebook and Google would be more difficult. Paolo del Vecchio, M.S.W. commented that Samsung recently announced a partnership to develop virtual reality using their gear technology for diagnostics, assessments, and treatment. They hope to launch this program to hospitals and other treatment settings as well as households starting next year. He was concerned about the rapid development of these approaches and encouraged NIMH to keep investing resources in this area. Dr. Gordon replied that NIMH is pursuing solutions to the challenge of testing and validating apps and other technology-based treatments promoted for commercial use.

Dr. Krystal stated that computational psychiatry has changed the way data is used and models are integrated. How to translate computational models into formal theory is an underappreciated gap in psychiatry research. Computational techniques are data hungry and rely on the broad dissemination of large data sets.    

Concept Clearances

Geetha Senthil, Ph.D., Office of Genomics Research Coordination

Development and Validation of Technologies for Rapid Isolation and Characterization of Extracellular Vesicles of Central Nervous System Origin

Dr. Geetha Senthil stated that her concept is a technology-focused initiative developed by the NIMH to contribute to the NIH Blueprint for Neuroscience Research . The NIH Blueprint is a framework for ICs with neuroscience portfolios to collaborate and to identify and support cross-cutting new science research by pooling resources and expertise. The Blueprint is comprised of 14 ICs and the Office of the Director. 

Dr. Senthil explained how extracellular vesicle signaling is a universal biological process. Extracellular vesicles are a heterogeneous mixture of nanoscale membrane-bound vesicles including micro-vesicles and exosomes. They are produced by a diverse range of cell types and ubiquitously found in many biofluids. These vesicles carry cell type-specific molecular cargoes of cytosolic and membrane-bound proteins, lipids, functional regulatory RNAs, and other molecules reflective of the cell origin. 

Once released into the extracellular space, extracellular vesicles can be taken out by target cells, either in distant or neighboring regions. The extracellular vesicles deliver their cargoes to the target cells, which can affect their physiological state.  They in effect act as an important mediator for cell-to-cell communication and signaling.

There is growing evidence that extracellular vesicle signaling plays an important role in the normal physiological functions and pathological processes of all major central nervous system (CNS) cell types. For example, exosomes can spread cancer proteins to healthy cells. More importantly, the exosomes produced by these different cell types in the CNS can cross the blood-brain barrier and enter the peripheral circulation. The exosomes derived from the extracellular vesicles in peripheral circulation offer a clinical opportunity to assay the health and function of the brain in a non-invasive manner.

The challenge is that not much is known about the cellular origin of the extracellular vesicles derived from the CNS cells in the peripheral circulation in humans. The key barriers are limited technologies and protocols for isolating and detecting these CNS-derived extracellular vesicles.

The concept seeks to enhance strategies and develop new methodologies for reliable and reproducible isolation, detection, purification, and characterization of extracellular vesicles derived from different CNS cell types from human biofluids. Examples of such techniques include cost-effective high-throughput methods for isolating and purifying these extracellular vesicles. Areas of interest could include the development of protocols optimized for CNS-EV isolation from different biofluid sources and the development of analytical technologies that can identify and track the cellular origin of these vesicles and their molecular cargoes. 

If successful, the concept will result in novel tools for non-invasive assays of CNS function and disorders, and these methods could potentially provide new strategies for prognosis, diagnosis, and intervention. The Common Fund extracellular RNA program is much broader. This CNS focused initiative complements and extends the Common Fund program. Dr. Senthil thanked her team, particularly Dr. Doug Meinecke, who co-led the development of the concept.

Discussion

Dr. Walsh expressed his enthusiastic support for the concept and stated that extracellular vesicles have important roles as mediators of extracellular signaling. Researchers are only just beginning to understand how expansive their roles are. The concept that exosomes from brain tumors are found in peripheral blood is very well documented. Dr. Huganir remarked that there are some unpublished studies showing that presynaptic terminals are releasing exosomes that then transfer to post-synaptic membranes. Because the tools for studying them are poor, he applauded the timeliness of the effort. 

The Council unanimously passed a motion to approve the concept. 

Michele Ferrante, Ph.D., Division of Translational Research and Division of Neuroscience and Basic Behavioral Science

Computational Models for Validating Dimensional Approaches to Psychopathology

Dr. Ferrante stated that the purpose of the concept is to stimulate the validation of dimensional approaches in psychopathology. Clinicians are attempting to classify patient populations into neuro-behavioral types (biotypes) utilizing different units of analysis. To better understand the physiological and psychological basis of psychopathology in terms of domains of function that do not solely rely on DSM diagnoses, the concept would focus on utilizing a Research Domain Criteria (RDoC) framework and sophisticated computational approaches to understand pathophysiology in clinical populations.    

Groups of scientists came together to define domains of function and the different constructs of the RDoC. If data-driven approaches could be applied to different behavioral assays on the same construct to validate and refine the hierarchical constructs within the same domain of function, some of these tests might be predictive of treatment outcome or classifications of patients. The concept proposes to stimulate collaboration between clinical and computational neuroscientists using theory and data-driven models for RDoC studies. This will require at least two levels of analysis, including high-resolution measurements of neural circuits and behavior activity using ecological momentary assessment and active and passive methods of data collection. The concept also includes three converging behavioral paradigms probing each RDoC dimensional construct of interest to get convergent validity for the measures and to test whether dimensional constructs can be identified, refined, merged, subdivided, hierarchically organized, and/or depend on each other through convergent mechanisms.

Dr. Ferrante stated that the concept will utilize accelerated longitudinal studies and predictive algorithms to make predictions as to which children might develop pathologies. They will appropriately include healthy subjects and patients from multiple diagnostic groups, as there are certain domains of function that may be predictive of future pathologies.

On a separate note, the concept will stimulate competitive supplements to analyze large-N data sets. These data sets could be merged to increase patient population heterogeneity or to supplement studies to add RDoC measures to models. It is important to develop new platforms for the integration of computational modeling into clinical research on psychiatric disorders, especially considering that new patients are being added to the All of Us Research Program. 

Discussion

Dr. Krystal mentioned that a computational psychiatry meeting in Germany identified conflicts between how categorical and dimensional approaches to psychiatry can be evaluated within single models. He considered whether the strong focus on dimensional approaches could be broadened to include dimensional and categorical approaches together to the extent that they are complementary and interacting perspectives, particularly within a Bayesian framework. Dr. Oquendo agreed that the categorical aspects are still important in medical decision-making. She also noted that when using high-dimensional data such as brain imaging, the pre-processing of the data is extremely expensive, and there is always a concern that important data is being discarded. The use of computational approaches can help evaluate whether they are retaining all relevant data to study the underlying basis of disease. She added that computational scientists have a lot to learn in terms of how to manage complex brain, environment, and behavioral data. 

Ian Gotlib, Ph.D. expressed concern about the requirement for three behavioral measures for each construct. He suggested using a behavioral approach to see what constructs come from computational psychiatry rather than become focused on existing constructs. Dr. Ferrante replied that the idea is to define the psychological construct based on the underlying biology. For example, if you want to define the brain circuits involved in a construct, you would need to have a broad range of behavioral constructs you want to explore, to reveal the plasticity of the system involved. The psychological construct would be redefined by the underlying biology using a data-driven approach. 

Dr. Gordon noted that this was an issue that the attendees at the computational psychiatry workshop struggled with. There is a need to incorporate the categorical approaches using mathematical methods. The challenge is merging the data-driven and theory-driven approaches. As Dr. Averbeck explained, even data-driven approaches are developed from some theoretical concepts about the behavior. Dr. Gordon explained that Dr. Ferrante’s concept implies that the use of the RDoC constructs is not mandatory. The challenge is to encourage approaches that are cross-domain and cross-construct, as well as categorical and dimensional. 

Gregory Miller, Ph.D. asked for clarification as to whether Dr. Gotlib was referring to three different constructs or three different measures of a given construct. Dr. Ferrante stated that the concept refers to three different measures of the same construct under evaluation. Dr. Miller noted that starting with three measures of reward responsivity will result in reward responsivity as the construct. Dr. Gordon agreed and mentioned that the next step might be deep behavioral phenotyping on a large scale to help break the divisions down. Dr. Ferrante explained that for each of the constructs of interest, they were interested in collecting multiple behavioral assays. 

Dr. Miller stated that the concept is compatible with RDoC because RDoC is not the matrix. The matrix is an example of RDoC. He is co-chairing a workgroup to propose changes in the matrix. The concept is therefore a supplement to the matrix as it currently exists. 

Dr. Brent stated that he is concerned about how to identify individual patients in the current data sets. Dr. Ferrante responded that the same behavioral assays should be performed on the same patients. Dr. Gordon added that investigators can submit patient level data with patient level identifiers to NIH that can link data from multiple data sets and subject areas. He is working with Dr. Brennan to accomplish this on an NIH-wide scale. Dr. Gotlib confirmed that as a contributor, his team works hard to identify every patient with every piece of data.

The Council unanimously passed a motion to approve the concept. 

Public Comment

Ms. Valerie Porr identified herself as the founder of the Treatment and Research Advancements National Association for Personality Disorder (TARA4BPD). She added to Dr. Linehan’s comments on the paucity of education and training for young researchers and clinicians and the need for NIH collaboration with universities to help prepare the future scientific workforce.

Adjourn Open Session

Dr. Gordon thanked the participants for attending and adjourned the open session of the NAMHC meeting at 12:37 PM. 

Joshua A. Gordon, M.D., Ph.D.

Appendix A

Summary of Primary MH Applications Reviewed

Council: September 2017

IRG Recommendation
Category	Scored #	Scored Direct Cost $	Not Scored (NRFC) #	Not Scored (NRFC) Direct Cost $	Other #	Other Direct Cost $	Total #	Total Direct Cost $
Research	590	$962,186,318	414	$502,997,734	7	$2,044,848	1011	$1,457,228,900
Research Training	0	0	0	0	0	0	0	0
Career	57	$41,676,947	25	$19,915,969	0	0	82	$61,592,916
Other	0	0	0	0	0	0	0	0
Totals	647	$1,003,863,265	439	$522,913,703	7	$2,044,848	1093	$1,528,821,816

Appendix B

Department of Health and Human Services
National Institutes of Health
National Institutes of Health
National Advisory Mental Health Council
(Terms end 9/30 of designated year)

Chairperson

• Joshua A. Gordon, M.D., Ph.D.
  Director
  National Institute of Mental Health
  Bethesda, MD

Executive Secretary

• Jean Noronha, Ph.D.
  Director
  Division of Extramural Activities
  National Institute of Mental Health
  Bethesda, MD

Members

• Tami D. Benton, M.D. Ad Hoc (Pending)
  Psychiatrist-in-Chief
  Department of Child and Adolescent Psychiatry And Behavioral Sciences
  Children’s Hospital of Philadelphia
  Philadelphia, PA
• David A. Brent, M.D. (17)
  Academic Chief
  Child & Adolescent Psychiatry
  Endowed Chair in Suicide Studies
  Professor of Psychiatry, Pediatrics and Epidemiology
  Director, Services for Teens at Risk
  University of Pittsburgh School of Medicine
  Pittsburgh, PA
• Randy D. Blakely, Ph.D. Ad Hoc (Pending)
  Professor
  Department of Biomedical Sciences
  Charles E. Schmidt College of Medicine
  Florida Atlantic University
  Jupiter, FL
• Benjamin G. Druss, M.D., M.P.H. (18)
  Rosalynn Carter Chair in Mental Health and Professor
  Department of Health Policy and Management
  Rollins School of Public Health
  Emory University
  Atlanta, GA
• Ian H. Gotlib, Ph.D. Ad Hoc (Pending)
  David Starr Jordan Professor and Chair
  Department of Psychology
  Stanford University
  Stanford, CA
• Alan E. Greenberg, M.D., M.P.H. Ad Hoc (Pending)
  Professor and Chair
  Department of Epidemiology and Biostatistics
  School of Public Health
  George Washington University
  Washington, DC
• David C. Henderson, M.D. Ad Hoc (Pending)
  Chair
  Department of Psychiatry
  Boston University School of Medicine
  Boston, MA
• Michael F. Hogan, Ph.D. (18)
  Consultant and Advisor
  Hogan Health Solutions LLC
  Delmar, NY
• Richard L. Huganir, Ph.D. (17)
  Professor and Director
  Department of Neuroscience
  Investigator, Howard Hughes Medical Institute
  Co-Director, Brain Science Institute
  The Johns Hopkins University School of Medicine
  Baltimore, MD
• Lisa H. Jaycox, Ph.D. Ad Hoc (Pending)
  Senior Behavioral Scientist
  Health Program
  Rand Corporation
  Arlington, VA
• John H. Krystal, M.D. (19)
  Robert L. McNeil, Jr. Professor of Translational Research
  Chair, Professor of Neurobiology
  Chief of Psychiatry, Yale-New Haven Hospital
  Department of Psychiatry
  Yale University School of Medicine
  New Haven, CT
• Marsha M. Linehan, Ph.D. (17)
  Professor and Director
  Behavioral Research and Therapy Clinics
  Department of Psychology
  University of Washington
  Seattle, WA
• Gregory A. Miller, Ph.D. Ad Hoc (Pending)
  Professor and Chair
  Department of Psychology
  University of California, Los Angeles
  Los Angeles, CA
• Maria A. Oquendo, M.D. (17)
  Ruth Meltzer Professor of Psychiatry & Chairman
  Department of Psychiatry
  Perelman School of Medicine
  University of Pennsylvania
  Philadelphia, PA
• Rhonda Robinson Beale, M.D. (19)
  Senior Vice President and Chief Medical Officer
  Blue Cross of Idaho
  Meridian, ID
• Elyn R. Saks, J.D., Ph.D. Ad Hoc (Pending)
  Orrin B. Evans Professor of Law
  Gould School of Law
  University of Southern California
  Los Angeles, CA
• Hyong Un, M.D. (17)
  Head of EAP & Chief Psychiatric Officer
  AETNA
  Blue Bell, PA
• Christopher A. Walsh, M.D. (19)
  Chief, Division of Genetics and Genomics
  Boston Children’s Hospital
  Bullard Professor of Pediatrics and Neurology
  Harvard Medical School
  Boston, MA

Ex Officio Members

Office of the Secretary, DHHS

Thomas E. Price, M.D.
Secretary
Department of Health and Human Services
Washington, DC

National Institutes of Health

Francis Collins, M.D., Ph.D.
Director
National Institutes of Health
Bethesda, MD

Department of Veterans Affairs

Amy Kilbourne, Ph.D., M.P.H. (Pending)
Deputy, Quality Enhancement Research Institute
Department of Veterans Affairs
Washington DC

Department of Defense

Steven E. Pflanz, M.D.
Air Force Director of Psychological Health
Mental Health Branch Chief
Air Force Medical Support Agency
Falls Church, VA

Liaison Representative

Paolo del Vecchio, M.S.W.
Director
Center for Mental Health Services
Rockville, MD