News‎ > ‎

BRAIN RFI. Vote on the future of phase II by Nov15

posted Nov 8, 2018, 2:35 PM by Bionic Lab   [ updated ]

What you need to know about Phase 2 BRAIN RFI.

The BRAIN Initiative was a 2013 White House initiative to support better understanding the brain. The Budget for the BRAIN Initiative has also expanded with recent bipartisan legislation securing funding to at least 2026 with an operational budget greater than NIBIB. There were some politics involved, but ultimately a decision was made to use the money in phase 1 to largely support brain related proposals that struggled in traditional mechanisms and traditional review panels. These ended up being grants that were engineering based (design driven proposals) as opposed to classical neuroscience hypothesis driven proposals that are funded through traditional mechanisms under NIBIB, NINDS, NIMH, NIA, NIE, etc (eg. parent R01s)

BRAIN Initiative Phase 1 is closing to an end, they the committee/working group is trying to pick a direction for Phase 2. There is considerable internal pushback against technology development and a push for funding “circuit function” research (ie research that is already funded through traditional mechanisms).

There also remains a major gap between "Technology Development" and "Circuit Function", which is not addressed under current BRAIN Initiative Mechanisms. For example, there is very little "Interface Science" research to help guide/inform technology development.


1)      Neural Interface research is not represented on the working group panel nor the leadership of the BRAIN Initiative. The major representation is made up of neuroscientists despite “The overarching vision of the BRAIN Initiative is best captured by Goal #7 (New Technology)”

2)      Effective increases to pay lines: by removing support for grants that struggle in normal review/study sections (ie technology development), it frees up money. Then, by supporting BRAIN grants that already have standard mechanisms (parent R01s) through the BRAIN Initiative, it frees up money in the traditional mechanisms (ie parent R01s). This effectively raises the pay line for standard neuroscience grants.

Lessons from 2017 GSI

The NIH announced the “Grant Support Index (GSI)” on May 2, 2017 to support new investigators (average age of 1st R01 has shifted from 35 (1980) to 47 (2017)). While I’ve been fortunate to receive my R01 at 32, this is a huge issue. The GSI was designed to help decrease that first R01 age, by capping funding to each investigator at a maximum of 3 R01s. However, the NIH received overwhelming feedback that led to abandoning the GSI on June 17, 2017. This is because graduate students, postdocs, and assistant professors that would benefit from it never responded to RFI/request for feedback, where was STRONG and overwhelming feedback was provided by a few directors and PIs with 5+ R01 equivalents.

Lesson: This is your opportunity to have your voices heard. You don’t have to be a PI or a faculty. You can be a postdoc, student, industry, or a barista. You just need to submit your opinions by Nov 15th 2018.

Here are some talking points that I think are important and within the scope of the BRAIN 2025 roadmap and priorities, but are overlooked missing in current RFAs and funding opportunities:


Tools and Technologies

  • ·       Tools and technology do not fair well in most standard study sections and mechanisms because they are peer reviewed by scientists that technology be developed in a hypothesis driven approach. 
  • ·       Neurobiology of Neurotechnology (Neural Interface Cybernetics: ie the science of communications and automatic control systems in both machines and living things): So much focus is on building “new” technology, there is very little effort to focus on the “science that governs functional communication between biology and interface”. This goes beyond biocompatibility and tissue scarring. How does biology govern communication with the technology and how does technology influence neural activity?
  • ·       Too much emphasis on channel count and not enough on intelligent design: there are no RFAs currently available to study what makes a good interface and what makes a bad interface. Instead, current focus is on “increasing channel count”. But maximizing channel count likely means a greater “observer principle” alteration of the natural circuit. A better approach would be evaluate the science of the technology interface to inform better design and optimization of different design parameters. See
  • ·       Technology will always be obsolete, but the knowledge generated to guide technology development will be immortal. Therefore, there should be emphasis on understanding the biology of neural interfaces. see
  • ·       Cross-training: The importance of understanding both the language of neuroscience and neural engineering is emphasized elsewhere. It also explains why you can’t “just collaborate, you do the engineering and I’ll do the science” (closed-loop engineering):
  • Momentum: But dropping funding for technology, the infrastructure for technology gets disassembled. The training, engineers, and development pipeline get disassembled and reset. This means the next round of technology development grant means the training and infrastructure has to be rebuild from scratch. Instead, NIH should maintain the pipeline to facilitate new innovation tomorrow. It's naive to say we're done with brain technology innovation; "we've solved it"

Circuit Function

  • ·       Observer Principle: U despite all the tools that are being developed,  we still have very little idea how using these tools alter natural circuitry function. How can you use these tools to study circuit function if you don't understand the "side effects" on native neural circuits from using these tools? (eg, electrodes cause silencing of nearby neuron (alive but not firing properly) because oxygen and nutrient delivery is damaged. This means that instead of increasing channel count for next generation electrodes, but we should be focusing on technology that helps repair the neurovascular unit around implants. See
  • ·       Assumptions: Many neurotechnologies are based on assumption. There is immerging research that suggests or proves that some of these assumptions are wrong (eg NeuN labeled neurons around electrodes are not silenced.). Instead, there should be RFAs that are designed to tease apart and test these assumptions.
  • ·       Standard Technology: There is no emphasis on studying how standard technology influences neural circuit function. How can we evaluate if “new” technology from BRAIN is better or not if there are no efforts to evaluate these impacts on neural circuits with standard technologies?
  • ·       Non-standard approaches for utilizing standard neurotechnology: when asking an adult “how many ways can you use a paperclip”, most will answer with 1-3 that are related to hold sheets of materials. Similarly, utilizing technology to understand neurodegeneration and regeneration/wound healing is not widely explored, especially their influence on neural (and gliovascular) circuit function.
  • ·       Disease/Degeneration: Many degenerative diseases and brain injuries share activation of similar pathways. The key differences are where and what triggers onset. What makes these degenerative disease hard to study is that their onset is difficult to identify and the focal point of activation is hard to pinpoint (some have multiple focal initiation points). With technologies, the time and location is precisely known and you get a bonus of having a sensor or effector at the epicenter of the trigger injury. This not only informs technology development, but also understanding disease progression and potential platform for treatments.
  • ·       So much emphasis is on Neural Circuits. Despite the fact that microglia, astrocytes, oligodendrocyte, NG2 glia, pericytes, etc regulate neural activity in different way and each express a number of different subtypes and phenotypes, AND outnumber neurons, there is disproportionately fewer RFA mechanisms supporting their understanding on neural circuit function.
  • ·       Without underlying basic science to understand how to regulate “non-neuronal cells” and how “non-neuronal cells” regulate neural circuits, how can we create gliotechnologies and neurovasculartechnologies?


Dissemination and Training

  • ·       Cross-training: The importance of understanding both the language of neuroscience and neural engineering is emphasized elsewhere. It also explains why you can’t “just collaborate, you do the engineering and I’ll do the science” (closed-loop engineering):



  • ·       Representation: No advocates for technology interface nor are there any individuals that appreciate the neural interface challenge serve on the working group. One technologist on the working group has, on multiple occasions, stated in talks, “I’m not really sure what happens to the brain [when this is implanted]. I’m not a biologist, it’s not important to me.”

Topics that do not do well in traditional NIH mechanisms

1)    1. Technology development

2)    2. Neural Interface Science/Cybernetics

3)    3. Regulatory Neurotechnology science

a.     What makes technology safe

b.     What makes technology effective

c.     We don’t even know which experiments to carry out to evaluate the above questions. We need research to determine what experiments and metric standards inform safety and effectiveness at both pre-clinical and clinical trial levels