A Tale of Two BRAINS: #BRAINI and DARPA's SUBNETS

Image credits. Left: SUBNETS program (DARPA). Right: BRAIN interim report presentation (NIH).


In April, the White House announced the $100 million Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative. The goals of this bold new research effort are to "revolutionize our understanding of the human mind and uncover new ways to treat, prevent, and cure brain disorders like Alzheimer's, schizophrenia, autism, epilepsy, and traumatic brain injury." A series of high-profile journal articles traced the genesis of this initiative from the Brain Activity Map idea to develop nanotechnologies and "image every spike from every neuron" (Alivisatos et al., 2012) to its current emphasis on neural circuits and systems neuroscience more broadly construed (Insel et al., 2013). In the first year (FY 2014),¹ $50 million will be allocated to DARPA and $40 million to NIH.²

The two federal agencies have taken starkly different approaches to the challenge, in terms of timing and scope. They also address different levels of nervous system function. Both are ambitious, but one surpasses earlier calls for a "moon shot" to the mind.  IF successful,³ it would render much of pre-clinical neuroscience research quaint and obsolete (except for providing mechanistic details).


The Tale of Two BRAINS

1. The National Institute of Health (NIH) sponsored a series of meetings and solicited public feedback. The NIH Director's BRAIN Advisory Committee issued its Interim Report (PDF) on September 16. The report focuses primarily on animal models, including improved technologies for recording neuronal activity and manipulating circuit function.⁴ Here are the high-priority research areas for FY 2014:

#1.  Generate a Census of Cell Types.
#2.  Create Structural Maps of the Brain. 
#3.  Develop New Large-Scale Network Recording Capabilities.
#4.  Develop A Suite of Tools for Circuit Manipulation.
#5.  Link Neuronal Activity to Behavior.
#6.  Integrate Theory, Modeling, Statistics, and Computation with Experimentation.  
#7.  Delineate Mechanisms Underlying Human Imaging Technologies.
#8.  Create Mechanisms to Enable Collection of Human Data.
#9.  Disseminate Knowledge and Training.

These are very ambitious projects, each delineated in more detail in the full report (PDF). However, the NIH has yet to issue a Request for Applications that outlines the requirements for grant proposals submitted through this program.


2. On the other hand, the Defense Advanced Research Projects Agency (DARPA) announced their goals for the BRAIN Initiative via the New York Times on October 24 and and issued a broad agency announcement (call for proposals) on October 25. The focus is on developing technologies and treatments that use deep brain stimulation (DBS), which has been highly successful in Parkinson's Disease.

There are 3 Technical Areas that are covered in the announcement. All applicants must address Area One, and teams of investigators are encouraged to address all three.

TA One is comprised of clinical trials in order to establish mechanistic models of awake, behaving human brain activity.

TA Two encompasses the hardware development component in order to create safe and effective sensing and stimulation systems.

In TA Three, investigators will use human-relevant animal models [primates, not rodents] to generate safety and efficacy data as well as establish preliminary theories and rapidly prototype hypotheses regarding the links between neural data and clinical outcomes. 

To elaborate, over a 5 year period, the successful applicants must conduct clinical trials in human patients with 7 specified psychiatric and neurological disorders (not including PD), some of which have never been treated with DBS. The successful teams will use devices that both stimulate and record neural activity, and provide real-time data that can be decoded as reflecting a particular behavioral state... basically, a futuristic implant that can adjust its own stimulation parameters based on how the patient is doing. At least, that's how I interpret it.

Brainstorm (1983). Brain-computer interfaces have improved a bit in the last 30 yrs.

Systems-Based Neurotechnology for Emerging Therapies (SUBNETS)

Let's take a closer look at TA One of DARPA's SUBNETS program.

SUBNETS is distinct from current therapeutic approaches as it seeks to develop the ability to create a closed-loop diagnostic and therapeutic system. Through measuring pathways involved in complex systems-based brain disorders such as depression, compulsion, debilitating impulse control, and chronic pain, SUBNETS will attempt to establish the capability to record and model how these systems function in both normal conditions, among volunteers seeking treatment for unrelated neurologic disorders, as well as impaired clinical research participants. SUBNETS will then use these models to determine appropriate therapeutic stimulation methodologies that meet guidelines for both safety and efficacy in human participants. These models will be adapted onto next-generation, closed-loop neural stimulators that exceed currently developed capacities for simultaneous stimulation and recording and provide a research investigator, a clinician, and a human research participant with the ability to record, analyze, and stimulate multiple brain regions for therapeutic purposes.

Seven disorders are targeted: Post-Traumatic Stress Disorder (PTSD), Major Depression, Borderline Personality Disorder (BPD), General Anxiety Disorder (GAD), Traumatic Brain Injury (TBI), Substance Abuse/Addiction, and Fibromyalgia/Chronic Pain. To the best of my knowledge, there is no published literature on DBS for PTSD, BPD, GAD [as opposed to OCD], or TBI [except for minimally conscious state]. At clinicaltrials.gov, a Pilot Study of DBS of the Amygdala for Treatment-Refractory Combat PTSD was withdrawn prior to enrollment. There's one DBS trial for TBI that aims to enroll 5 patients over a 4 year period. I couldn't find anything for BPD or GAD (although these disorders might possibly be comorbid in some patients treated for depression or OCD).

With this starting point in mind...

All proposers must address each of the conditions described above by the end of Phase Two of this program. If, through the course of Phase One, a given condition is demonstrated to not be amenable to modeling and intervention as described in this BAA, performers will be allowed to describe an alternative condition.

Here are the milestones for Phase One (24 months):

• Generate models for two DSM diagnosis categories chosen from: (A) Major Depression, (B) PTSD, (C) General Anxiety Disorder, (D) Borderline Personality Disorder.
• Generate models for one disrupted neurologic of physiologic system chosen from: (A) Fibromyalgia/Chronic Pain, (B) TBI, (C) Substance Abuse/Addiction.
• Demonstrate ability to intraoperatively relieve symptom severity with computer in the loop (measured through neural signature of disease as well as phenotypic presentation).
• Evaluate the effects of intraoperative stimulation on symptomology at four hours, 24 hours, and 72 hours through neural recordings and behavioral experiments.

For Phase Two (36 months):

• Refine models from Phase One.
• Generate models for remaining two DSM diagnosis categories and two neurological/physiological systems.
• Demonstrate ability to intraoperatively relieve symptom severity with device in the loop (measured through neural signature of disease as well as phenotypic presentation).
• Demonstrate ability to chronically relieve symptom severity over 14 day window.
• Demonstrate ability to provide therapy in response to stimuli in free-living environment.

So perhaps every major DBS center (e.g., Emory, University of Toronto, UCLA, Oxford, Brown, Butler, Cleveland Clinic, Mayo Clinic, Stanford, Johns Hopkins, Bonn, Magdeburg, Amsterdam, etc.) can work together to develop appropriate models for all the disorders and choose precise target locations. Then they'd need to collaborate with the brain decoding crowd, BCI developers, and William Gibson.


Footnotes

¹ Stanford professor and working group co-chair Bill Newsome said:

"The government shutdown will very definitely affect BRAIN--will bring it to a complete halt in fact. To write good proposals, to get them evaluated, to get the money committed for this next year flowing, that’s a long process--even with the NIH process moving at warp speed, it takes the better part of a year. We on the working group, we delivered our end of the bargain. NIH wants to deliver on its end of the bargain, but they simply can’t do it if they’re sitting at home on an unwanted furlough."

² Plus $20 million to NSF (Samuel et al., 2013) and over $120 million from private foundations.

³ I hate to be a wet blanket, but don't see how the entire DBS research community could achieve some of the Phase 2 goals within 5 years. I'm not alone in this. Dr. Helen Mayberg, one of the researchers most qualified to submit a proposal to DARPA, said:

“Is it overambitious? Of course,” said Dr. Mayberg, adding that working with the brain is “a slow process.” But she said that it was an impressive first investment and that the clear emphasis on human illness was “stunning.”

⁴ One section of the report mentions Devices for Monitoring and Stimulating the Human Brain, but mostly in the context of recruiting patients as research participants (not in developing technologies or treatments... which is covered by DARPA).


Further Reading

Agency Initiative Will Focus on Advancing Deep Brain Stimulation

BRAIN Initiative Interim Report: A Readers Guide

BRAIN Initiative Links at Empirical Planet

BAM and BRAINI links at Nucleus Ambiguous

From BAM to The BRAIN Initiative: A clearer view of a major neuroscience enterprise

Neuroscience thinks big (and collaboratively).

Scientific priorities for the BRAIN Initiative.

The challenge of connecting the dots in the B.R.A.I.N.


References

Alivisatos AP, Chun M, Church GM, Greenspan RJ, Roukes ML, & Yuste R (2012). The brain activity map project and the challenge of functional connectomics. Neuron, 74 (6), 970-4 PMID: 22726828

Insel TR, Landis SC, & Collins FS (2013). Research priorities. The NIH BRAIN Initiative. Science, 340 (6133), 687-8 PMID: 23661744