Technology is the backbone of the National Institutes of Health’s (NIH) Brain Research Through Advancing Innovative Neurotechnologies® Initiative, or The BRAIN Initiative®. In a virtuous cycle of technology and knowledge, new tools open doors for discovery, which in turn drive the design of new and better tools. New advances in human-centered brain technology offer hope for cures sooner than later.
Deep brain stimulation (DBS) is a case in point: a tried-and-true technology success story that’s been around for decades for some conditions, like Parkinson’s disease and epilepsy. Thanks to DBS-based treatments, affected individuals have regained control of previously uncontrollable body movements like tremors and seizures. Although we’re still in the early days of applying DBS to other conditions, there’s a lot to be excited about for the DBS-based treatments currently being tested for obsessive-compulsive disorder, binge eating disorder, post-traumatic stress disorder, substance use disorder, and depression. This research provides real hope for people living with challenging long-term conditions.
Last month, another groundbreaking study came out—researchers identified a pattern of brain activity, or “biomarker,” related to clinical signs of recovery from treatment-resistant depression. As reported in this study, after six months of targeted DBS, 90% of research participants showed a significant improvement in depression symptoms, and 70% were no longer depressed. What’s more, the biomarker predicted depression symptoms more than a month before symptoms surfaced. This work comes from a BRAIN Initiative-funded team led by Dr. Helen Mayberg (a neurologist at the Icahn School of Medicine at Mount Sinai in New York City) and Dr. Christopher Rozell (a computer engineer at Georgia Institute of Technology in Atlanta). It’s notable that the research team includes a mix of disciplines working together: computer science, psychiatry, and neurosurgery. This collaboration embodies the ethos of the BRAIN Initiative: doing better science through diversity of thought and experience.
In a remarkable series of recent advances, artificial intelligence (AI) has played a starring role in the ongoing clinical improvement of DBS, including in this new work, by identifying clinically relevant activity “signatures” from huge volumes of data captured from brain activity, imaging, and other measurements. Since Mayberg's DBS experiments on treatment-resistant depression about two decades ago, we’ve also seen improvements in hardware and software, which collectively make stimulation more accurate and allow a longer battery life for implants.
Quantifying Brain Data
Determining and adjusting an effective “dose” for a brain-based stimulation treatment is anything but straightforward. Complicating matters even further, neuropsychiatric disorders like depression are caused or worsened by a range of factors such as biology, social environment, and other external influences that can vary over time or circumstance. They are also notoriously difficult to measure. Self-reports from patient interviews vary a lot from person to person. That makes it tough to know what is a normal, transient mood variation and what is a more serious problem requiring attention or intervention. It’s also challenging to know whether and when treatment is having a measurable effect.
Capturing objective data can help fine-tune technologies like DBS to patient-specific needs. For example, an individual’s appearance and behavior are always key components for disease diagnosis and progression. The Mayberg and Rozell study’s AI algorithm tracked changes in facial expression to help differentiate between short-lived, resolvable mood shifts and authentic depressive symptoms. Analyses of brain imaging data allowed the team to correlate longer recovery times with the amount of damage to depression-related brain regions.
Observing and quantifying patient-specific responses to DBS and other therapies will one day allow precise health adjustments guided by biomarkers. Earlier this year, for example, I wrote about separate BRAIN Initiative-funded research that is taking a similar approach and has identified a chronic pain biomarker that is being applied to ongoing patient-centered research.
The Smallest Sample Size
You’re likely noticing that the sample sizes in the research I’ve been talking about seem impossibly small—a handful of research participants or even just a single individual in some studies. Yes, that is true. Collectively, though, they reflect a new paradigm for scientific investigation. Instead of starting with a one-size-fits-most treatment and testing it in thousands of people, precision medicine strategies start with one person, gaining a readout and targeting treatment, and setting the stage for randomized trials to demonstrate efficacy before broader deployment.
It’s important to note that the outcome of patient-centered research won’t apply to only one person. Rather, by gathering millions of brain-computer interactions and using AI algorithms to sort and identify actionable patterns from the data, we are generating a set of clinically valid features that can be applied to many individual humans. The ability to objectively measure disease-related brain activity via biomarkers is a game-changer for brain-based clinical research.
It will be a major challenge to scale up invasive and expensive treatment technologies like DBS into routine clinical practice for neuropsychiatric disorders. That integration into clinical care will take time and ingenuity. But again, we can look to the power of technology to help. It may be, for example, that AI can also be used to further refine brain biomarkers to the extent that scalp-based, non-invasive brain stimulation becomes accurate enough to eliminate the need for neurosurgery altogether. That type of treatment will also take time, but you never know how and when technology will surprise us.
And while the science proceeds at an accelerating pace, we’ll need to establish guidelines and guardrails to protect the privacy and agency of the individuals from whom we are collecting these precious data. I will continue to explore the rapid pace and impact of human-centered neurotechnology in my next post. Without question, we need an intentional approach toward involving the human voice and experience in the evolution of brain-machine interactions. Stay tuned!
With respect and gratitude,
John Ngai, Ph.D.
Director, NIH BRAIN Initiative