Real-time neural diagnostics and data-enabled neurotechnologies are here. Welcome to the world of theranostic neurotechnology.

Neuromodulation/BCI devices will soon not only treat disease but reveal the inner workings of biology in real-time—merging diagnosis with therapy. This intersection gives way to a new class of device-based ‘theranostics,’ technologies that can treat diseases while offering companion diagnostics and previously inaccessible biomarkers.

Neurotechnology devices of the past have been placed into one of two categories. First is those that ‘read’ i.e. typical motor and speech interfaces that connect humans with technology, move bionic arms, or play games via computer interaction. Second is those that ‘write,’ typically neuromodulation devices which preceded commercial BCI efforts by many decades and could be used to modify some electrically driven disease or disorder. Well-known examples of ‘write’ technologies are Deep Brain Stimulation (DBS) for Parkinsons or epilepsy management, Cochlear implants for restoring hearing, or spinal cord stimulation (SCS) for pain management. Data and sensing enabled therapeutic neurotechnology is about to kickstart a tech-med revolution in healthcare. We’re coining it ‘theranostic neurotechnology’. Theranostic neurotech devices feature both therapeutic ‘write’ capability alongside companion, real-time diagnostic and biomarker readouts.

Many diseases have unique electrical 'signatures’—some cell or circuit level function that can be measured electrically. By implanting a device, we can switch on recording functionality and extract real-time data to investigate trends and patterns during disease progression, awake/sleep cycles, adjacent and complementary treatment approaches, and practically anything else that a human being does in their day when not sat inside of an MRI scanner. That is to say, we’ve never seen this kind of active neural data before.  

And this is critical because biology has a data problem. As it stands today, most biological data is inconsistent, often scattered across many small datasets generated by artisan biologists. Poor control is already driving a reproducibility crisis in the field, but to slide this into the current tech narrative, it makes most bio data unsuitable for training AI. On the other hand, our never-before-seen longitudinal recordings allow us to train completely new AI models of biology. The impact we’ll have on the wider healthcare ecosystem is unprecedented.

The first generation of theranostics are decades in the making  

While read/write neurotechnologies are no longer science fiction, they’re not exactly new either. In contrast to common perceptions, the BCI titans of today (most think Neuralink) did not lead the charge. Instead, neurotech incumbents (Medtronic, Boston, and Neuropace) forged the first generation of theranostic platforms. It’s just that due to the long commercialization timelines, we’re only now beginning to see these devices out in the world.  

A recent article published in nature reviews speaks to the convergence of BCI and ‘conventional’ Neuromodulation devices. Though there remains a disconnect at this nexus. BCI companies are still searching for therapeutic markets to expand into, and conventional neuromodulation devices, though approved and well characterized for specific indications, have limited sensing and processing capability.

Closed-loop vs diagnostic devices

A closed loop system is automated to self-optimize, either removing or reducing clinician input to enhance or speed up efficacy and/or reduce side effects. Though imaging technologies do not necessarily need to be ‘closed-loop’ to be useful. For example, an MRI is already a powerful and lucrative tool which has saved and extended millions of lives. MRIs achieve this without also controlling the radiotherapy dose or holding a scalpel.  

But taking it one step further, an MRI-like device that also provides therapy is a rather alluring proposition. Devices that provide us with weekly reports of biological drift, disease biomarkers, and other real-world data will become some of the most powerful ‘imaging’ devices we’ve ever seen. We expect these closed loop to be as transformative as the MRI itself.

A scanner in the brain

A second generation of neural interfaces are coming. Speak to enough brain-computer-interface builders and you’ll observe that almost every company has a closed-loop disease strategy; some more baked than others. We are truly seeing the convergence of BCI and medical neurotech.

There are two areas where theranostic platforms will have high impact:  

1. In diseases where empirical measures of disease progression/recurrence/treatment response do not exist, are not trustworthy, or cannot be captured at sufficiently high frequency to unlock rapid intervention

2. Areas where insights can integrate with the healthcare system at large, e.g. by reducing the frequency of hospital scans, remotely monitoring disease state, tracking objective symptoms (as opposed to relying on patient testimonies), strengthening ‘gold standard’ imaging techniques, or improving the precision and efficacy of complementary (e.g. drug) therapeutic approaches  

‍

The multi-billion-dollar healthcare questions are where will impact happen first, and where will the largest benefits to human health reside? What will be the breakthrough application?

There are clear targets in psychological health; OCD, depression, Alzheimer’s/neurodegenerative disorders, brain injury, stoke, and oncology. All areas in which we have a grainy understanding of disease progression or recovery, where we lack clear, objective markers of disease severity, and where neural/electrical recordings can act as new biomarkers for disease. But these are also areas where there is a massive disease burden to both the patient and the healthcare system itself—a burden which this second generation of neurotechnologies is set to remedy.

Novel technology brings novel ethical questions  

While second generation neurotechnologies are set to dramatically improve and extend human life, questions around data privacy, ethics, and clinician intervention naturally arise. On the clinician side, delegating treatment to an algorithm is a significant change in how doctors practice. Even if it is more effective, one could see parallels to current debates around autonomous vehicles. To what extent clinicians should be in the loop remains an open question. What access should we provide and how should we present high-density, high frequency data to a practitioner?  

Another open question concerns user access to data. How accessible should the data be to the patient? How much control over their devices can, or should, they have? Neuro-ethics and neuro-data rights are both topical areas of debate. While it is true that my watch, car, and phone report way more data back to server rooms than I’m comfortable with, I do still have the agency to remove these items from my life without a medical procedure.  

Broader technical questions encompass the tools, data, and trials we need to validate our neural insights. In my view, we need more data across more timepoints; continuous and consistently collected data is ideal. We also need greater collaboration between those producing the data and all other touch points of the healthcare ecosystem. From clinical trial access to drug companies, to software layers, to the newest healthcare models.

Final thoughts

Closing the loop on neuromodulation and implantable devices is exciting and inevitable. I’m sure that we will see improvements to neuromodulation when we can optimize its delivery against a relevant biomarker [saluda!], but I am even more bullish on on-board diagnostic and data capabilities that will provide a scale, temporal resolution and long-term evolutionary understanding of disease in the near term.

Crunching this data with advanced reasoning models and integrating these with supplementary biological data sets from a more wholistic healthcare infrastructure will drive dramatic improvements in healthcare. More people will move from being treated in the hospital to their home; reducing costs, enabling personalized treatment, and reducing clinical trial failures. The implants of the future will be tiny diagnostic labs worn inside the body, constantly monitoring and reporting biological trends. We will treat and report on illness from within.

Where Coherence fits into the future  

The Coherence platform was built with data-led intelligence at its core. We have a unique insight into the electrophysiology of cancer, and it is where we can have the largest impact on human health at large. After all, cancer is the second leading cause of death worldwide. This is why we’ve built smart electrotherapy systems that don’t just treat disease, but measure the electrical language of cancer itself. Using this previously inaccessible language, we can predict drug responsiveness, record symptoms, predict growth-rates, and track how cancer spreads. Ideally, we do all this so that we can prevent cancer spreading in the first place.  

In 2024, from electrical cues alone, we successfully trained a cancer prediction and classification model to identify glioma progression, presence, and severity. We are the first in the world to do this, with a publication expected later this year. A new paradigm of companion, device-based theranostics is just around the corner—Coherence is leading the charge.

-- Dr Ben Woodington, co-founder & CEO Coherence Neuro