This Clinician’s View is written by AdventHealth neurophysiologist and Scientific Director of the MEG Laboratory Eduardo Martinez Castillo, PhD.
Despite decades of demonstrated clinical value, magnetoencephalography (MEG), a noninvasive neurophysiological imaging technique, remains available in only about 25 centers across the United States. While its clinical use has focused on presurgical mapping in epilepsy and brain lesions, advances in artificial intelligence and data analysis are enhancing how we interpret complex brain activity, and the role of MEG in clinical decision-making is evolving.
At AdventHealth, our focus extends beyond simply providing access to this technology. We are equally committed to how MEG is applied in practice, integrating it with other imaging modalities and emerging analytical approaches to better inform patient care and improve outcomes.
By enabling real-time mapping of brain function, MEG is informing more precise, personalized and data-driven decisions around surgical planning and treatment selection. Just as importantly, it can reduce or even eliminate the need for more invasive diagnostic procedures, lessening risk and easing the burden on patients and their families.
At AdventHealth, we opened Florida’s first MEG lab in 2013 as part of an ambitious initiative in neurosciences led by what is now AdventHealth for Children, our pediatric hospital. Since then, we have expanded access to include adult patients with epilepsy as well as individuals with brain lesions, including vascular malformations. While the core MEG technology has remained relatively constant since 2013, its clinical applications and the value it delivers to patients have continued to grow.
Understanding the Power and Advantages of MEG
At its core, MEG captures the magnetic fields generated by neuronal electrical activity in the brain. This allows us to directly observe function and accurately identify abnormal activity, such as epileptic discharges. It also plays a critical role in presurgical planning by mapping essential functions like language, movement and vision.
With this information, we can help surgeons design approaches that spare critical functions, minimizing the risk of functional morbidity, including patients’ ability to speak, move and engage with the world, while still effectively treating the underlying condition. In many cases, MEG can reduce or even eliminate the need for more invasive procedures such as intracranial EEG monitoring and the Wada test, lowering risk and minimizing hospitalization.
In addition to being completely noninvasive with no radiation exposure, MEG offers several key advantages:
- Direct measurement of cortical brain activity
- Millisecond-level temporal resolution
- Precise localization of function-specific brain regions
- Signal of interest unaffected by volume conduction, increasing the accuracy of localization and the sensitivity to cortical activity in comparison to other modalities
Perhaps most importantly, MEG allows us to understand brain function at the individual level. No two brains are exactly alike, nor do they respond similarly to disease or treatment, and this level of insight helps tailor care in a way that reflects each patient’s unique anatomy and condition.
Advancing the Use of MEG to Deliver More Patient-Centered Care
Throughout my decades of working with the MEG, how we’ve utilized the data it provides has drastically evolved. Following my neuroscience training in Madrid, Spain, I moved to the University of Texas (UT) Health Science Center at Houston to complete a fellowship at one of the field’s pioneering centers, the MEG Laboratory at UT-Houston Medical School led by Dr. Andrew Papanicolaou. Under his mentorship, I had the chance to participate in studies that helped move MEG from being a research tool to being a clinically valid instrument at the forefront of advancing neuroimaging.
At that time, we needed to understand which aspects of brain activity could be reliably characterized in clinical populations and which metrics were clinically valid. The clinical value of MEG was already evident from early studies where we compared the localization accuracy of MEG to invasive video electroencephalography (VEEG) for localizing seizure onset zones in patients undergoing evaluation for epilepsy surgery. We found that MEG offered equivalent accuracy, an important step in validating it as a less invasive alternative.
Over the years, I had the opportunity to contribute to several multicenter studies funded by the National Institutes of Health (NIH), National Science Foundation (NSF) and Department of Defense. Collectively, this work helped establish MEG as a valuable clinical tool, particularly in presurgical planning for epilepsy and brain tumors.
As these protocols were tested and validated, we were able to better describe the impact of MEG in patient care. It has helped to reduce reliance on invasive procedures and has also accelerated evaluation and treatment selection in complex cases of refractory epilepsy.
The Power of Collaboration and Artificial Intelligence to Expand MEG’s Capabilities
While MEG has always offered exceptional temporal resolution, recent advances, including network-level analysis, multimodal integration and improved source modeling, have expanded its role well beyond localization. Today, it functions as a comprehensive functional brain mapping system. Clinically, this means we can track when and where functions like language processing unfold with greater accuracy, even when the brain’s anatomy is distorted by lesions. The fact that MEG is completely non-invasive and silent makes it possible to now study how brain functional maps change over time in pediatric populations. This has opened the door to address critical questions like how language representation changes due to brain lesions or recurrent seizures.
Emerging artificial intelligence (AI) and data-driven analysis are accelerating this progress. Machine learning tools can now assist in detecting epileptiform activity and modeling abnormal brain networks with increasing precision. In many ways, AI has always been part of MEG’s evolution as we’ve long relied on sophisticated algorithms to interpret complex datasets and support diagnostic and treatment decisions. What’s different now is the scale and depth of insight these tools can provide, helping us better understand patterns and relationships that would otherwise be difficult to detect.
At AdventHealth, we’ve already partnered with researchers at the University of Central Florida (UCF) and together, have proven the benefits of incorporating artificial intelligence algorithms to fine tune our protocols for language mapping in patients with epilepsy.
However, more extensive AI development requires availability of large datasets. Given the limited number of MEG scanners with patient data to fuel its development, additional collaboration remains essential to unlocking MEG’s full potential.
Looking ahead, we aim to expand research partnerships, including with groups like the ENIGMA Consortium, which brings together researchers in imaging genomics, neurology and psychiatry from around the globe. Efforts such as ENIGMA-MEG are examining how brain activity patterns vary across populations, linking imaging data with other clinical findings. For patients, this means more evidence-based care driven by data that no single center could generate alone.
How MEG is Forging New Frontiers in Patient Care
Advances in MEG are starting to enable a more precise, individualized approach to care that benefits patients in several ways:
- Mapping language function in younger children with epilepsy and other conditions
- Informing surgical planning and predicting functional recovery
- Identifying brain phenotypes to guide treatment selection
In pediatric epilepsy, MEG is addressing a long-standing challenge. We previously had limited insight into how language networks are organized in young children. As a result, surgical decisions often relied on highly invasive procedures and/or generalized assumptions such as left-hemisphere dominance. However, epilepsy can significantly reorganize a child’s developing brain, and with MEG, we can noninvasively map each child’s unique language network, helping guide surgical strategies that minimize the risk of functional deficits and better determine who may benefit from surgery versus non-surgical treatment or additional diagnostic evaluation.
MEG is also helping shift care from reactive to predictive. Historically, we’ve had few noninvasive tools to assess which brain regions might support functional recovery after surgery. Now, advanced MEG-derived metrics allow us to estimate the likelihood of functional improvement. For example, in a child with language decline due to ongoing seizures, we can map both the language network and the epileptic activity. This helps us evaluate whether there is partial overlap between networks and if removing the seizure focus may not only control seizures but also enable recovery of lost function. These are new pieces of information that we are still testing to validate against standard approaches, but we are starting to add information that facilitates the predictions on whether a functional recovery will happen following the surgical treatment.
Finally, MEG is currently being explored as a tool for predicting treatment response. Even within the same diagnosis, an individual’s brain connectivity pattern is patient-specific, like a fingerprint, and can be considered the patient’s brain phenotype, resulting from the interaction of its genotype and the environment. So, it makes sense to investigate whether specific features present in the patient’s brain activity can predict treatment response (i.e., predictive biomarkers) instead of treating all patients with the same diagnosis identically. In principle, once the biomarkers are identified and validated, MEG data could support the use of treatments tailored to the neural physiology rather than generic diagnostic categories. These are questions we hope to answer in the coming years. The more we investigate the volumes of data we derive from the MEG scanner, the more subtle differences we are identifying.
Although MEG does not directly measure neurotransmitters, it provides insight into the brain’s balance of inhibitory and excitatory activity through measurements of neural activity and brain network connectivity. In epilepsy, this excitatory/inhibitory balance is broken. Sometimes the disruption is local, affecting limited areas, and sometimes it is distributed across extended networks. Having a clear, patient-specific picture of the affected network can help to plan the most effective therapy with the goal of regaining this excitatory/inhibitory balance. This perspective adds to the effort of helping patients reach effective therapy sooner, avoiding unnecessary side effects and ultimately, achieving better outcomes through more personalized care.
To fully realize these advances, MEG cannot function in isolation. Its greatest clinical value emerges when it is integrated into a broader, collaborative framework of care that brings together multiple disciplines, imaging modalities and perspectives to form a more complete understanding of each patient’s condition.
Interdisciplinary Collaboration to Provide Comprehensive Evaluation
At AdventHealth, MEG is part of a comprehensive, integrated diagnostic approach that can include other functional and structural neuroimaging modalities as needed, including MRI/fMRI, PET/SPECT, and EEG. Each modality provides a different perspective on brain structure and/or function. In complex cases, it is the converging evidence of these data that allows clinicians to validate diagnostic hypotheses and customize treatment plans.
When it comes to the MEG lab specifically, multiple clinicians, technologists and neuroscientists are involved in its success, and I am especially grateful for the contributions of the clinical teams from the Pediatric Outpatient Procedures and Sedation Unit, Anesthesiology, and Special Imaging/Radiology as well as the following individuals:
- Billie Pullum, MHA, BSN, RN, Executive Director, Clinical & Support Operations
- Michelle Curtier, Epilepsy/MEG Program Coordinator
- Evelyn Hernandez, MEG/EEG Technologist
- Po-Ching Chen, PhD, Biomedical Scientist
- Elakkat Dharmaraj Gireesh, MD, PhD, Neurologist
- Ammar Hussain, MD, Neurologist
- Holly Skinner, DO, Neurologist
- Angel Claudio, MD, Neurologist
- Julia Henry, MD, Neurologist
Aligning with the Future of Brain Health
After decades of clinical validation, expanding applications and deeper integration with advanced analytics, MEG is no longer simply an emerging technology. In fact, it is becoming an essential component of how we understand and treat complex neurological conditions.
For years, much of our best guidance for neurological care came from invasive imaging modalities. Today, MEG is helping us get there noninvasively, and I believe we are just scratching the surface of its capabilities.
As AI advances, it will help us uncover new applications for MEG, facilitating clinically meaningful insights that can guide faster, more precise care. The MEG technology itself is evolving as well. Next-generation sensors promise greater sensitivity and perhaps even wearable devices that will bring brain mapping into patients’ daily lives at home, making it easier to capture real-world data without the burden of hospital visits.
Ongoing research is also expanding where MEG can potentially make a difference, with emerging studies in conditions like autism, dyslexia and Alzheimer’s disease.
I am excited to see how MEG will continue to provide us with a more complete picture of brain health, supporting earlier diagnosis, more precise treatment and truly individualized care for each patient.
