Time for Brain Medicine

When people have a medical problem with their heart, they see a cardiologist. Lungs? Pulmonologist. Kidneys? Nephrologist. Brain? It’s complicated. If it’s a behavioral problem, they are referred to a psychiatrist—unless it’s degenerative, then they are referred to a behavioral neurologist. Motor or sensory issues require a neurologist, unless the etiology is not “organic,” such as catatonia or a “functional” tremor. Such divisions are long-standing and date back to the late 1800s when pioneering neurologists Jean-Martin Charcot and his student Sigmund Freud separated “hysteric” (i.e., “functional”) etiologies from “organic” disease (i.e., detected by instrumentation) (1, 2). Perhaps the most accurate way to differentiate neurology from psychiatry is “structure” versus “function,” although no definition exists that can be universally applied. Clinically, there are no defining rules to divide brain disease into neurological versus psychiatric (e.g., destruction versus dysfunction or sensorimotor versus behavioral). So why do physicians practice this way? More importantly, what happens when a patient has brain disease that crosses this imaginary boundary? How many diagnoses have been missed or harmfully delayed? How often is the best treatment outside of our scope of specialty training because that treatment is used to treat another aspect of the brain? In essence, how can we know what we don’t know?

We authors suggest that the current segregated training in neurology or psychiatry is not sufficient, and we propose a new paradigm combining the skill sets of those trained in neurology and those trained in psychiatry (3). This new program would promote a broad scope of training that integrates the brain sciences in the first years, with a capacity for subspecialized depth of training in the later years, without increasing the overall length of training. We call this new paradigm “brain medicine.”

Now Is the Time for Brain Medicine

As physicians, we are witnesses to the greatest neuroscientific era in history. The treatment of brain disease is transforming. With the same electroencephalography that long since changed epilepsy management, we can now use cutting-edge analytics to detect unique patterns in disorders ranging from concussion to various childhood psychiatric problems (4, 5). Structural brain imaging, which revolutionized stroke management, can now be used to identify network-based subtypes of depression that are amenable to distinctly targeted brain stimulation treatments (6, 7). Such examples of emerging management approaches for brain disorders number in the hundreds or even thousands and span levels of analysis from the genome to networks (8–11). Yet, despite advances in knowledge and technology, rigid clinical structures remain stuck in the past and isolate these two medical approaches to the brain. Symptom-based diagnoses created by clinicians provide a useful “language,” but they do not map onto the underlying neurobiology. This creates problems for neuroscientists trying to understand diseases with heterogeneous etiologies and localizations.

The National Institute of Mental Health responded to this problem by creating the Research Domain Criteria initiative. They expect that “identifying syndromes based on pathophysiology will eventually be able to improve outcomes” (12). This expectation is based on the assumption that “dysfunction in neural circuits can be identified with…electrophysiology (and) functional neuroimaging” (12). But do psychiatrists have the necessary training to implement and interpret these tests? As new diagnostic breakthroughs uncover the biological bases of psychiatric disorders (which then effectively become “organic” etiologies; e.g., N-methyl-d-aspartate receptor encephalitis), will neurologists have the necessary training to care for these patients? The skill sets of neurologists and psychiatrists are complementary. Both are essential to best diagnose and treat the complex disorders of the brain.

Being a skilled diagnostician, like an effective laboratory test, requires sensitivity and specificity. Understanding both the neurological and psychiatric aspects of the brain enhances diagnostic specificity (e.g., “I know this is not new-onset psychosis because I know this is normal pressure hydrocephalus”). Being trained in one discipline is like having a hammer, where each patient presentation is either a nail or it is not. Having a screwdriver available can enable new diagnostic specificity, not only by recognizing that the hammer isn’t working, but because the screwdriver is working. Of course, manifestations of brain disorders are not so clear cut. Therefore, the ability for one individual physician to have both perspectives is the key.

Combined neurology-psychiatry residency programs in the United States are 6 years long and currently average only four graduates annually. However, our contention is that the high prevalence of comorbid conditions and overlapping treatments indicate that the majority of neurological and psychiatric patients would benefit from the care of someone with combined training. Academic tertiary care centers are accustomed to patients with complicated and comorbid conditions. One can anticipate that patients with complicated conditions will become more prevalent in community private practice. The role of nonphysician providers can be expected to expand because these individuals meet a crucial need in providing care to a growing patient base. The practice of these providers will largely involve more common diagnoses, that is, the “horses.” Consequently, the less common diagnoses, the “zebras,” will require physician consultation or referral, which means a greater proportion of uncommon presentations. We authors anticipate an inevitable shift in the complexity of patients seen by residency-trained physicians in all settings and emphasize the need to prepare for this shift. A thorough development of the argument for the need for training in all aspects of the brain is beyond the scope of this article, but we refer the skeptical reader to several references for further review (13–16).

Challenges of Breadth, Depth, and Feasibility

Here, we have highlighted the need for breadth of training across neurology and psychiatry. However, we also consider it self-evident that subspecialization will become increasingly necessary as neuroscientific advances continue to expand the essential knowledge base of genes, biochemical pathways, circuits, biomarkers, and treatments. These advances will translate into an increasing need to interpret findings with approaches, such as advanced imaging, genomics, proteomics, and pharmacogenetics, in the context of classical neurological and psychiatric symptoms. From this context, precision medicine may emerge and involve, for example, assigning dorsomedial prefrontal cortex transcranial magnetic stimulation treatment to those with connectivity-based anxious biotypes of depression (6, 7) or prescribing particular medications guided by genetics or proteomics (17).

Meanwhile, additional years in residency (6 years for existing combined programs) are not fiscally or temporally feasible. Prolonging the time spent as an underpaid resident, without the promise of a corresponding increase in salary after residency, would likely deter many candidates. Finally, any solution that addresses the above challenges must be amenable to realistic integration with the current structure of clinical practice and departmental organization. How can these challenges be overcome to harness the potential of neuroscientific discoveries for the benefit of brain medicine patients?

Training in Brain Medicine

We propose a novel encompassing approach to clinical training with respect to structure and function of the brain. Rather than training almost exclusively in a field that addresses only particular aspects of brain disease (Figure 1A-B), we propose a foundation of training across multiple facets of brain disease that erases the barrier between psychiatry and neurology (Figure 1C-D). Importantly, this approach does not undo the development of any specialty or field but rather builds upon this underlying foundation. We will compare and contrast this approach to the current paradigm, as shown in Figure 1.

A brief description of current residency curricula in neurology and psychiatry, by year, is provided here as a frame of reference for the proposed changes.

Year 1.

Neurology programs today include at least 8 months of general medicine—it is actually 1 year, but many programs include up to 4 months of electives that can include neurology. Psychiatry programs typically include 6 months of general medicine and 6 months of core psychiatry rotations.

Year 2.

Neurology programs are primarily dedicated to inpatient requirements. In years 2–4, a 3-year longitudinal continuity clinic is held one half-day per week. Most programs include the single month of required psychiatry during this year. Psychiatry programs are also primarily inpatient (some programs do this in year 3) and typically include the required 2 months of neurology during this year.

Years 3–4.

“Senior resident” status is obtained in both neurology and psychiatry: the resident is now considered capable of supervising students and junior residents while on call and under the direction of the attending physician while on inpatient services. Required outpatient rotations, along with continuity clinics, are completed. Often, a year of electives can be compiled between these 2 years.

What Is Similar Between the Existing and Proposed Residency Programs?

The following features are common to existing programs and the proposed program:

• The duration of both programs is 4 years.

• Both programs include a general medicine requirement.

• The first 2 years are inpatient predominant and lead to “senior resident” status.

• The last 2 years include outpatient and elective rotations targeted toward residents’ interests and goals.

• Brain medicine residents would fulfill the required rotations for both psychiatry and neurology put forth by the Accreditation Council for Graduate Medical Education (ACGME) (18, 19).

• Goals and competencies for both psychiatry and neurology put forth by the ACGME would be fulfilled (18, 19).

• Continuity clinics for both specialties would be included.

• Options to pursue fellowship training would be available.

• Graduates would become board eligible, with the ability to practice independently immediately after graduation.

• Graduates would be eligible to practice in all of the specialties and subspecialties that are currently available.

• Board certification through the American Board of Psychiatry and Neurology (ABPN) would remain the same.

What Is Different Between the Existing and Proposed Residency Programs?

The proposed program would differ from existing programs in several ways.

Reduced general medicine requirements.

Rotations would be specific to the educational needs of a brain medicine resident and independent of internal medicine staffing needs. For neurologists and psychiatrists, general medicine is important for learning how to function as a physician. Moreover, it is essential to understand the relationship between the body and the brain to contextualize brain disease within the framework of general medicine. However, we contend that greater yield would result from more than 1 or 2 months of psychiatry and neurology for a neurologist and a psychiatrist, respectively. Reduced exposure to general medicine is not a radical proposition—few specialties still hold to the traditional precedent of a full intern year of general medicine, which is the equivalent of a fifth year of medical school. To meet our objectives, we propose four selected medicine rotations (recommendations provided in Figure 2). In our experience, additional months beyond these provide diminishing returns.

Inpatient neurology and psychiatry.

The first 2 years would include the inpatient requirements of both neurology and psychiatry, rather than the requirements of only one specialty (compare top panels to bottom panels in Figure 1).

Subspecialty tracks.

Training in the last 2 years would be more individualized. Rather than rotating through a broad range of clinics, which are largely determined by each institution and not by the ACGME, residents would have tracks that they could select. Tracks may be subspecialties (an example is shown in Figure 1C) or general (neurology or psychiatry) (an example is shown in Figure 1D). A general brain medicine track would also become available.

Brain medicine boards.

Board eligibility and board certification in brain medicine would encompass psychiatry and neurology.

Fellowships.

It is beyond the aim of this article to determine board-certification requirements of specialty boards. However, we do not think that it is beyond reason that some boards might consider a track sufficient to obtain eligibility. In any case, trainees will be well prepared for their subspecialty fellowship training.

Exposure rotations.

Year 1 would include exposure rotations (one half-day per week) to help identify tracks of interest (some programs already have versions of these rotations) (green bars in Figure 2B).

Summary of Brain Medicine Residency Curriculum

The first 2 years will be used to build a foundation of competency involving many aspects of brain medicine and disease (Figure 2A). Brain medicine residents would become capable of diagnosing and managing common and acute neurological and psychiatric conditions through traditional inpatient rotations (Figure 2B). The last 2 years would prepare trainees for their discipline of choice through tracks. These tracks may be specific subspecialties, such as addiction medicine, epilepsy, movement disorders, or interventional psychiatry. These tracks may also be more generalized, such as general neurology or general psychiatry, or provide the novel ability to train broadly across all of brain medicine. Some specialties are beyond the scope of a 4-year training program, such as child psychiatry or neuromuscular medicine, and will require additional fellowship training, just as they do today. Other fellowships would still be required unless subspecialty boards accepted training from a track. Of course, employers could hire subspecialists from a track even if these individuals are not board eligible, just as employers can today. A sample schedule of what a 4-year brain medicine residency might look like is shown in Figure 2B.

A New Name

“Brain medicine” is the practice of medicine involving disease or dysfunction of the brain. Brain medicine encompasses neurology and psychiatry. The name distinguishes itself from surgery and from the practice of medicine in the rest of the body. It is a simple, clear, intuitive label for a particular area of practice. In our experience, other encompassing names, such as “neuropsychiatry,” tend to confuse both the medical and nonmedical community alike. Moreover, neuropsychiatry is commonly misattributed to neuropsychology, a nonmedical field. While “psychiatry” and “neurology” will continue to describe clinical practices and departments, a new name places emphasis on the biological basis of function and behavior. This may even help destigmatize “psychiatry.” In so doing, these types of physicians would join internal medicine, family medicine, emergency medicine, and other “medicine” specialists. In short, we propose brain medicine as the name for a new residency program to not only cross the traditional boundaries of neurology and psychiatry but also encompass and unify these disciplines.

Discussion and Limitations

The neuroscientific revolution is erasing the line drawn between disorders of structure and function. Medical fields were defined on the basis of what could be seen with the naked eye or with a microscope at autopsy. With advanced brain imaging, neurophysiological techniques, sophisticated animal models, and cell cultures, it has become apparent that all brain disorders have their basis in brain structure and that all brain disorders affect brain function. However, clinical training remains stuck behind these boundaries. In response, we propose a novel residency program to address the growing need to provide training that captures the multifaceted nature of the brain—with flexibility for early subspecialization to keep up with advances—without increasing the duration of training. This is accomplished through foundational training in brain medicine during the first 2 years, followed by 2 years of subspecialization or more extensive generalization (Figure 1C–D).

Although we are proposing this as a novel training paradigm, it is actually not new at all. Internal medicine residents become board eligible after 3 years of training across organ systems, followed by optional subspecialization. The proposed brain medicine training program would mirror this paradigm, except that trainees would complete all 4 years before board eligibility. We will now address several notable challenges and limitations.

Competency in Only 4 Years

Time in training is a precious commodity. In an era of reduced resident work hours, bloated electronic medical records, increasing regulatory requirements, and expanding medical knowledge and subspecialties, is it possible to train in two specialties without increasing the duration of training? Even our current single-specialty system grapples with attaining proficiency in the face of time limitations (20, 21). How, then, can one possibly become competent in two specialties in only 4 years? We contend that it can be done through reprioritization. The brain medicine concept is not an accelerated program for the most gifted, nor is it a shortcut. Current combined neurology and psychiatry programs (in which all of the authors trained) reduce the combined 8 years of training into 6 years through overlapping shared requirements and not duplicating the intern year. In the proposed brain medicine program, as illustrated in Figure 1, several months of general medicine in the first year are reallocated to brain medicine. Furthermore, once a senior-resident level of competency is attained, various rotations less critical to one’s chosen field are reallocated and dedicated toward one’s chosen subspecialty (Figure 1).

We concede that achieving expertise in all brain subspecialties in 4 years is not possible. We contend that because subspecialists are increasingly practicing only within their subspecialty, it is not necessary to spend precious months of training in less relevant subspecialties. Even now, there are more subspecialties than there are months to rotate through them. The trend toward subspecialization is a product of the incredible expansion of knowledge in this last frontier of neuroscience. As a result, a misguided goal to maintain a traditionally accepted ratio of individual knowledge to available knowledge in one’s specialty (as a metric of competence) established 80 years ago could arguably result in a 30-year program today. It is therefore an untenable position to strive for the same depth across the entire breadth without continuously increasing training duration. Rather, creative solutions to the already restrictive time limitations must be conceived and implemented (22).

In current practice, a stroke specialist typically does not manage patients with new seizure onset or read the electroencephalogram, but rather he or she consults the epileptologist. After years of practicing only within one’s subspecialty, those skills not actively in use will fade. With this inherent reality in mind, what should be the goal of training? Is it to be capable of fully managing every patient (which is not currently accomplished), or is it to develop expertise in one’s subspecialty while appropriately referring with awareness of urgent issues, initial work-up, and treatment? This is the level of competency expected of a senior resident, which is currently supposed to be accomplished after 2 years of training. Brain medicine, therefore, is a deliberate shift from trying to become expert in all subspecialties to gaining a senior-resident level of competence and familiarity across aspects of the brain, followed by focused training related to one’s subspecialty. Currently, decisions of subspecialty are typically made in the third year, which would fit within the time frame of the proposed brain medicine program (Figure 1).

We have focused on subspecialties to this point, but neurologists and psychiatrists who have a more general practice constitute many of those in the two fields and are critical to the public health mission. Brain medicine easily incorporates training programs that remain broad (Figure 1D) and can be favorably compared with current structures (Figure 1A) but with an additional foundation that covers aspects of the whole brain. In this sense, brain medicine training is broad: it is more akin to internal medicine training than the narrow subspecialty of behavioral neurology & neuropsychiatry, for which combined neurology and psychiatry training is often confused.

It is worth emphasizing that in only 3 years, our internal medicine colleagues develop the clinical competence to diagnose and treat disease across all other organ systems and are actually board eligible without any required fellowship training. Our internal medicine colleagues are well respected and competent. Similar to the situation with internal medicine, all current neurology and psychiatry subspecialties would be available through brain medicine. Both fields are already governed by the same board, the ABPN, and there is already substantial overlap in tested material. We can also look to other combined programs: internal medicine-pediatrics programs combine two 3-year specialties into a 4-year program. Adding only one extra year is achieved through substantial overlap. By contrast, combining fields that merely intersect, such as child neurology, leading to board eligibility in pediatrics and neurology, as well as triple-board programs, leading to board eligibility in pediatrics, psychiatry, and child psychiatry, are both 5-year programs. Both add only 1 year to the standard 4 years of neurology or psychiatry despite only intersecting knowledge and practice bases. This raises the question of whether psychiatry and neurology have enough overlap (a single-organ system) to be accomplished in 4 years through the proposed changes described herein or whether, like the combined programs above, an additional year will be necessary.

With our collective experience from combined neurology and psychiatry training, we believe that 4 years may be sufficient to accomplish our stated goals of training, with additional expertise still available through fellowships. However, this is merely our opinion. We are proposing an idea that must be tested before it can be considered for widespread implementation. Trials and testing may reveal that 5 years is needed to achieve competence in brain medicine. The important consideration is that both 4-year and 5-year programs are tested, with detailed outcomes of competency and feasibility being assessed. Only then can the optimal length of a curriculum never before attempted be known with any certainty.

Our field is changing and requires a fresh view. For example, fellowship training may expand to include modes of practice, similar to existing nonapproved neurohospitalist fellowships or interventional psychiatry fellowships (23–25). Brain medicine provides the flexibility to focus on the most relevant training (Figure 1) while coming to understand manifestations of the whole brain.

Physician-Scientists

Physician-scientists are key to the advancement of our clinical practice. Many existing neurology and psychiatry programs allow up to 1 year of dedicated research time. Given the breadth and depth of clinical training in the brain medicine model, protected research time would not be feasible. How will physician-scientists receive the necessary research training in brain medicine? In most cases today, aspiring trainees will go on to do additional postdoctoral training or fellowships, even in situations with built-in protected time during residency. Additionally, some medical and surgical programs address a similar challenge with optional extensions of residency while building protected time into the residency. Specialty tracks also have potential to reduce clinical time in fellowships, leading to earlier independence. Ultimately, we who are physician-scientists contend that the benefit of clinical training across the brain sciences has outweighed the potential cost of postdoctoral research training.

Implementation

Change is challenging. How can a change of this magnitude actually be implemented? The first–and arguably most important–point is that we are proposing a change in residency training but not in departmental structure or clinical practice models. In other words, while this training model allows for change in response to neuroscientific advances to improve patient care, the only direct change is in residency rotations. Leaving most of the current infrastructure intact is an advantage of the brain medicine training model and may help facilitate this experiment. None of the following would have to be changed: ACGME requirements, trainee competencies, ABPN board-certification examinations, existing clinics and services, faculty positions, promotion structures, and so forth. The difference comes not in what physicians practice but in how we practice and, ultimately, in how well we practice. As an example, if a patient is seeking a movement specialist, he or she will be referred to a movement specialist in a neurology department. The difference will be that this movement specialist would have a greater background in psychiatry from the first 2 years of training, with which to manage comorbid conditions and more keenly recommend medication changes if iatrogenic contributions are recognized. Additionally, this movement specialist would have had increased exposure to his or her chosen subspecialty before a fellowship (Figure 1). The individual would otherwise be hired and promoted exactly as he or she would in the current structure.

All of this being the case, changing a residency structure is not trivial. Multiple levels of stakeholders have vested interests in these programs and must be included in the process of actually implementing brain medicine: the ACGME and its Review and Recognition Committee over neurology and psychiatry; the ABPN as the governing board for both specialties; and governing entities of both specialties, which include the American Psychiatric Association, the Association of University Professors of Neurology, the American Academy of Neurology, and the American Neurological Association. Change is difficult, and change should not be made only for the sake of change. The benefits of a brain medicine training program must prove beneficial in smaller pilot programs before widespread support is solicited.

The ACGME has a platform for novel pilot residency programs (26). Thus, a stepwise integration “where the rubber meets the road” is possible and appropriate. We expect that the model would be tested before full-scale implementation. It would likely start with a resident from one program, then another, then perhaps a small cohort, then numerous cohorts, and so forth. As with all things, the idea must be tested, and if it proves to be superior and feasible, it will naturally emerge as the predominant training paradigm. As it is in science, a novel idea may begin with a single subject before larger studies are warranted and before eventual widespread acceptance, if merited. One such scientific example of this principle comes from one of the authors of the present article (M.S.G.) (27).

Smaller and Community Programs

What if there are insufficient resources at a given program for combined training? Many smaller community programs may be able to build collaborative training models in the same way that categorical programs are started today. In cases where this is still not feasible, brain medicine programs need not exist at the exclusion of individual neurology or psychiatry programs alone. Similarly, while large academic programs typically have expertise in most or all subspecialties, smaller academic programs may not have sufficient exposure for all subspecialty tracks. Community-based residencies typically aim to train generalists, which still fits in the framework of brain medicine (Figure 1D). In short, the brain medicine model does not take away from our current training system, but rather it adds to it.

Conclusions

Physicians don’t know what they don’t know. Our experiences as current and graduated trainees of combined programs in neurology and psychiatry have taught us that there is an immediate and growing need for physicians to have an understanding of the whole brain in order to become more aware of what is not known. The exponential expansion of neuroscientific information also necessitates a greater depth of understanding to remain current. This is not a proposal to change current practice, but it is one to change training and thereby improve current practice. We have designed and described a 4-year brain medicine residency program in response to this need. It is time to break down centuries-old barriers and to let science guide the next generation of physicians dedicated to the brain. It is time for brain medicine.