JAMA AI will publish research spanning the full spectrum of AI in health care, from fundamental advances in AI capabilities to clinical applications, rigorous evaluation, and real-world implementation. The journal’s stated areas of focus include generative and agentic AI, clinical decision support, AI-enabled diagnostics and therapeutics, and digital health technologies, along with the methodological frameworks needed to evaluate their safety, effectiveness, and impact on patients, clinicians, and health systems. That breadth reflects a recognition that the scientific questions surrounding medical AI are no longer confined to computer science laboratories; they now sit at the heart of clinical medicine, health policy, and the daily practice of physicians worldwide.

The journal will be led by inaugural Editor in Chief Roy H. Perlis, MD, MSc, a prominent figure in computational psychiatry and clinical research methodology. Perlis currently serves as Director of the Center for Quantitative Health and holds the Ronald I. Dozoretz, MD, Endowed Chair at Massachusetts General Hospital, and is Professor of Psychiatry at Harvard Medical School. On February 1, 2027, he will begin a new appointment as Director of the Bakar Computational Health Sciences Institute and as the Priscilla Chan and Mark Zuckerberg Distinguished Professor at the University of California, San Francisco. His dual grounding in clinical psychiatry and quantitative methods signals the journal’s intent to bridge technical innovation with bedside relevance.

“As a JAMA Network journal, JAMA AI will be firmly rooted in medicine,” Perlis said in the announcement. “We care about advances in health AI models, trials of new interventions, and real-world clinical implementation, as well as the policy and education that can shape the responsible use of AI in medicine worldwide. I am looking forward to publishing the science that will help clinicians and health systems understand where these technologies add value and best serve patients.” The statement underscores a guiding philosophy for the journal: that the measure of medical AI is not algorithmic performance in isolation but demonstrable value in clinical practice.

The urgency of that mission is difficult to overstate. As AI capabilities advance and adoption accelerates across medicine, the need for rigorous evidence to guide safe and effective use has become increasingly pressing. Evaluating these technologies presents scientific challenges that traditional clinical research methods were never designed to address. Performance and reliability must be assessed across diverse populations and clinical settings, and researchers must understand how these tools affect clinical decision making, care delivery, and patient outcomes. Models that excel in one hospital or demographic group may fail in another, and the speed of AI development often outpaces the slower cadence of clinical validation studies.

The launch of JAMA AI reflects the rapid growth of the field and the importance of distinguishing technological promise from demonstrated improvements in patient care. As a member of the JAMA Network, the new journal will join a portfolio of premier publications that inform clinical practice worldwide. It is expected to uphold the standards that define the Network’s journals, including scientific rigor, transparency, patient-focused investigation, and ethical conduct and reporting of research, while benefiting from the Network’s broad global reach and influence. For researchers developing AI tools, publication in the journal will signal that their work has met the same evidentiary bar as any other clinical science.

“The steps we take today to build new technologies and harness the full potential of AI will shape the health care landscape of tomorrow,” said John Whyte, MD, MPH, CEO of the American Medical Association, which publishes the JAMA Network. “But innovation alone does not build trust—evidence does. For generations, the JAMA Network has set the standard for scientific research, discovery, and thought leadership on which physicians rely. This new journal builds on that legacy, helping ensure AI delivers on its promise to make medicine more effective, more personalized, and more responsive to the needs of every patient.” His comments frame the journal as a trust-building instrument at a moment when public and professional confidence in AI tools remains fragile and uneven.

Perlis brings substantial editorial experience to the role. He is also the founding Editor in Chief of the JAMA+ AI channel, which highlights AI research and commentary from across the JAMA Network journals. Since its inception, the channel has expanded to include resources and training materials, a podcast, a bimonthly email newsletter, and a search portal that allows readers to explore AI-related JAMA Network content by journal and topic. The channel will continue to curate and amplify this content for readers, creating a complementary ecosystem in which original research published in JAMA AI can be contextualized, discussed, and disseminated to a broad clinical audience.

“The opportunities created by AI are extraordinary, but realizing their potential requires rigorous evaluation,” said Kirsten Bibbins-Domingo, PhD, MD, MAS, Editor in Chief of JAMA and the JAMA Network. “The JAMA Network is an exceptional platform for authors and readers around the world to publish, discover, and debate the science that is shaping clinical practice. JAMA AI is a welcome addition to the JAMA Network, providing a trusted forum for research, education, and opinion that will help clinicians, health systems, and the public better understand how these technologies can improve care and, most importantly, improve the health of the patient.”

The arrival of a dedicated JAMA Network journal for AI arrives at a pivotal moment for the field. Generative models are being tested in documentation, diagnosis, and treatment planning; agentic systems are being explored for autonomous clinical workflows; and health systems are deploying AI-enabled tools faster than the evidence base can keep pace. By creating a peer-reviewed home for trials, validation studies, implementation research, and policy analysis, JAMA AI aims to supply the evidentiary foundation that clinicians, regulators, and patients will need to judge which technologies truly deliver. Whether the journal can keep up with the velocity of AI innovation while maintaining the JAMA Network’s exacting standards will itself be one of the defining scientific questions of the coming decade. Additional information, including an accompanying JAMA Editorial and podcast, is available at jama-ai.com.

Subject of Research: Launch of JAMA AI, a new peer-reviewed journal for clinical artificial intelligence and digital health research

Article Title: Announcing JAMA AI – a new journal from the JAMA Network

Article References: Announcing JAMA AI – a new journal from the JAMA Network. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: JAMA AI, JAMA Network, artificial intelligence, digital health, clinical decision support, generative AI, agentic AI, peer review, medical publishing, health technology evaluation, Roy Perlis, clinical research

Abstract collections occupy a distinctive place in biomedical publishing. Unlike a conventional review or original research article, an abstract supplement is a snapshot of work in progress: hundreds of concise summaries, each typically describing a single study, case series, quality-improvement project, or technological innovation. The format trades depth for breadth and immediacy. A reader can scan the entire landscape of a field in a single sitting, identifying which questions are attracting attention, which methods are gaining traction, and where the field’s energy is heading. For meeting attendees, the collection functions as a companion volume to the oral and poster sessions; for those who cannot attend, it offers a structured window into conversations they would otherwise miss entirely.

Neurocritical care itself is one of the youngest subspecialties in medicine, and its rapid maturation is part of what makes annual abstract collections from the society so closely watched. The field emerged from the recognition that patients with severe brain injury, whether caused by trauma, stroke, hemorrhage, seizures, or infection, require intensive management that addresses not only the primary insult but also the secondary injuries that unfold in the hours and days afterward. Elevated intracranial pressure, disrupted cerebral blood flow, metabolic derangements, and inflammation can each compound the original damage. Preventing those secondary insults is the central task of the neurocritical care unit, and it demands expertise that spans neurology, neurosurgery, anesthesiology, emergency medicine, nursing, and pharmacy.

The scope of work typically presented at the society’s meetings reflects that interdisciplinary character. In past cycles, abstracts have ranged from clinical trials and large observational cohort analyses to bedside physiology studies, investigations of novel monitoring technologies, surveys of practice variation across institutions and countries, and analyses of outcomes and health equity. Quality-improvement and implementation science projects are well represented as well, since translating evidence into consistent bedside practice remains one of the field’s persistent challenges. The 2026 abstracts continue this tradition, gathering contributions from investigators at every career stage, including a substantial share from trainees whose first scientific presentations often begin as meeting abstracts.

For researchers, the publication of meeting abstracts in a peer-reviewed journal serves several practical functions. It establishes priority by timestamping an idea or finding in the permanent record. It allows the work to be cited while the full manuscript is being prepared, which matters enormously in fast-moving areas where a two-year lag between meeting presentation and journal publication can render a finding stale. It also provides early-career investigators with a citable entry on their records at a stage when authorship on a full paper may still be months or years away. Journals in the field, including Neurocritical Care, have long treated these supplements as an integral part of their annual publishing cycle rather than an afterthought.

Readers and clinicians, meanwhile, use abstract collections in a different way: as a forecasting tool. Patterns across hundreds of summaries can reveal where consensus is forming and where controversy persists. When multiple groups from different countries independently report on the same question in the same year, that convergence often signals that a shift in practice or in trial design is imminent. Conversely, persistent disagreement across abstracts can highlight areas where the evidence base remains thin and where large collaborative studies are needed. Meeting abstracts therefore function as an informal but valuable barometer of the field’s direction, one that complements the slower and more definitive signal of fully published trials.

The twenty-fourth annual meeting arrives at a moment of genuine ferment in brain injury medicine. Questions about how best to monitor the injured brain, how to individualize targets for blood pressure, oxygenation, and intracranial pressure, and how to integrate artificial intelligence and advanced analytics into bedside decision-making are all active areas of investigation across the neurocritical care community. Multicenter collaboration has become increasingly central to the field’s progress, because many of the conditions it treats, while devastating for individual patients, are individually rare enough that no single institution can recruit adequate numbers for definitive studies. Annual meetings and their published abstracts are where those collaborations are forged, refined, and announced.

It is worth emphasizing what an abstract collection is not. The summaries in a supplement have typically undergone review for relevance and scientific soundness, but they are not full peer-reviewed papers, and their findings should be regarded as preliminary until they survive the more demanding scrutiny of a complete manuscript. Sample sizes in abstracts are often small, analyses are sometimes interim, and conclusions are appropriately hedged. Responsible reporting and reading of these collections means treating each summary as a hypothesis-generating contribution to an ongoing conversation rather than as settled evidence. That caution applies with particular force to any abstract touching on treatment effects, where premature changes in practice based on preliminary data can do real harm.

Even with those caveats, the appearance of the 2026 abstracts in Neurocritical Care marks the formal opening of the scientific record for the society’s twenty-fourth annual meeting. Clinicians and scientists can now locate, cite, and build upon the work presented there, and the wider public can glimpse the questions that the specialists who care for patients with the most severe neurological injuries are asking next. The full articles that many of these abstracts will become are already in preparation in laboratories and clinical research units around the world, and over the coming months and years they will carry the field’s findings from the meeting hall to the bedside. For a subspecialty defined by the race against secondary brain injury, that pipeline from preliminary result to validated practice is not an administrative detail. It is the mechanism by which the discipline improves, one meeting, one abstract, and one published paper at a time.

Subject of Research: Publication of the Neurocritical Care Society's 24th Annual Meeting 2026 abstracts in the journal Neurocritical Care

Article Title: Neurocritical Care Society Abstracts from the 24th Annual Meeting 2026

Article References: Neurocritical Care Society Abstracts from the 24th Annual Meeting 2026. (2026). Neurocritical Care. https://doi.org/10.1007/s12028-026-02638-5 (Author names are not included in the publisher’s citation export.)

Image Credits: AI Generated

DOI: 10.1007/s12028-026-02638-5

Keywords: neurocritical care, Neurocritical Care Society, annual meeting, conference abstracts, brain injury, Springer Nature, medical publishing, intensive care, clinical research, scientific meetings, Neurocritical, Care

The stakes of this question are anatomical as much as they are social. Dermatology is perhaps the most visually dependent specialty in medicine. A diagnosis of erythema, for example, relies on the recognition of redness, yet redness manifests very differently on pale skin than on deeply pigmented skin, where inflammation may appear violaceous, hyperpigmented, or nearly invisible to the untrained eye. Conditions as common as acne, eczema, psoriasis, and cutaneous lupus can present with subtle colour cues that vary systematically with the amount and distribution of melanin in the epidermis. Melanoma, too, may look markedly different on dark skin, and acral lentiginous melanoma, which arises on the palms, soles, and nail beds, disproportionately affects people with darker complexions. A clinician trained predominantly on images of light skin is therefore at a measurable diagnostic disadvantage when treating the majority of the world’s population, most of whom have intermediate to deeply pigmented skin.

The authors of the new study frame their inquiry within a growing body of literature that has documented persistent representation gaps in mainstream dermatology education. A widely cited 2021 analysis by Adelekun, Onyekaba, and Lipoff published in the Journal of the American Academy of Dermatology evaluated skin colour in dermatology textbooks and found that images of dark skin were substantially underrepresented relative to the composition of the population. That analysis echoed and extended earlier concerns raised by researchers and trainees who noticed that the visual canon of the specialty, accumulated over decades of publishing, skewed overwhelmingly toward lighter complexions. Subsequent work has examined the problem at multiple levels, from medical school curricula to residency training programs to the published literature itself, and has repeatedly found the same pattern: darker skin tones appear less often, appear later in training materials, and appear with less clinical variety.

One response to those documented gaps has been the creation and proliferation of dedicated textbooks on ethnic dermatology and skin of colour. These volumes were conceived precisely to correct the imbalance in general references, assembling clinical photography, pathophysiology, and treatment guidance relevant to patients with richly pigmented skin. There are now several well-established texts in this niche, some of them now in multiple editions, and they have become standard resources for dermatology residents and practicing clinicians who want to improve their diagnostic acumen across the full range of human skin. Given that these books exist specifically to serve populations that general textbooks have underserved, the new research letter poses an important internal question: do these dedicated resources themselves deliver balanced representation across the spectrum of darker skin tones, or do they too have blind spots?

The study, published as a research letter in volume 318 of Archives of Dermatological Research, was designed and conducted by a team spanning two Canadian institutions. Li and Liu, based at the University of British Columbia’s Faculty of Medicine and its Department of Dermatology and Skin Science, carried out the material preparation, data collection, and analysis, while Shi, of the Division of Dermatology at The Ottawa Hospital and the University of Ottawa, contributed to the conception and design of the study. According to the author contributions published with the article, all three authors participated in developing the study concept, commented on earlier versions of the manuscript, and approved the final text. The article was received in October 2025, revised in September 2026, and published on 23 September 2026, following peer review at the long-running dermatological research journal.

Methodologically, research letters in this field typically follow a content-analysis approach: the investigators identify the relevant textbooks, catalog the clinical images they contain, and classify those images by skin tone, often using an established classification system such as the Fitzpatrick scale or the more gradated tools developed specifically to reduce the tendency of observers to lump all darker skin into a single category. Finer-grained classification matters, because treating skin of colour as a monolith can obscure the distinct needs of patients across a wide range of pigmentation levels, from light brown to deep ebony tones. The detailed classification scheme used in the study, along with the underlying data, is provided within the manuscript and its supplementary information files, allowing readers to examine the criteria the authors applied and to replicate or extend the analysis.

Whatever the precise distribution the authors report, the conceptual significance of auditing dedicated textbooks is easy to articulate. Representation gaps in medical imagery are not merely cosmetic. Studies of clinical decision-making suggest that visual pattern recognition is a core component of diagnostic expertise, and that expertise is built through repeated exposure to examples. If a trainee sees twenty images of psoriasis on light skin and two on dark skin, the mental template that forms will be calibrated accordingly. The consequences are tangible: delays in diagnosis, misdiagnosis, and disproportionate disease burden among patients with darker skin have all been documented for conditions ranging from inflammatory disorders to skin cancer. Canadian survey data published in 2022 by Mahmood and colleagues on skin of colour representation in postgraduate dermatology education underscored that trainees themselves perceive these gaps in their formal programs.

The new study also connects to a broader reform movement that has gained momentum in the past several years. Syder and colleagues published a two-part analysis in the Journal of the American Academy of Dermatology documenting gaps in skin of colour content across medical school, residency, and continuing education, describing curricula that treat pigmented skin as an occasional special case rather than an integral part of dermatologic practice. Wilson and colleagues, writing in the International Journal of Women’s Dermatology, analyzed diversity content in the dermatologic literature and issued what they called a call to action, urging journals, educators, and publishers to treat equitable visual representation as a basic quality standard rather than an aspirational goal. Chang and Lipner catalogued resources for skin of colour images to help educators find better material. Against that backdrop, auditing the dedicated textbooks themselves is a logical next step, because these books are the very resources educators turn to when they try to close the gap.

The researchers report that they have no relevant financial or non-financial interests to disclose, and the study was conducted without external funding, which the authors state in their declarations. That independence matters in a conversation that touches on publishing industries with commercial stakes in their back catalogs. Publishers of medical textbooks face real costs in commissioning new clinical photography, obtaining patient consent, and reprinting new editions, and critics have noted that these economic realities can slow the pace at which visual content is refreshed. Independent audits such as this one provide the empirical evidence base that educators, professional societies, and publishers can use to prioritize where new imagery is most needed and to track progress over time.

For clinicians, educators, and patients, the message of the study is ultimately constructive. The existence of dedicated ethnic and skin of colour dermatology textbooks represents real progress in a specialty that for most of its modern history centered its visual teaching on a narrow slice of humanity. Auditing those texts is not an indictment of them but a way of ensuring that the tools built to correct one imbalance do not quietly introduce another. As dermatology’s global patient population continues to demand equitable, expert care, the granularity with which textbooks represent the full spectrum of skin tones will increasingly be seen as a marker of educational quality. The new research letter in Archives of Dermatological Research, available through its publisher with the data supplied in the manuscript and supplementary files, gives the field a rigorous, transparent framework for holding its own most specialized resources to that standard.

Subject of Research: Skin colour representation in dedicated ethnic and skin of colour dermatology textbooks

Article Title: Skin colour representation in dedicated ethnic and skin of colour dermatology textbooks

Article References: Li, M., Liu, W., & Shi, J. (2026). Skin colour representation in dedicated ethnic and skin of colour dermatology textbooks. Archives of Dermatological Research, 318(1), Article 475. https://doi.org/10.1007/s00403-026-04979-9

Image Credits: AI Generated

DOI: 10.1007/s00403-026-04979-9

Keywords: dermatology, skin of colour, medical education, textbooks, health equity, clinical imaging, diagnosis, representation, skin pigmentation, research letter, medical publishing, curriculum gaps