<strong>Background:</strong> Cervical sensorimotor control (SMC) is often disrupted in individuals with chronic neck pain, contributing to persistent symptoms and functional limitations. Traditional cervical SMC assessments are limited by complex setups, single-domain testing, and examiner dependency. Virtual reality (VR) technology offers a promising platform for multidimensional, standardized, and user-friendly assessment. <strong>Objective:</strong> This study aimed to develop and evaluate the validity and reliability of a VR-based system for assessing cervical SMC in healthy adults. <strong>Methods:</strong> A cross-sectional observational study was conducted in 30 healthy adults (aged 18-60 years). The custom-developed VR system (HP Reverb G2 Omnicept Edition, Unity engine) incorporated 5 SMC tests: cervical range of motion (ROM), joint position error (JPE), head-tilt response, figure of eight (FOE), and postural sway (PS). Test-retest reliability was assessed across 2 sessions, separated by 1 week, using intraclass correlation coefficients (ICCs). Concurrent validity was examined by comparing VR-based measures with gold standard optical motion capture or established clinical tools using Pearson correlation coefficients. <strong>Results:</strong> The VR system demonstrated good to excellent test-retest reliability across most outcome measures. The ICC for cervical ROM ranged from 0.851 to 0.968 across movement directions. The ICC for JPE in each direction ranged from 0.813 to 0.827. The ICC for the FOE test’s deviation frequency and task duration were 0.810 and 0.913, respectively. The ICC for the head-tilt response was 0.742 and ranged from 0.720 to 0.843 for PS under both visual conditions. The VR-based assessments for ROM, JPE, FOE, and PS showed strong correlations with reference measures (<i>r</i>=0.723-0.980), supporting concurrent validity. <strong>Conclusions:</strong> This VR-based assessment system provides a valid, reliable, and user-friendly multidimensional evaluation of cervical SMC. It offers a standardized, integrated, and clinically feasible alternative to conventional assessments, with potential applications in both clinical diagnostics and rehabilitation monitoring. <strong>Trial Registration:</strong> ClinicalTrials.gov NCT06474130; https://clinicaltrials.gov/study/NCT06474130
In this cross-sectional study of 300 board-style questions across physical therapy, occupational therapy, and speech-language pathology, we evaluated reasoning types and found high overall accuracy with variation by discipline and reasoning category; the strongest performance was in deductive and analytical reasoning and the lowest accuracy was in evaluative reasoning.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/63417171c0f2224d0fee995648371543" />
Background: Internet-based interventions for insomnia show promise, but understanding the process variables, such as knowledge acquisition and psychological acceptance, is crucial for enhancing digital adherence and clinical effectiveness. Objective: This study aimed to evaluate the feasibility, adherence, and preliminary clinical signals of 2 therapist-assisted interventions—internet-delivered cognitive behavioral therapy (iCBT) and internet-delivered acceptance and commitment therapy (iACT)—for insomnia in a primary care setting. Methods: This was a pilot randomized controlled trial. Adults seeking help for insomnia (n=18) were recruited via primary care and randomized to either a 5-module iCBT or iACT program delivered via a secure digital platform with weekly therapist feedback. Blinding of participants and therapists was not possible due to the nature of the interventions. Primary outcomes included the Insomnia Severity Index; secondary outcomes included the 9-item Patient Health Questionnaire, 7-item Generalized Anxiety Disorder, and WHO Disability Assessment Schedule. A novel sleep knowledge test was used as a process variable. The data were analyzed using split-plot analyses of variance (intention-to-treat or last observation carried forward and complete case analysis) and nonparametric Friedman and Kruskal-Wallis tests. Results: A total of 18 participants were randomized (iCBT: n=9; iACT: n=9). High attrition was observed, with only 33.3% (n=3) of iCBT and 55.6% (n=5) of iACT participants completing all modules. The iACT group demonstrated a significant within-group reduction in insomnia severity (=.01, Friedman test), whereas iCBT results were nonsignificant (=.10, Friedman test). No significant between-group differences were found for any clinical or process variables. Participants rated both treatments as credible (Credibility/Expectancy Questionnaire scores remained stable), though qualitative feedback indicated a need for more flexible, less burdensome content. Conclusions: This pilot study demonstrates that while internet-delivered insomnia treatments are feasible and credible in primary care, high attrition remains a significant barrier. Preliminary signals suggest that iACT may be a viable alternative to iCBT, potentially offering better adherence. Larger, fully powered pilot randomized controlled trials (estimated N=404) with refined recruitment and automated retention strategies are required to determine definitive comparative efficacy and the mediating role of sleep knowledge and acceptance. Trial Registration: Research and Development in the Västra Götaland Region (FoU i VGR) 272866;
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/afbaf4554a19449ff56c7c7894d41783" />
AI companies want to build systems that understand the external world and overcome the limitations of LLMs. Recent developments have brought world models to the forefront of the AI discussion.
Watch a conversation with editor in chief Mat Honan, senior AI editor Will Douglas Heaven, and AI reporter Grace Huckins exploring how AI might enter the physical world.
BOSTON — Thomas Bartlett’s life changed in 2019 when he was diagnosed with late-onset myasthenia gravis (MG). The 15-year veteran of the Bio-IT World Conference and Expo took to the stage on Tuesday as one of seven speakers in the opening plenary session of this year’s conference, which focused on various aspects of rare disease research and treatment.
The session offered a poignant, but often uplifting, launch to the annual conference, which celebrated 25 years from its inception in 2002, when the event was produced by the IDG World Expo Group.
Speaking with Susan Ward, PhD, founder and executive director of the Collaborative Trajectory Analysis Project (cTAP), Bartlett described an active life and fulfilling work prior to his diagnosis, and some of the debilitating physical and emotional impact of his disease. “I have to plan everything. If I’m going to go out, I plan ahead of time where I’m going to go [and] the amount of time,” he said. “I have to plan recovery.” Bartlett’s MG has prevented him from working a full-time job, as he would need a full day of rest just to recover from each day. “The math doesn’t work.”
Bartlett is now an ambassador for MG Uniter Myasthenia Gravis, an online platform designed to support some 70,000 patients living with the disease in the United States alone. Though there are some treatments that alleviate disease symptoms, there currently is no cure. Bartlett’s disease was diagnosed early thanks to a quick-thinking primary care provider. A recurring theme in the session was the stark reality that many in the rare disease community wait many years for a diagnosis.
Of the estimated 7–8,000 rare genetic diseases, many not well understood. About one in 10 people in the U.S. “either has or will have a rare disease at some point,” Ward noted. In a conference of about 2,700 attendees, “there are going to be about 270 people on average who might have a rare disease. So the magnitude of the problem is huge, even though the numbers of people are quite small.”
Given the computational nature of the Bio-IT conference, Bartlett and Ward soon turned to data and the challenges with collecting and aggregating information from rare disease populations. Ward noted that centers of excellence in rare disease may have several patients but “no one center that has enough data for anybody to really learn much.” Aggregating data from multiple centers and across geographies is one possibility but due to the differences that exist between centers across states and countries, “you need a really rich and deep ontology” as well as “context for what those data mean,” she said.
And that’s not the only challenge. In many cases, rare diseases can present and progress differently in patients with the same condition. “Imagine you’re trying to design a clinical trial. You’ve got patients who are fluctuating [while] you’re really looking for patients who are slowly declining” and “patients who have intermittent remissions,” Ward noted. With that mix, “you’re going to have a very noisy trial.”
There are also other data sources that could provide value. Bartlett noted that wearable devices such as the Apple watch capture useful health-related data, but as it is not clinical data, physicians cannot use it. As someone with decades of tech experience—including a stint at Apple—Bartlett asked: “How do we change that?” How do we prove the patient’s experience with the data that we can collect, and work with companies and legislation” to “include real world data and evidence and compare that … with the clinical data and get a much broader picture.”
Bartlett’s closing comments focused on hope for people living with rare diseases. And with good reason given recent successes in development of gene therapies and other therapeutics. As a patient, “you need to have something that you can look forward to today,” he said. For now, MG is incurable but “we will find ways to get to that end game and ultimately have a cure or at least a high level of quality of life.”
Shortening the rare disease diagnostic journey
Sebastien Lefebvre, head of technology, data and AI at Aurelis Insights, has spent 10 years in the rare disease space in different capacities including developing platforms for digital decision support or “data-driven and AI-assisted support decisions.” Speaking with William Van Etten, PhD, co-founder, CEO and principal scientist, StarfleetBio, Lefebvre described Rare Answers, a clinical decision support platform for rare disease diagnostics that he worked on while at Alexion Pharmaceuticals. The platform was designed in collaboration with two children’s hospitals as well as a number of technology and data science companies.
Sebastien Lefebvre, head of technology, data and AI, Aurelis Insights (right) and William Van Etten, PhD, co-founder, CEO & Principal Scientist, StarfleetBio (left). [Uduak Thomas]
He also described a second project in 2022 with the Rare-X program that analyzed data from public databases of rare and inherited diseases, drugs, and genes. Lefebvre and his colleagues hoped to produce an accurate assessment of the total number of rare diseases. Following extensive data cleaning and normalization—now made much simpler with advances in AI and machine learning—they arrived at a figure closer to 12,000. Of that number, between 80–87 percent have a genetic basis and about 80 percent had at least three associated phenotypic descriptors. That kind of information provides a viable starting point for applying AI-assisted data-driven diagnostic approaches. This is important because, as Bartlett noted, many patients with rare diseases wait years for a diagnosis. “It all starts with a diagnosis,” Lefebvre said. “[If] you don’t known what you’ve got, how can you [treat it].”
Van Etten is focused on making individual genomes truly private. The emergence of commercial personal genomics companies created a data privacy problem. Customers pay for their genomes to be sequenced by a company that holds the data, reads it, and sends periodic reports. He has developed an app called DNAVault that lets people host their genomes on their smartphones, putting data control back into their hands.
Van Etten’s new company, StarfleetBio, is partnering with his former consulting firm, BioTeam, and the Hubbard Center for Genomic Studies at the University of New Hampshire, to provide sequencing services.
“It used to be that we needed to centralize all the human genome data because you need a lot of compute to perform the analysis, but it’s really not required anymore,” he said. “We decided to decentralize it, where your genome is on your phone, you can generate your own reports, and nobody has access to it but you.”
Each encrypted genome is only accessible with a key unique to the individual’s phone. This way, only they individual can download their data and read it. Some audience members clearly approved of Van Etten’s app, with shouts of “Bravo!” from the back of the hall. (The app was later named one of three “Best of Show” winners at this year’s meeting.)
Among the features in the app is a fun kinship feature, which lets two people determine if they are related by placing their phones in close proximity as if sharing a Wi-Fi password. Another feature dubbed “origins” lets people track their ancestry over thousands of years via their Y-chromosome or mitochondrial DNA. Van Etten was particularly moved by the kind of insights this feature revealed about human relationships. “We found that all humans are far more closely related than we thought,” he said. “We all really [came from] the same 5,000–10,000 people from 50,000–70,000 years ago.”
Another app feature screens for the 81 ACMG medically actionable genes to provide health reports, while a final feature lets people ask questions about their genome and get answers much the same way one might enter a question into Google or ChatGPT.
Tying this to rare diseases, Van Etten is working on ways for app users to opt into participating in relevant research studies and clinical trials. The idea is that users could “toggle a switch” that would let alert the relevant researchers and then answer questions to help gauge eligibility. Importantly, this would all be done without people having to share their primary information.
Learning from rare diseases to treat common conditions
Another plenary conversation took place between Morgan Cheatham, MD, partner, head of healthcare & life sciences, Breyer Capital, and Catherine Brownstein, PhD, manager of the Molecular Genomics Core Facility at Boston Children’s Hospital and scientific director of the Manton Center for Orphan Disease Research Gene Discovery Core.
Morgan Cheatham, MD, Partner, Head of Healthcare & Life Sciences, Breyer Capital (left), and Catherine Brownstein, PhD, Manager of the Molecular Genomics Core Facility at Boston Children’s Hospital and Scientific Director of the Manton Center for Orphan Disease Research Gene Discovery Core (right). [Uduak Thomas]
Brownstein and Cheatham use OpenAI’s generative AI to help diagnose patients who in some cases had been waiting decades for answers. Importantly, “this was a zero-shot model,” Cheatham noted. “We didn’t do any specialized training of GPT-3, we just deployed the existing models and we’re able to return answers to families who have been waiting for sometimes over a decade.”
He also acknowledged the contributions of people living with rare diseases to many major drug modalities including CAR Ts and RNA medicines. “Many of those modalities were actually validated” with “the help of rare patients who were willing to participate in trials that allowed us to show the efficacy, the safety, and the durability of these modalities.”
According to Brownstein, a deeper understanding of rare conditions often has implications for more common conditions. “As someone who spent six years studying hypophosphatemic rickets … it’s these extreme cases, these rare presentations of disorders where you don’t know the underlying etiology [that] inform the common diseases,” she said. Understand the biology behind hypophosphatemic rickets “has implications for bone density and osteoporosis that affects a ton of us in this room.”
Many opportunities were highlighted where some form of AI is already being used or could be applied. One company that Cheatham mentioned is applying AI to colonoscopies to characterize inflammation levels in the bowel in a standardized way with an eye towards connecting patients with ulcerative colitis and Crohn’s disease to relevant clinical trials. There are also opportunities in cardiology, neurology, pathology and more.
Giving more patients the right to try
In the closing conversation, Van Etten spoke with Dylan Livingston, founder and president of The Alliance for Longevity Initiatives (A4LI). Livingston is at the forefront of efforts aimed at implementing policies in different states that allow patients with rare diseases to try treatments that may benefit before they have been approved.
The story of how Livingston, still in his 20s, got involved in healthcare policy is interesting. As a college senior during the Covid-19 lockdowns, “I started thinking about COVID as it relates to age [and] why … [I] would be pretty much completely unaffected by COVID and why my grandfather at 92 would most likely die,” he recalled. “It all comes back to aging, your immune response to these diseases and your immune response to chronic diseases overall.” That got him interested in the field of aging and longevity more broadly.
Dylan Livingston, founder and president of The Alliance for Longevity Initiatives (A4LI) (left) and William Van Etten, PhD (right) in conversation at Bio-IT World 2026 [Uduak Thomas]
Livingston and his group have worked to pass laws in the state of Montana that extend eligibility under the Right to Try Act, a piece of federal legislation that lets people with terminal illnesses try therapeutics that may help them which are not yet fully approved. The issue with the Right To Try Act as it stands is that “the definition of who is eligible is very narrow” and restricted to people with months left to live “which made no sense to me” Livingston said. From his perspective, people just diagnosed with conditions like Alzheimer’s or Parkinson’s should also have the chance to access treatments which could potentially help them earlier in their journeys as those who are further along in their journeys.
Expanding the Right to Try provides a possible pathway to those treatments without requiring approval from the U.S. Food and Drug Administration, which may be years away.
Livingston and his team have been successful in expanding the law in Montana to cover people with age-related ailments as well as people with rare diseases, people recently diagnosed with terminal diseases, and people with disease that will eventually become terminal. Now he and his team are working on getting similar changes in place in New Hampshire. There are safeguards in place: the proposed treatment has to be prescribed by two physicians, pass through IRB review, and the therapy must have passed a Phase I testing. “What we’re trying to do is create a system that is safe enough to prevent as many tragedies as possible while also opening up access to as many people as possible,” Livingston said.
As an example of the benefit of changing the law, Livingston shared a story of a father whose son had died from a rare mitochondrial disease. The father has since had the genomes of his two other children sequenced, only to discover that they carry the same mitochondrial mutation. In this scenario, Montana’s model would allow the father in this instance to bypass the strict requirements of a drug trial and access treatments that could potentially help his children.
“Maybe it’s not as great in terms of a data collection standpoint for companies, but what we’re offering here [are] options for people that don’t have any other options.”
*Bio-IT World Conference & Expo, Boston; May 19-21, 2026.
Scientists at the UCLA Health Jonsson Comprehensive Cancer Center have developed a cytokine-armored CAR T-cell therapy that helps the immune system better attack aggressive brain tumors in mice. Their study showed that the treatment also reduced dangerous side effects that have long limited immune-based treatments for glioblastoma, which is one of the deadliest and most treatment-resistant brain cancers.
The therapy works by reprogramming CAR T cells to release immune-stimulating proteins, IL-12 and DR-18, which activate the body’s own immune system, strengthening the overall anticancer response. This treatment approach improved tumor control in mouse models, including those carrying cancers made up of mixed cell populations that often escape treatment. Researchers also found that pairing the treatment with a second CAR T strategy targeting VEGF helped reduce side effects while preserving strong anti-tumor activity.
The findings point to a potential new strategy for treating recurrent high-grade gliomas and other solid tumors that historically have been difficult to target with CAR T-cell therapy. Research lead Yvonne Chen, PhD, co-director of the Tumor Immunology and Immunotherapy Program at the UCLA Health Jonsson Comprehensive Cancer Center, is senior author of the study, which is published in Cancer Research, in a paper titled “Armored Chimeric Antigen Receptor-T Cell Therapy Targets Antigen-Heterogeneous Glioma.”
Glioblastoma remains extremely difficult to treat because tumors suppress immune responses, contain diverse cancer cells, and create abnormal blood vessels that limit the effectiveness of immunotherapy. “Two features of glioblastoma pose formidable barriers to effective immunotherapy: tumor-antigen heterogeneity and a highly immunosuppressive tumor microenvironment (TME),” the team wrote. While CAR T-cell therapy has transformed treatment for certain blood cancers, success in solid tumors has been limited.
“Early data from clinical evaluation of chimeric-antigen receptor (CAR) T-cell therapies for glioblastoma show a strong safety profile and promising signs of response, but durable efficacy remains elusive,” they continued. Chen added, “A key challenge in treating brain tumors, particularly glioblastoma, is that the tumor cells are often antigen heterogeneous, meaning they do not all express the same proteins that can be recognized by a given targeted therapy.” The researchers further stated, “The glioblastoma TME is characterized by a variety of dysfunctional tumor-associated cell types that support tumor growth and metastasis.” The most abundant of these are tumor-associated macrophages (TAMs), which can directly suppress immune-cell function and promote tumorigenesis.
“We hypothesized that effective immunotherapy against brain tumors would have to engage naturally occurring immune cells, which can recognize a wide variety of target antigens, in the fight against cancer,” Chen noted.
Because brain tumors are considered immunologically cold, meaning they do not naturally trigger a strong immune response, the researchers designed “armored CAR T cells” to activate immunity against the tumor. These CAR T cells were built to recognize a tumor antigen called IL-13Rα2, a protein commonly found on glioblastoma cells, while also secreting immune-stimulating proteins that recruit and activate the body’s immune cells.
The team then tested multiple combinations of these “armor” molecules in immunocompetent mouse models of glioblastoma, using head-to-head comparisons to evaluate how each design affected tumor growth and immune activity. The CAR T cells were studied in several orthotopic glioma models, including tumors engineered to vary in antigen expression to better reflect the heterogeneity seen in human disease. After testing multiple combinations, researchers identified one especially potent pairing: IL-12 and decoy-resistant IL-18 (DR-18). “Through head-to-head in vivo comparisons of potentially synergistic armor combinations, we demonstrated that T cells expressing a CAR plus IL-12 and the decoy-resistant form of IL-18 (CAR-12.DR18 T cells) show strong efficacy against antigen-heterogeneous glioma in immunocompetent mice,” the investigators reported.
The team showed that the therapy demonstrated the ability to eliminate tumors containing cancer cells that lacked the target recognized by the CAR T cells, a major hurdle in glioblastoma treatment because tumors can evolve and escape single-target therapies. “IL-12 and DR-18 work synergistically to activate the immune system, resulting in a dramatic influx of immune cells into the tumor-bearing brain,” stated Chen, who is also a professor of microbiology, immunology, and molecular genetics at UCLA and a member of the UCLA Broad Stem Cell Research Center. “The diverse immune-cell population recruited into the brain contributes to attacking the tumor, including ones that cannot be directly recognized by the CAR T cells themselves.”
Because IL-12 can trigger dangerous inflammation, the researchers also explored ways to reduce side effects while maintaining anti-tumor activity. They found that adding a second engineered CAR T approach targeting VEGF—a protein that drives abnormal blood vessel growth and contributes to swelling in glioblastoma—helped reduce treatment-related toxicity while maintaining strong tumor control in mice. “Robust anti-tumor efficacy with effective toxicity mitigation was achieved via combined administration of CAR-12.DR18 T cells with CAR T cells that secrete an anti-vascular endothelial growth factor (VEGF-A) single-chain variable fragment. This combination therapy presents a clinically applicable strategy to overcome key barriers to effective treatment of glioblastoma,” the authors stated.
“When developing novel therapies, we always have to balance considerations for safety and efficacy,” Chen said. “Potent cytokines such as IL-12 and DR-18 have toxicity potential, which is why we performed in-depth studies to understand the nature and severity of the toxicity and devised ways to counteract safety concerns while maintaining anti-tumor activity.”
The findings point to a potential new strategy for treating recurrent high-grade gliomas. The researchers are now completing the necessary preclinical studies and raising funds to launch a Phase I clinical trial in patients with the disease.
“We are very encouraged by the ability of our cytokine-armored CAR T cells to kill not only tumor cells that express IL-13Rα2, but also tumor cells that are not directly recognizable to the CAR T cells,” Chen said. “We are excited to have developed a clinical protocol that would allow us to bring this therapy to the clinic while also providing a detailed toxicity management plan to ensure patient safety.”
<![CDATA[Explore how integrated psychotherapy—CBTp, family support, and humane alliance—reduces relapse and restores meaning beyond medication in schizophrenia.]]>
If you’ve spent any time talking to other parents about sleep, you’ve probably heard about melatonin. One person swears by it. Someone else warns against it. And if your child is struggling to fall asleep, it can be hard to know what to believe.
Melatonin is widely available, often marketed as a “natural” sleep aid, and increasingly used to help kids of all ages. But it may not be the right solution. The key is knowing when it makes sense, when it doesn’t, and how to make it work best as part of a bigger sleep plan.
What is melatonin?
Melatonin is a hormone the body produces naturally to regulate sleep. As it gets dark, the brain releases melatonin to signal that it’s time to wind down. Light exposure at night — especially blue light from screens — can disrupt the body’s natural rhythm so you don’t feel sleepy even when it’s bedtime.
“Melatonin supplements can help facilitate that circadian rhythm, that 24-hour sleep cycle,” says Rohn Nahmias, DO, a child and adolescent psychiatrist at the Child Mind Institute. Because it’s sold over the counter, melatonin seems to be harmless. In practice, clinicians are much more cautious. While they tend to consider it relatively low risk, with few side effects, they still don’t view it as something to take casually or indefinitely.
That’s because there are real gaps in what we know about melatonin — especially when kids take it regularly over long periods. “The longest study was about four years, and they did not find any issues,” says Judith Owens, MD, MPH, a pediatric sleep expert and professor of neurology at Harvard Medical School. “But the data are very limited in subject numbers and long-term follow-up.”
So, the question parents often ask — is melatonin safe for kids? — doesn’t have a simple yes or no answer. The more useful question may be whether melatonin is right for your kid and what guidelines should you follow if you’re going to give it to them.
That’s why it’s especially important to consult your child’s pediatrician or tell them if you’ve already started your child on melatonin. “Because melatonin is over the counter, families often forget to mention it,” says Dr. Nahmias. But doctors need to know about supplements, especially if a child is on other medications or has additional health concerns.
Melatonin can be helpful in the right situations, but using it without looking deeper can be “kind of slapping a Band-Aid onto a problem,” Dr. Nahmias says. In his practice, he always wants to rule out anxiety, mood concerns, or a medical issue that may be impacting sleep (such as snoring, breathing problems, or pain) before introducing something new like melatonin.
When melatonin can help
Melatonin is most useful when the issue is falling asleep, not staying asleep. For example, if your child has persistent trouble falling asleep — not just occasional bedtime resistance — and you have already tried using behavioral strategies.
The immediate-release melatonin sold in the United States is primarily helpful for sleep onset. (A prolonged-release form approved in the UK and EU may help with staying asleep, but it’s not available here.) Melatonin can also help with circadian rhythm issues like delayed sleep-wake phase disorder, when a child or teen’s natural sleep and wake times are much later than their schedule allows. “That is the other real indication for melatonin,” Dr. Owens says.
How to use melatonin thoughtfully
If melatonin does make sense for your child, experts agree: Start low and go slow. Dr. Nahmias recommends beginning with 1 to 2 milligrams for kids four and up and increasing the dose only if needed — up to 3 milligrams for kids ages 6–10. Many children respond well to low amounts. More than 5 milligrams, he says, isn’t much more effective and is more likely to cause side effects like grogginess or irritability. Dr. Owens emphasizes a similarly cautious approach: Use the lowest effective dose, monitor whether it’s actually helping, and reassess regularly rather than letting it quietly become an open-ended routine.
Dr. Nahmias recommends using melatonin in “clusters” — short, purposeful stretches of nightly use to help realign the circadian rhythm — and then pulling back to use as needed. A cluster might make sense for a few weeks after travel, during a tough transition, or while the family works on behavioral changes.
If a child needs melatonin every night for months, that’s a signal to dig deeper. “That tells me that there’s something likely going on underneath that’s not being addressed,” Dr. Nahmias says.
Side effects and safety
Melatonin is generally well tolerated, but knowing the possible side effects can help you catch problems early. The most common is grogginess the next morning. Others include irritability, headaches, dizziness, stomach upset, and — in toddlers and younger children — increased bedwetting. Some children may also experience vivid dreams or nightmares, though these tend to be mild.
Check in with your child’s pediatrician if side effects appear, your child needs melatonin frequently or for more than a short period, the dose keeps creeping up, sleep problems get worse, or if you notice signs of anxiety, depression, or ADHD that might be driving their sleep struggles.
One important safety note: melatonin gummies look like candy, and Dr. Owens describes an “astronomical increase” in calls to poison control centers and emergency room visits related to melatonin in children, largely due to accidental ingestion. Store melatonin — especially gummies — like any other medication: out of reach, ideally in a locked cabinet. Never present them to children as a treat.
Dr. Owens describes the variability as “huge” and says she was genuinely shocked by the findings. When shopping, look for products with a USP Verified mark. USP (United States Pharmacopeia) is an independent nonprofit that tests supplements to confirm the product contains what the label says, in the correct amount, and without harmful contaminants.
Melatonin for kids with ADHD or autism
There’s solid evidence that melatonin can benefit children with neurodevelopmental conditions, particularly autism and ADHD, who are more likely to have disrupted sleep-wake cycles. “There is a pretty robust literature supporting efficacy, without a lot of side effects,” Dr. Owens says about children on the autism spectrum.
But Dr. Nahmias encourages parents of kids with autism or ADHD to take a closer look at bedtime routines before turning to melatonin. “Both of those populations of kids do best when there is structure put into their day,” he says. He recommends having a posted list of the sleep routine that a child needs to accomplish as a helpful visual reminder that will start to become second nature as it is built into their evening.
If getting to sleep is still an issue, trying melatonin makes sense. But for kids with autism and ADHD, it may take more time to see an effect. “It can be helpful to give it a bit longer to re-right the sleep cycle,” Dr. Nahmias says. “Trying it for two to three months may be more beneficial than just a few weeks.”
Many neurodivergent kids may also need to stay on melatonin longer than neurotypical children. “This makes it absolutely imperative that administration of melatonin for these children is under the supervision of a health professional who can monitor efficacy and side effects, and recommend periodic ‘off-drug holidays,’” Dr. Owens says. The goal is to get sleep back on track and consistent for some time. Once that goal is achieved, Dr. Nahmias adds, “it is important that there be attempts to take breaks from the medication or try lower doses.”
When melatonin is not likely to help
Melatonin is often used in situations where it’s unlikely to make much difference. That’s not a criticism of parents — sleep deprivation is exhausting, and it’s natural to reach for something that seems gentle and accessible. But if the real issue is an inconsistent routine, untreated anxiety, or an unidentified medical problem, melatonin may only be a temporary solution. The real problem will persist and need diagnosis and treatment.
For younger children — kids under five, especially — sleep problems are almost always better addressed with changes in habits and routines rather than supplements. For kids under two, “there’s no reason to use melatonin… and really, honestly, under five, for the most part,” Dr. Owens says. At those ages, behavioral approaches almost always work better.
What to try before using melatonin
Small, consistent changes in sleep habits can make a real difference — and unlike a supplement, these strategies can help children build skills that support sleep for years to come. Dr. Nahmias focuses on what’s often called sleep hygiene: a consistent bedtime and wake time, a predictable wind-down routine, and a sleep environment that actually supports rest, like keeping the bedroom dark, cool, and quiet. “The bed is meant for one thing and one thing only, and that is for sleep,” he says, emphasizing no TV or cellphones before bed. Even small amounts of light can interfere with the body’s natural melatonin production.
He also suggests addressing anxiety, rumination, or bedtime fears directly. “Behavioral interventions, time and time again, have really been shown to be very effective,” Dr. Owens says. For younger children, that might mean learning to fall asleep without a parent in the room. For older kids and teens, it often means setting limits around devices and making sure the schedule they’re keeping is actually realistic.
Making a melatonin plan
Melatonin is a tool that works best with a plan behind it. Before starting, write down what problem you’re actually trying to solve. Is your child unable to fall asleep before 11pm? Waking during the night? Scared to sleep alone? Anxious about school?
A simple sleep log can make your conversation with your child’s doctor much more useful. For a week or two, track bedtime, approximate time asleep, any night wakings, morning wake time, screen use, and whether melatonin was used and, if so, at what dose. Dr. Owens calls sleep diaries “invaluable” because they reveal patterns that are nearly impossible to see when everyone’s tired and running on memory. That information helps you and your child’s doctor decide whether melatonin is worth trying, whether it’s working, and when it might be time to stop — keeping the focus not just on getting through tonight, but on helping your child build the sleep habits they’ll carry with them for years.
Frequently Asked Questions
Is melatonin safe for kids?
Melatonin is generally considered low risk and can be helpful for some children who have trouble falling asleep, especially when used short term and under a doctor’s guidance. But experts caution that there are still gaps in research about long-term use in kids, so it shouldn’t be treated as harmless or used casually without looking at underlying issues.
Can you OD on melatonin?
Accidental overuse can happen, especially because melatonin gummies may look like candy to young children. In recent years, there has been a major increase in poison control and ER visits related to accidental ingestion, so melatonin should always be stored like any medication — out of reach and ideally locked away.
Is melatonin bad for kids?
Melatonin is not inherently “bad” for kids, but it’s not the right solution for every sleep problem. If poor sleep is caused by anxiety, inconsistent routines, or a medical issue, melatonin may only mask the problem.
How much melatonin is safe for kids?
Experts recommend starting with the lowest effective dose. Some clinicians suggest starting with 1–2 milligrams for children ages four and up, and generally not exceeding 3 milligrams for kids ages 6–10 unless advised by a doctor.
Scientists at the Icahn School of Medicine at Mount Sinai have created a new artificial intelligence (AI) model that helps reveal how genes function together inside human cells, offering a powerful new way to understand biology and disease. Their study, headed by Avi Ma’ayan, PhD, professor of pharmacological sciences and director of the Mount Sinai Center for Bioinformatics at the Icahn School of Medicine at Mount Sinai, introduces a gene set foundation model (GSFM) designed to learn patterns in how genes are grouped and function across thousands of biological contexts.
The work draws inspiration from advances in large language models (LLMs) such as ChatGPT, which learn how words gain meaning depending on their context. In a similar way, a GSFM learns how genes behave differently depending on their cellular “context.”
The model provides a new way to understand the structural and functional organization of genes and their products inside human cells. This improved understanding could eventually support the development of better diagnostics, biomarkers, and therapies. By mapping how genes relate to one another across many biological situations, the GSFM creates a reference framework that can help scientists interpret complex multiomics datasets more effectively, say the investigators. “The organization of genes within cells remains one of the major unsolved questions in biology,” Ma’ayan noted. “The GSFM helps address this by learning from millions of gene groupings derived from published research and gene expression datasets.”
In their paper the authors explained, “Genes are a bit like words, and gene sets are a bit like sentences, because words are reused in different contexts to express unique meanings, and cells reuse genes to carry out different biological functions.”
“Genes rarely act alone,” Ma’ayan further noted. “Instead, they participate in multiple biological processes, forming different molecular groupings depending on where and when they are active in the cell. A single gene can play different roles in different settings, much like a word can have different meanings in different sentences. Just as modern language models learn the meaning of words from context, we asked whether AI could learn the ‘meaning’ of genes in the same way. Our GSFM was designed to do exactly that.”
To build the model, the researchers compiled millions of gene sets from published scientific studies and gene expression datasets. In total, the system learned from hundreds of thousands of independent research efforts.
The AI model was trained in a way similar to solving a puzzle: it was given part of a gene set and asked to predict the missing pieces. Over time, it learned underlying patterns that describe how genes are grouped and interact.
The AI model was then benchmarked against other approaches and demonstrated strong performance, including the ability to identify gene-gene and gene-function relationships before they were confirmed experimentally. To evaluate this, the model was trained using gene sets from publications up to a defined cutoff date, and then tested on whether it could predict discoveries reported in studies published after that cutoff date.
“Unlike previous biological AI models that primarily rely on gene expression data, our GSFM is uniquely trained on gene sets, a different and largely underused type of biological information,” Ma’ayan stated. “This approach allows the model to integrate diverse data from many diseases, experimental methods, and research conditions, creating a unified representation of gene relationships across biology.”
The team’s studies showed that the new model can help identify the function of poorly understood genes without immediate laboratory experiments, highlight genes involved in disease processes, and suggest potential new drug targets and biomarkers. The model offers a reusable knowledge system for many types of biomedical research data analysis tasks—for example, improved gene set enrichment analysis. In essence, the researchers suggested, GSFM offers a new “map” of how genes work together in different contexts. “Unlike prior methods that are mainly based on similarity of all genes to annotated genes, GSFM’s architecture can capture the more complex non-linear and multi-modal relationships between genes and the gene modules these genes constitute,” the investigators wrote. “GSFM’s ability to predict genes held out from known gene sets can be useful for many applications in computational systems biology.”
GSFMs could enhance existing bioinformatics tools and improve the interpretation of data collected with omics technologies. One immediate application is in gene set enrichment analysis, a widely used method in molecular biology research. By improving how scientists interpret gene groupings, the model may help uncover new biological insights from both existing and future datasets.
“Like the way LLMs predict the next word in a sentence, GSFM guesses the next missing gene when presented with a gene set,” the scientists stated. “With this power, GSFM can be used to reliably assign the most likely functions to understudied genes, and make gene set enrichment analysis more precise, ranking the most relevant enriched terms when presented with any query gene set.”
The research team plans to expand the system by combining GSFM with other AI foundation models. One goal is to integrate it with language-based models to generate natural-language explanations of gene functions. Another future direction is combining GSFM with drug-focused AI models, with the long-term aim of predicting how drugs interact with cells and supporting the design of new therapeutics.
“In summary, GSFM’s ability to distil knowledge from large amounts of unlabeled gene sets automatically, and to do so successfully across multiple sources of knowledge, can be translated into many ‘‘low-hanging fruit’’ hypotheses that could be tested in wet lab experiments to rapidly advance knowledge in biomedical research,” the investigators concluded.
Storytelling is core to humanity’s DNA, stemming from our impulse to express ideals, warnings, hopes, and experiences. Technology has always been woven through the medium and the distribution: from early humans’ innovation of natural pigments and charcoals for cave paintings to literal representation by the camera.
The landscape of storytelling continues to shift under our feet. Social and streaming platforms have multiplied, audiences have fragmented, and our demand for fresh, unique media is insatiable. A recent McKinsey podcast cites that we are watching upwards of 12 hours of video content daily, often on multiple devices and multiple platforms.
All this content is expensive to produce: With a baseline budget of $150M, a Hollywood feature runs $1M per minute of finished film; prestige streaming content is in the hundreds of thousands per minute. And since consumers want to engage with authentic, original material, every company is now effectively a media company. That means we all face the same pressure: more content, with the same time and budget constraints.
There is no longer a question whether to use AI for content; the math doesn’t work any other way. What leaders need to focus on now is how to adapt responsibly, protect brand integrity, uplift team creativity, and build customer trust.
A few things worth holding onto as this era accelerates:
AI amplifies what’s already there, both good and bad. Weak strategy stays weak.
Responsible adoption means knowing what’s in your tools and models. Provenance and transparency are the foundation, not the finish line.
Scale without taste is just noise. Investing in your team’s judgment is what makes more content matter.
Fundamentals of great storytelling have not changed. Regardless of format or channel, what makes audiences lean in are still characters, arc, ingenuity, and surprise.
The permanent sprint
Creative teams are trapped on the endless hamster wheel of production, and it’s not slowing down. According to Adobe research, content demand will grow 5x over the next two years. Social content shelf life is now measured in hours, not weeks. Keeping fresh work in the pipeline is a permanent sprint, requiring teams to rethink how creative production functions.
The first move is freeing creative teams by having AI absorb the repetitive work so they have space for the strategic creative decisions that require human ingenuity. In a recent study from Adobe, 94% of creatives report that AI helps them produce content faster, saving an average of 17 hours per week. That recovered time is not a productivity metric; it is renewed creative capacity.
As a use case, Nestlé offers a useful blueprint. Its teams operate across 180 countries with a portfolio of iconic brands including Nescafé, KitKat, and Purina. Using Adobe Firefly Custom Models embedded in existing content workflows allows teams to generate assets in a brand-informed style without disrupting creative flow. At Nestlé, workflow cycle times dropped 50%. “With Firefly Custom Models, we can react at the speed of culture. It’s the closest thing we’ve had to magic.” says Wael Jabi, global strategic comms lead for KitKat.
As we move into the agentic era, the possibilities expand further. Adobe’s Creative Agent thinks in systems, not tasks, orchestrating across workflows, apps, and processes to close the gap between idea and execution, and get teams out of the production cycles that consume their productivity.
Build for your brand, not every brand
A company’s brand is how the world recognizes and connects with them. And it’s more than a collection of assets—it is dynamic, subjective, and expressed in thousands of micro-decisions made every day by the people who know it best. As production scales, keeping everything tuned to the brand gets more challenging. Off-the-shelf AI cannot replicate the level of nuance creative teams bring to content, and there’s a real cost to getting it wrong; diluting a brand in market with almost-right output is not an acceptable option. Customer trust is fragile.
Starting with a bespoke AI model built with Adobe Firefly Foundry addresses this directly. Firefly Foundry starts with a commercially safe base model and trains further on a company’s IP, making it possible to produce content that genuinely reflects the team’s vision.
And to ensure that Firefly Foundry models truly represent the creatives at the helm, Adobe has partnered with film studios like Wonder Studios, Promise.ai, and B5 Studios, and the “big three” talent agencies CAA, UTA, and WME to deeply understand what it means (and what it takes) to build an IP-immersive model that keeps creatives at the center as these film studios and talent agencies scale their visions. These brand ecosystems can accelerate nearly every phase of the production process, from ideation and storyboarding to production and promotion, all while preserving artistry and authorship. And to power the next generation of creativity and content, Adobe has recently announced a strategic partnership with NVIDIA, delivering best-in-class creative control along with enterprise-grade, commercially safe content at scale.
Generic AI gives teams a starting point. But a model trained on a brand’s own IP gets them to the finish line, while still leaving room for the creative calls that matter most.
When agents become the audience
AI is not only reshaping how we create; it is reshaping how customers find and engage with brands entirely. According to Adobe Digital Insights, AI-powered shopping has surged 4,700%. Agentic web traffic is up 7,851% year over year. Yet, most businesses still have significant gaps in AI-led brand visibility. If content is invisible to AI agents, then a brand is invisible to customers.
Major League Baseball is ahead of this curve. Using Adobe LLM Optimizer, the league monitors how its content surfaces across AI interfaces and makes real-time adjustments to maintain visibility. As fans search for tickets, stats, or game-day experiences, the league ensures its brand shows up wherever that search is happening. And with Adobe’s recent acquisition of Semrush, brand visibility goes even further.
The agentic web created an entirely new content surface that did not exist two years ago, and this exponential proliferation of content illustrates precisely why scaled, on-brand content production has become a strategic imperative. A well-built agentic foundation offers full visibility into (and control over) every piece of content, from production to performance.
How to prepare for AI integration
Here are a few steps to get started:
Audit before automation. Content supply chains usually include duplicated processes, unclear ownership, and assets living in many different places. Before AI can accelerate anything, develop a clear map of how content moves through the organization today: who creates it, who approves it, where it lives, and where it breaks down. AI applied to a broken process just breaks it faster.
Walk through workflows. Resist the urge to overhaul everything at once. Start with production tasks that are high-volume, low-stakes, and well-defined: asset resizing, localization, and background generation. Use those wins to build internal confidence before expanding into more complex creative territory.
Build responsible governance from the start. Governance added as an afterthought becomes a bottleneck. Building it in from the beginning creates a competitive advantage that lets teams move fast with confidence. And this means clear policies on model training, content provenance, human review thresholds, and communicating AI use to customers. The brands that earn lasting trust will treat transparency as a feature, not a footnote.
This content was produced by Adobe. It was not written by MIT Technology Review’s editorial staff.
![An image showing Sebastien Lefebvre, head of technology, data and AI, Aurelis Insights (right) and William Van Etten, PhD, co-founder, CEO & Principal Scientist, StarfleetBio (left) in conversation at Bio-IT World 2026. [Uduak Thomas]](https://www.genengnews.com/wp-content/uploads/2026/05/Seb-Bill-300x233.jpg)
![Morgan Cheatham, MD, Partner, Head of Healthcare & Life Sciences, Breyer Capital (left), and Catherine Brownstein, PhD, Manager of the Molecular Genomics Core Facility at Boston Children's Hospital and Scientific Director of the Manton Center for Orphan Disease Research Gene Discovery Core (right). [Uduak Thomas]](https://www.genengnews.com/wp-content/uploads/2026/05/Cheatham-Bownstein-300x232.jpg)
![Dylan Livingston, founder and president of The Alliance for Longevity Initiatives (A4LI) (left) and William Van Etten, PhD (right) in conversation at Bio-IT World 2026 [Uduak Thomas]](https://www.genengnews.com/wp-content/uploads/2026/05/Livingstone-Van-Etten-300x225.jpg)
