The scale of the underlying problem is enormous. Rare diseases are individually defined as affecting fewer than 1 in 2,000 people, but more than 7,000 distinct types are currently recognized, and recent estimates suggest the true number may exceed 10,000. Collectively, rare diseases affect roughly 3.5 to 5.9 percent of the global population, equivalent to between 263 and 446 million individuals. Nearly half of all rare diseases are neurological in nature, and around 90 percent of rare childhood diseases involve major neurological manifestations. Approximately 80 percent of rare diseases are genetic in origin, nearly 70 percent of those present during childhood, and about 30 percent of affected children die before their fifth birthday. These figures underscore why rare neurological diseases in children represent a disproportionate share of the global rare disease burden.

To isolate this population, the research team applied three criteria to the GBD 2021 cause hierarchy: relevance to children aged 0 to 14, a prevalence threshold of no more than 50 per 100,000 people, and availability as a distinct disease category permitting population-level estimation. Four groups emerged: motor neuron disease, multiple sclerosis, neuroblastoma and other peripheral nervous cell tumors (NPNCT), and a residual category of other neurological disorders that largely comprises inherited neuromuscular conditions such as muscular dystrophies and myopathies. The team computed age-standardized prevalence, mortality, and disability-adjusted life-year (DALY) rates, stratified by age, sex, and sociodemographic index (SDI), and quantified temporal trends using the average annual percentage change (AAPC).

The headline numbers tell a story of expanding prevalence. Globally, the number of prevalent cases rose by 21.98 percent, from roughly 80,785 in 1990 to 98,541 in 2021. The age-standardized prevalence rate climbed 7.17 percent, from 4.60 to 4.93 per 100,000 population, with an AAPC of 0.21 percent. DALYs increased even faster, rising 30.46 percent from 1.01 million to 1.32 million, with the age-standardized DALY rate up 14.92 percent. In contrast, the age-standardized mortality rate fell 6.98 percent, from 0.43 to 0.40 per 100,000. Children under five carried the heaviest burden in 2021, with the highest prevalence, DALY, and mortality rates of any age group, and prevalence rates rose across all pediatric age groups over the study period.

The four disease groups followed strikingly divergent trajectories. NPNCT had the highest prevalence rate in 2021 at 2.62 per 100,000 and the steepest increase, with an AAPC of 0.54 percent, and it was the only condition whose mortality rate rose. Other neurological disorders, dominated by monogenic neuromuscular diseases, were the leading contributor to disability alongside NPNCT, with a DALY rate AAPC of 0.73 percent. Multiple sclerosis showed a statistically significant but epidemiologically modest rise in prevalence. Motor neuron disease, by contrast, declined across all three measures, with its prevalence rate falling 6.37 percent. The authors attribute this decline partly to advances in prenatal diagnosis, newborn screening, and disease-modifying therapies for conditions such as spinal muscular atrophy, though they caution that uncertainty in GBD estimates for rare pediatric motor neuron diseases is considerable.

Boys consistently bore a higher burden than girls across prevalence, DALYs, and mortality, and the gap widened over time, with boys’ prevalence rate rising faster than girls’. The sex disparity echoes broader GBD 2021 findings across 37 neurological conditions, but this analysis confirms it emerges early in childhood. The authors note that the difference may partly reflect biological factors but also unequal healthcare access and diagnostic ascertainment between boys and girls, particularly in low-SDI countries, and they urge caution in interpretation.

The geographic patterns are perhaps the most consequential finding. High-SDI countries recorded the highest overall rates in 2021, at 9.33 per 100,000 for prevalence, yet all their rates have declined steadily since 1990, plausibly reflecting improvements in early diagnosis, multidisciplinary care, and supportive treatment. The opposite occurred in lower-development settings: middle-, low-middle-, and low-SDI countries all experienced rising prevalence, disability, and mortality, with the pace of increase inversely related to SDI. Low-middle-SDI countries saw the fastest prevalence growth, with an AAPC of 0.87 percent. Across countries, SDI values correlated strongly with prevalence rates, at r = 0.91. South Asia recorded the sharpest regional rise in prevalence, while Western sub-Saharan Africa saw the largest increases in disability and mortality. At the national level, 116 of 204 countries and territories showed rising prevalence, with Guyana experiencing the steepest increase, and India, China, and Pakistan reporting the highest absolute case counts in 2021.

Decomposition analysis revealed what is driving these increases. Population growth accounted for 73.33 percent of the global rise in prevalent cases, with true change in prevalence contributing 35.28 percent. In low- and low-middle-SDI countries, demographic expansion was the dominant force, whereas in middle-SDI regions and several high-case countries, rising prevalence itself played the leading role. The authors interpret the rising recorded prevalence largely as a product of diagnostic progress: newborn screening, next-generation sequencing, and single-molecule sequencing have dramatically improved case ascertainment, with roughly 250 to 280 newly described rare disorders reported annually since 2010. Longer survival among affected children, enabled by expanding targeted therapies, further inflates measured prevalence and DALYs even as mortality falls.

Looking forward, the team built a customized forecasting framework combining Poisson regression on SDI with an autoregressive integrated moving average model applied to regression residuals, integrating four models through root-mean-squared-error-weighted averaging, with ridge regularization and a quadratic SDI term to improve stability. Validated against withheld 2011 to 2021 data, the model achieved a mean absolute percentage error of just 0.62 percent. Its projections indicate that by 2050, prevalent cases will reach approximately 110,200, an 11.83 percent increase from 2021, with the age-standardized prevalence rate climbing to 5.62 per 100,000. Children under five will see the fastest growth, and Western sub-Saharan Africa, Southern sub-Saharan Africa, and South Asia the sharpest regional rises, with 132 countries expected to show increasing prevalence trends. The authors acknowledge that future gene therapies and expanded genomic screening could alter these trajectories in ways the model cannot capture.

To complement the epidemiological picture, the researchers mined ClinVar for genetic evidence of rare monogenic neurological diseases, compiling a curated list of 520 disease-causing genes. Between 1990 and 2024, the database recorded 65,331 pathogenic or likely pathogenic variants in these genes. The acceleration is dramatic: only 1,633 variants were reported between 1990 and 2010, compared with 18,851 in the following decade and 44,847 in just 2021 to 2024, alongside 55 newly reported disease-causing genes in the most recent period alone. The largest genetic evidence base lies within the other neurological disorders category, dominated by inherited neuromuscular diseases, of which more than 600 monogenic forms have now been identified. The authors stress that these figures reflect testing availability and submission practices, which skew toward high-income countries, rather than the true underlying genetic burden. Even so, the convergence of rising prevalence, accelerating genetic discovery, and widening global inequality paints a clear picture: rare childhood neurological diseases are becoming a growing and increasingly tractable global health challenge, one that will demand strengthened surveillance, equitable diagnostic access, and sustained investment in the low- and middle-income countries where the burden is now rising fastest.

Subject of Research: Global epidemiological and genetic burden of rare neurological diseases in children

Article Title: Global epidemiological and genetic burden of rare neurological diseases in children

Article References: Xie, Z., Sun, C., Liang, W., Zhao, X., Miao, L., Liu, X., Lu, Y., Liu, C., Yin, Z., Li, J., & Xie, Z. (2026). Global epidemiological and genetic burden of rare neurological diseases in children. iScience, 29(11), Article 117791. https://doi.org/10.1016/j.isci.2026.117791

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117791

Keywords: rare diseases, neurology, pediatrics, Global Burden of Disease, genetics, ClinVar, neuromuscular diseases, neuroblastoma, health inequality, epidemiology, DALYs, disease forecasting

The study, published in the International Journal of Data Science and Analytics, starts from a candid acknowledgment of a problem that plagues much of agricultural machine learning: the labels. True pathological observations of infection are rare and expensive, so the researchers instead derived their disease-risk annotations from fungicide-treatment records, the practical decisions growers made to protect their vines. This makes the labels management-oriented risk proxies rather than direct measurements of pathogen presence. The team is explicit about this distinction, and it matters, because a treatment record reflects a grower’s judgment as much as the biology of the disease itself. Any model trained on such data learns to predict management behavior conditioned on weather, not infection events in the strict epidemiological sense.

The proxy-label problem had a second wrinkle. The dataset contained separate annotations for downy mildew and powdery mildew, but the two agreed on 98.9 percent of observed days, and in the validation and test years they were identical. Rather than pretending the two diseases share an epidemiology, which they demonstrably do not, the researchers collapsed the labels into a single operational any-mildew target. This pragmatic choice reflects how growers actually experience the threat: either mildew is a concern today or it is not. But the authors are careful to note that the combined target cannot support claims about distinguishing the two diseases, and biological interpretation of model outputs must remain cautious.

With the target defined, the team built an event-based formulation. A new risk event is counted only after a minimum disease-free gap has elapsed, preventing a single prolonged risky period from fragmenting into dozens of trivial episodes. Models are then asked a genuinely anticipatory question: will a new event begin within the following three to seven days? This framing, borrowed from the early-prediction literature on time series, directly measures what daily classification cannot, namely whether a model can see an episode coming before it starts. The data were split chronologically, with 2020 and 2021 for training, 2022 for validation, and 2023 for testing, ensuring that models were always forecasting into the future rather than interpolating within a shuffled mixture of seasons.

The model comparison spanned the standard spectrum of tabular and sequence learners. Logistic regression provided a linear baseline, XGBoost represented gradient-boosted decision trees, and two deep sequence architectures, long short-term memory networks and temporal convolutional networks, captured temporal dependencies in the agro-meteorological streams. Crucially, the researchers compared three input representations: calendar-only features such as day-of-year, environmental-only variables from the vineyard sensors, and the two combined. This design allowed them to isolate how much predictive power genuinely comes from the weather, a question that many published disease-forecasting studies never ask.

The results were humbling for the environmental data. Calendar-only models were highly competitive, with calendar-based LSTM and TCN models detecting 7 of the 8 test events, while a simple monthly climatological baseline, essentially a lookup table of historical risk by month, still detected 6. Environmental-only models were substantially weaker across the board, and adding environmental variables to calendar information produced inconsistent gains that varied by model family rather than delivering a uniform improvement. In other words, much of what the sophisticated models learned may have been the seasonal rhythm of vineyard disease risk, a pattern that a calendar already encodes. The authors’ SHAP analysis, which attributes model predictions to input features, confirmed that the fitted models used both seasonal and meteorological information, but it did not reveal a single dominant or consistently interpretable environmental precursor, no silver-indicator variable that reliably signals an impending outbreak.

This seasonal confounding is arguably the study’s most important cautionary finding. Mildew risk in temperate vineyards follows the growing season, so any model with access to the date can achieve respectable performance without understanding humidity, leaf wetness, or temperature at all. Without explicit seasonal baselines, reported accuracies can flatter models that are merely reproducing climatology. The Skopje team’s insistence on comparing against calendar-only and climatological baselines offers a template for how disease-forecasting studies should be evaluated, and their results suggest that several celebrated successes in the literature may deserve a second look under this stricter light.

The event-based formulation also revealed trade-offs that daily classification obscures. When the researchers compared their event-based models against conventional daily-status classifiers, they found the two objectives capture different things and exhibit different profiles in point-wise accuracy, event coverage, and alert burden. A model can score well on daily labels while missing the onset of episodes entirely, or catch most onsets while flooding the grower with false alarms. Neither formulation proved uniformly superior, which means the choice between them should be driven by the operational goal: a grower deciding when to scout needs anticipation, while a system scheduling routine monitoring may prefer stable daily estimates. Sensitivity analyses reinforced this picture, showing that results depend materially on the historical window length, the prediction horizon, and the event-separation gap, parameters that previous studies often fixed arbitrarily.

For the wine industry, the practical implications are sobering but useful. The best models anticipated most risk episodes several days ahead, which is genuinely valuable lead time for scheduling sprays and scouting, and targeting interventions more precisely can reduce fungicide use, cost, and environmental impact. But the finding that environmental sensors added inconsistent value over the calendar suggests that in-vineyard IoT weather stations, increasingly common in precision-agriculture deployments, must justify their cost through demonstrable gains over seasonal baselines on each new dataset, not through generic claims. The study also underscores the value of the underlying agro-meteorological dataset introduced by Arvanitis and colleagues in 2025, whose IoT-derived environmental records made the analysis possible, and the Skopje team states that their processed data, event annotations, engineered features, and experimental scripts will be released in a public repository, with interim availability from the corresponding author.

More broadly, the work contributes to a growing movement in machine learning toward event-oriented evaluation, in which systems are judged not on per-instance accuracy but on their ability to anticipate discrete, consequential occurrences, whether disease outbreaks, forest fires, or equipment failures. By defining events explicitly, enforcing minimum gaps, and measuring event coverage against alert burden, the vineyard study offers a methodological blueprint that extends well beyond oenology. Its honest treatment of proxy labels, its insistence on seasonal baselines, and its refusal to overclaim interpretability from SHAP attributions model the kind of rigor that agricultural AI will need as it moves from papers to practice. The next mildew wave may still be hard to predict, but thanks to this work, we now know better how to ask the question, and better how to tell whether an answer is real.

Subject of Research: Event-based machine learning early warning of vineyard mildew disease risk from agro-meteorological time series

Article Title: Event-based early warning of vineyard disease risk from environmental time series

Article References: Dimitrovski, I., Kitanovski, I., Davcev, D., Kalajdziski, S., & Mitreski, K. (2026). Event-based early warning of vineyard disease risk from environmental time series. International Journal of Data Science and Analytics, 22(1), Article 317. https://doi.org/10.1007/s41060-026-01282-8

Image Credits: AI Generated

DOI: 10.1007/s41060-026-01282-8

Keywords: machine learning, viticulture, downy mildew, powdery mildew, early warning systems, agro-meteorological time series, LSTM, temporal convolutional networks, XGBoost, SHAP, precision agriculture, disease forecasting