NIH highlights three measures of invisible health risk: metals, genes, speech

Wildfire metals linger

The National Institutes of Health highlighted three recent findings that share a theme: health risks that cannot be seen directly, measured through metals in the body, gene activity in the brain and patterns in speech. The first concerns the 2023 Maui wildfire, which burned through both wildlands and urban areas. Burning built infrastructure releases potentially toxic metals into the environment, and a research team led by Drs. Ruben Juarez and Alika Maunakea studied 1,400 Maui adult residents six to 18 months after the fire. m:…0648

The exposures were large. Antimony levels in study participants were nearly 40 times higher than national averages, manganese levels were nearly four times higher, and barium and arsenic levels were more than twice as high. Participants with the highest metal concentrations lived in the regions of Maui most affected by the wildfire, and the elevated metal burden lasted up to 18 months after the disaster. m:…0648

The health link is lung function. Participants with higher metal mixture levels were more likely to have poor lung function, and arsenic, cadmium and copper contributed the most to the observed effect. The study results were published in the Proceedings of the National Academy of Sciences on September 28, 2026, and the authors' message is that wildfires at the wildland-urban interface contribute to prolonged metal exposures that can harm lung function, so longer-term health monitoring and environmental sampling beyond air quality are necessary to capture wildfire health effects. m:…0648

A brain atlas across disease

The second finding is a single-cell atlas of gene activity in the prefrontal cortex across the human lifespan and eight brain diseases. The study analyzed gene activity in more than 6 million cells from almost 1,500 people, and the research results were published on September 23, 2026, in a series of eight papers. The eight diseases studied were Alzheimer's disease, diffuse Lewy body disease, vascular dementia, Parkinson's disease, tauopathy, frontotemporal dementia, schizophrenia and bipolar disorder. m:…6835

The atlas found gene activity changes common to all eight studied diseases, primarily involving genes related to basic cellular functions. Alzheimer's disease, diffuse Lewy body disease, vascular dementia and Parkinson's disease showed the most similar associated gene activity changes, and severe Alzheimer's disease is associated with fewer neurons and more immune and vascular cells compared with less severe disease. Psychiatric symptoms in Alzheimer's disease are linked with more deep-layer excitatory neurons. m:…6835

The lifespan pattern is equally striking. Healthy brains undergo a period of rapid gene activity remodeling from birth to age 24, followed by midlife stability and renewed changes after age 65. The PsychAD Consortium developed a computational framework to identify cell states associated with Alzheimer's disease, cognitive decline, resilience and psychiatric symptoms, and aims to build a resource representing approximately 10,000 individuals. The atlas is a shared reference, not a treatment, but it gives every future brain-disease study a common map to compare against. m:…6835

Aging in speech

The third finding is the most portable: a speech clock that uses machine learning to estimate biological age. Speech-based clock estimates of biological age were similar to those derived from brain scans or DNA chemical markers, and speech features such as pitch, timing and word choice help detect accelerated aging and dementia. The study analyzed speech recordings of 2,928 native Spanish speakers in five Latin America countries. m:…5012

The clinical signal is specific. Larger speech age gaps correlated with higher levels of the Alzheimer's-related protein p-tau217 in the blood, pitch and emotional content were the most important features for identifying people with predicted biological ages older than their chronological age, and people with Alzheimer's disease had larger speech age gaps than those with mild cognitive impairment. Adversity in the social exposome is associated with higher speech-predicted biological age compared with chronological age, which ties the clock to lived conditions as well as biology. m:…5012

The authors are careful about the limits: long-term studies are needed to determine how well speech-based clocks predict aging-related health problems. Read together, the three NIH-highlighted findings suggest a practical direction for health measurement: the risks that matter most after a fire, across brain diseases and in aging are the ones that linger quietly, and the tools that catch them are the ones that measure bodies, cells and speech over time rather than waiting for symptoms to declare themselves. m:…5012 m:…0648 m:…6835