TL;DR: Baby teeth grow in weekly layers that record everything circulating through the body. A Science Advances study used that record to reconstruct nine-metal exposure week by week from 20 weeks before birth through 40 weeks after — and found two early-infancy windows where higher metal mixtures tracked with later behavioral symptoms and MRI signs of altered brain development.
Key Findings
- Two postnatal vulnerability windows: Weeks 4–8 and weeks 32–42 after birth carried the strongest behavioral signal — not pregnancy, the period most environmental research focuses on.
- β = 0.15 (95% CI 0.004–0.28) for late infancy: Higher metal mixture exposure in weeks 32–42 linked to higher behavioral symptom scores at ages 8–14.
- Three converging MRI signals: Higher exposure tracked with smaller brain volume, lower global network efficiency, and weaker white matter integrity.
- Nine metals modeled together: Mixture analysis instead of single-element studies — closer to actual lived exposure but harder to attribute to one metal.
- Weekly resolution from a tooth: The team reconstructed exposure layer by layer across 60 weeks of development, far sharper than parental recall.
- 489 children, PROGRESS cohort, Mexico City: Prospective birth-cohort design with tooth chemistry, parent-reported behavior, and MRI on the same kids.
Source: Science Advances (2026) | Rechtman et al.
Baby teeth usually end up in a memory box. They also happen to be one of the best chemical diaries a person carries out of infancy. Each tooth grows in layers, and the metals circulating in the body get incorporated week by week — including weeks that began before the child was born.
Most environmental-health research has to rely on memory, blood snapshots, or zip-code-level estimates. A baby tooth is closer to a hard drive.
This Science Advances paper used that hard drive to reconstruct nine-metal exposure across a tightly defined developmental window, then linked the resulting timeline to adolescent MRI and behavior data on the same children.
What Teeth Can Do That Blood and Urine Cannot
Blood and urine measure recent exposure. They are snapshots, and they are blind to anything that happened a year ago, let alone before birth. Teeth solve that problem by accident of biology. As the dental tissue grows, trace metals get layered into the structure in chronological order — fetal layers first, postnatal layers later — at roughly weekly resolution.
The team used that natural archive to reconstruct exposure to a mixture of nine metals across 20 weeks before birth through 40 weeks after birth. That is the developmental window during which brain growth, sensory learning, immune calibration, and circuit refinement do not move at one steady pace. They surge and pause and lock in at different times.
The children came from the PROGRESS birth cohort in Mexico City, a long-running study built to connect early exposures with later health. By adolescence, the researchers had three unusually well-aligned datasets on the same children: tooth-based exposure histories, parent-reported behavior, and MRI scans.
The Strongest Windows Were Not in Pregnancy
Most environmental-developmental science centers on pregnancy. Pregnancy genuinely matters. But this study’s headline result was that timing carried information — and the strongest behavioral links lived after birth, not before.
Two windows stood out. One around weeks 4–8 after birth, when the developing brain is rapidly elaborating early sensory and emotional circuitry. Another around weeks 32–42 after birth, late infancy, when motor, language, and attention systems are integrating in ways that depend on a specific neurochemical and metabolic environment.
The cleanest single number came from that late-infancy window. Higher metal mixture exposure during weeks 32–42 was associated with higher behavioral symptom scores in adolescence (β = 0.15, 95% CI 0.004 to 0.28). Only about 4% of children scored in the clinical range, so this is not destiny — it is a small population-level shift that matters because exposure is widely distributed.
MRI Tracked the Same Windows
The behavioral findings would be easier to dismiss in isolation. They did not stand alone. The same metal-exposure windows linked to MRI measures pointing to altered brain organization across three converging signals:
- Smaller brain volume: a broad structural signal, not an isolated regional hit.
- Lower global network efficiency: a graph-theory measure suggesting information moves less easily across the brain network.
- Weaker white matter integrity: structural disruption in the long-range fiber tracts that connect distant regions.
The convergence is what makes the result more than an exposure screen. The same children with higher early-infancy metal exposure showed both behavioral symptoms and brain-organization differences in adolescence — measured independently, lining up against the same exposure windows.
Mixtures, Not Just Lead
Lead gets most of the public attention in conversations about childhood metal exposure, and for understandable reasons. But children do not encounter exposures one variable at a time. They encounter mixtures, through food, drinking water, air, housing, soil, and the built environment. Some of those metals are toxic at low levels; others are essential nutrients that become disruptive at the wrong dose, timing, or balance.
This study modeled the exposure profile as a mixture, which is closer to lived biology. That choice also makes single-metal blame harder. A mixture analysis can identify vulnerable windows and combined risk, but it cannot cleanly say “this metal caused this MRI feature.” That limit is not a flaw — it is a fairer reflection of what the developing brain actually faces.
Why This Study Is Hard to Wave Away
Environmental-health studies are usually hampered by imperfect memory. Parents can know that a child lived near traffic, drank well water, or spent time in older housing, but they cannot reconstruct chemical exposure week by week during infancy. The tooth method removes that ceiling. It creates a biological timeline rather than a retrospective estimate.
The method is not infallible — it captures metals incorporated into dental tissue, not every exposure route directly, and it requires careful calibration. But it gives researchers a much sharper temporal lens than most early-life exposure designs, sharp enough to move from “early metal exposure matters” to “which weeks appear most sensitive.”
That distinction is policy-relevant. Prevention is often won by timing. A narrow infancy window is easier to protect than an undefined childhood exposure span.
Possible Implications of this Study
This is an observational cohort study, so it cannot prove that reducing one exposure during one specific week would prevent a later behavioral symptom. Families should not read it as cause for panic about any single test result, and clinicians should not extrapolate to individual cases beyond their evidence.
The actionable message is structural. If exposure timing matters at this resolution, then environmental protection during pregnancy and infancy needs sharper windows than broad early-life slogans. Reducing avoidable metals in water, food, housing, and consumer environments stops being only a toxicology issue and becomes a brain-development issue.
For clinicians evaluating attention, mood, or anxiety symptoms in children, the paper adds another reason to take environmental histories seriously.
For researchers, it shows that baby teeth may become a powerful archive for connecting early exposures to later brain outcomes — a chemical diary that can still be read years after the tooth has fallen out.
Citation: Rechtman et al. Fetal and postnatal metal metabolism-related changes in brain function are associated with childhood behavioral deficits. Science Advances. 2026;12(17):eadz1340. DOI: 10.1126/sciadv.adz1340
Study Design: Prospective birth-cohort analysis linking tooth-based exposure reconstruction, behavioral assessment, and MRI.
Sample Size: 489 children aged 8–14 from the PROGRESS Mexico City cohort.
Key Statistic: Higher metal mixture exposure in weeks 32–42 after birth associated with higher behavioral symptom scores (β = 0.15; 95% CI 0.004–0.28); MRI changes in volume, network efficiency, and white matter integrity tracked the same windows.
Caveat: Observational design — identifies exposure windows and associations but cannot prove single-metal causation.
