“Drinking diet soda during pregnancy linked to autism,” the New York Post and other media outlets similarly declared last month, after the University of Texas Health Science Center at San Antonio put out a press release that touted new findings by some of its researchers. According to the study, boys whose mothers reported drinking diet soda—or otherwise consumed a comparable amount of the artificial sweetener aspartame—every day during pregnancy or breastfeeding were about three times more likely to have been diagnosed with autism. (Researchers found no statistically significant trend among girls, who are diagnosed with autism at far lower rates.)

But a few days after the press release, the London-based Daily Mail ran an article in which experts from Harvard and Bournemouth University in the United Kingdom called the San Antonio researchers’ data “extremely flawed” and their conclusions “irresponsible.” Among other criticisms, they noted that the data relied on each mother’s recollection rather than real-time observation or contemporaneous notes. The critics also pointed out that the sample size was small and that observing a correlation doesn’t prove that artificial sweeteners cause autism.

These and other limitations had already been disclosed in the study and recognized in the press release. “There is no proof of causation,” agreed Sharon Fowler, the study’s lead author and an adjunct assistant professor of medicine at UT Health San Antonio. Nevertheless, Fowler, who has studied the health impacts of chronic exposure to non-nutritive sweeteners for more than a decade, is convinced that aspartame poses serious health risks. She maintains that this latest study’s findings should lead mothers to think twice before consuming artificial sweeteners when they’re pregnant or breastfeeding.

The debate over the significance of these findings may have more to do with how science gets communicated to the general public, and how media outlets report the nuances of that information, than with the specifics of the study. “I don’t think we’re communicating as effectively as we should be,” says Robin Kochel, an associate professor at Baylor College of Medicine and associate director for research at Houston’s Texas Children’s Hospital Autism Center, who was not involved in the San Antonio study. “We always tell families that scientists believe autism is likely caused by a combination of genetic and environmental factors. But what we know about environmental factors is really kind of in its infancy, compared to what we know about genetic influences.”

Genetic studies have long demonstrated that DNA plays a major role in a child’s development of autism. With the help of advancing technologies, medical geneticists can order relatively straightforward tests that indicate whether a child carries genetic mutations commonly found in autism patients. It is much more difficult, bordering on impossible, for scientists to control for the countless other environmental factors that the parent or child has been exposed to—foods, beverages, air quality, medications, etc.—in order to isolate a single element that may have increased a child’s risk of developing autism. “The bulk of the evidence is pointing towards genetics,” Kochel says.

The UT Health San Antonio findings, which were first published in August in the journal Nutrients, were drawn from the written recollections of 235 mothers of children with autism spectrum disorder, along with a control group of 121 mothers of children with typical neurological development. Kochel notes that the study’s modest sample size was made even smaller when researchers drew conclusions by gender and by severity of diagnosis. Additionally, some of the study participants were mothers of adult children with autism and trying to recall daily behavior from many years earlier—even decades before, in some cases.

Kochel also says that the results may have been muddied because about half the control group of children (those without autism spectrum disorder) had a sibling with autism participating in the study, meaning that in some instances the same mother was providing responses as both a “case” and a “control”—in an ideal study these would be separate individuals. Furthermore, she adds, since children’s neurological development is considerably different when they’re in the womb as opposed to after birth, the study would have been improved by focusing on sweetener consumption either during pregnancy or during breastfeeding, rather than lumping the two categories together.

Fowler, for her part, recognizes that this study, like many of its kind, is small and has significant limitations. She emphasizes that further research is needed and agrees that the results would be more convincing if researchers had access to more comprehensive data, but she doesn’t think that’s an argument against releasing the findings. “If there’s some kind of warning on the horizon that something is happening, do you go ahead and share that warning, even though it’s imperfect, as a precaution for women who are drinking these substances now?” she asks rhetorically. “Or do you wait until the perfect study has been done?”

She and her colleagues are working on related research, using a much broader set of data that was collected in real time, that she hopes will be published in the next several months. To bolster her argument for the avoidance of aspartame, Fowler cites past studies in both animals and humans that found a link between artificial sweeteners and poor neurological or cardiometabolic outcomes. In July, the International Agency for Research on Cancer determined, based on “limited evidence,” that aspartame is “possibly carcinogenic to humans.” But the U.S. Food and Drug Administration disputed that classification.

Research related to the causes of autism has led to heated controversy since at least 1998, when a since-debunked study linked autism to the measles, mumps, and rubella vaccine. That study sparked a decades-long misinformation campaign that experts say resulted in widespread vaccine hesitancy, leading to devastating health outcomes, including hundreds of thousands of deaths during the pandemic. Fowler admits to “internal wrestlings” over when was the appropriate time to announce her team’s findings, but, particularly considering what she views as a convincing body of research related to aspartame and poor health outcomes, she felt the risk of further harm from artificial sweeteners far outweighed the potential for the spread of misinformation.

Given the limitations of this study, the difficulty in identifying environmental risks related to autism, and the ambiguous information coming from national and international authorities on artificial sweeteners, Kochel recommends “everything in moderation.” She says to talk to your health care providers if you have specific concerns. Don’t worry too much about an occasional diet soda, but more importantly, keep in mind what we know to be components of a healthy lifestyle: plenty of water, whole foods, exercise, and sleep.