This article has been reviewed according to Science X's editorial process and policies . Editors have highlighted the following attributes while ensuring the content's credibility:
American football participation was associated with worse cognitive and neuropsychiatric function in later life. Credit: PxHere
What we learned this week: Left-handed people may have a psychological edge in competition. Humanoid robots can now do creepy parkour through the uncanny valley. And if you've ever cared for an elderly cat, a new study highlights a biological quirk that could explain why they're so prone to kidney disease.
Additionally, archaeologists studied the impressive and sophisticated Mayan water-management system; a new study finds that even nonconcussive head impacts can lead to brain damage over time; and researchers report that LLMs are already sophisticated enough to identify anonymous social media users with reasonable accuracy:
Don't drink this civilization-threatening beverage
Over a six-year period, researchers studied the ancient water system of Ucanal, a once densely populated pre-industrial Mayan city, finding surprisingly sophisticated engineering and water management that endured for 1,500 years—but with a hidden threat the Mayans could not have known about. The researchers studied three excavated reservoir sites in northern Guatemala. The Aguada 2 reservoir served the wealthy side of the city; Aguada 3 served lower classes; Piscina 2 was a reservoir connected to the city drainage system.
The researchers found no evidence of cyanobacteria contamination at the sites, indicating that the Mayans understood biological water contaminants and mitigated the growth of the blue-green algae. Aguada 2 was built on high ground and used rock-filled inlets to filter sediment and waste naturally. And the researchers believe the site was surrounded by vegetation, which would have kept it cooler and thus free of bacterial growth. Analysis of biomarkers at the three sites indicate effective human waste management.
However, there was one threat the Mayans couldn't detect: mercury contamination. The use of cinnabar, a bright red, mercury-sulfide-based pigment, was widespread by the Terminal Classic period and mercury contamination at the site increased by 300%. Since mercury poisoning is not detectable via the senses, the Mayans would have been unaware of the threat.
Are you ready for some chronic traumatic encephalopathy?
We're well into decade three of studies into repetitive head impacts in American football and their enduring effects on brain health, and a new project involving former players over age 40 finds significantly poorer performance on memory and thinking tests. Additionally, these participants had greater concerns about their cognitive abilities and reported higher levels of depression symptoms than subjects with no history of repetitive head injury. The participants played at all levels, from youth and high school teams to college and professional leagues. Outcomes worsened with more exposure to football.
The risk extends beyond diagnosed concussions. Researchers believe repeated low-level impacts, called subconcussive hits, jeopardize long-term brain health and contribute to chronic traumatic encephalopathy. The significance of this study lies in the deeper field of participants, as previous studies have focused primarily on players at the professional level and thus produced findings that aren't easily generalizable to the broader population. The findings suggest a dose-response relationship between years played and progressively worse outcomes.
Discover the latest in science, tech, and space with over 100,000 subscribers who rely on Phys.org for daily insights. Sign up for our free newsletter and get updates on breakthroughs, innovations, and research that matter—daily or weekly.
They will know you by your reposts
A new study suggests that artificial intelligence could bring the era of online anonymity to a rapid close. To determine whether a large language model could de-anonymize social media users, researchers developed a model to replicate and automate the decision-making processes of a human investigator. The AI trains on a specific user's posts on either Reddit or Hacker News, gathering comments, jokes, biographical details like education history and individual writing quirks. The researchers knew the identities of many of the users, stripping names, links and other identifiers from the training data.
The model then develops a mathematical representation of the person's profile based on this accumulated microdata and searches for candidate matches on sites like LinkedIn, assigning a confidence score to predicted matches. Their framework successfully linked accounts with 67% accuracy.
The authors conclude, "Pseudonymity does not provide meaningful protection online. Users who post under persistent usernames should assume that adversaries can link their accounts to real identities or to each other, and that the probability rises with each piece of micro-data they post."
Written for you by our author Chris Packham, edited by Gaby Clark, and fact-checked and reviewed by Robert Egan—this article is the result of careful human work. We rely on readers like you to keep independent science journalism alive. If this reporting matters to you, please consider a donation (especially monthly). You'll get an ad-free account as a thank-you.
© 2026 Science X Network
— Source: Phys.org (https://phys.org/news/2026-03-saturday-citations-bad-news-footballers.html)
