Conventional studies may overestimate the range of long-term COVID symptoms
A Stanford-led analysis of 14.4 billion American health insurance claims found that conventional study designs may overestimate how many long-term health problems can be attributed to COVID-19.
The study examined two sources of bias in such studies: comparing positive COVID testers with people who were never tested, and grouping negative-testing subjects in hindsight instead of at the time of the test. These biases caused studies to attribute unrelated symptoms to COVID, researchers found, such as a 76% higher risk of future injury from causes including lightning and drowning.
By enrolling subjects at their first polymerase chain reaction (PCR) test, then grouping them by positive or negative result, the study found 98% fewer COVID-attributable outcomes from four to 12 months. Myopathy, hair loss, pulmonary embolism and respiratory failure stood as the only effects with substantially elevated risk. By one year, nothing was clearly attributable to COVID.
The authors stress that these findings do not mean long COVID is 98% less common, or that certain symptoms are not genuine; only that it is unclear whether COVID causes them.
Proceedings of the National Academy of Sciences (PNAS) published the study on Oct. 27. Computer science professor Jure Leskovec co-senior-authored the paper.
New findings propose model for thermal noise communication
In a new study, Stanford engineers offered a first-principles model for thermal noise communication. This understanding contributes to the development of sensors used in environments with limited energy supply, such as agricultural sensing.
Thermal noise communication uses background static in electrical conductors, produced by electrons jiggling with heat, to carry data. A thermal noise transmitter doesn’t need to generate its own signal, thereby using much less energy than a radio transmitter.
The new study found that the receiver acts like an “electron refrigerator,” making thermal noise appear cold. Meanwhile, the transmitter alternates between an “absorbing” and “reflecting” state. Depending on the state, the receiver reads either a “warm” or “cold” signal, without either part actually heating or cooling. Using this explanation, the study proposed how to best design devices using these signals.
Proceedings of the National Academy of Sciences (PNAS) published the study on Oct. 20. Senior author Zerina Kapetanovic, an electrical engineering assistant professor, demonstrated thermal noise communication in previous research before contributing to its explanation in this study. Geneva Ecola and Jida Zhang, two Stanford electrical engineering grad students, first-authored the paper.
‘Hard’ X-ray pulse measurements allow for precise electron motion imaging
Researchers at the SLAC National Accelerator Laboratory, which Stanford operates for the Department of Energy, measured the shortest high-energy X-ray pulses ever recorded at an X-ray free-electron laser. The record marks a milestone in attosecond science, the effort to observe electron motion over ultrafast timescales.
Nature Photonics published the paper on Oct. 20 after accepting it in September. Ichiro Inoue, a professor at the University of Tokyo, led the study. Co-author and SLAC professor Agostino Marinelli led a 2019 study that generated attosecond X-ray pulses under 5 keV, referred to as “soft” pulses, before recording the “hard” pulses over 5 keV in this study.
While X-rays normally damage their samples, these ultrashort pulses capture them before damage occurs, allowing researchers to image electron motion deep inside atoms and dense plasmas.
The team found that “hard” pulses last between 100 and 400 attoseconds, or billionths of a billionth of a second. This measurement confirmed that these pulses have the right duration for attosecond imaging, giving researchers the confidence that resulting images aren’t smeared or unclear.