PTSD prediction, phishing while tired, an exoplanet : NPR

The regular news roundup with NPR’s science podcast Short Wave covers a possible use for AI in emergency rooms, a distant planet and the link between cybersecurity and Daylight Saving Time.

SCOTT DETROW, HOST:

It’s time now for our science news roundup from Short Wave, NPR’s science podcast. I am joined by the show’s hosts, Emily Kwong and Regina Barber. Hello to both of you.

REGINA BARBER, BYLINE: Hey.

EMILY KWONG, BYLINE: Hey.

DETROW: You were in stereo just now.

BARBER: Yeah.

DETROW: It was great.

BARBER: We planned that.

DETROW: So you’ve brought us three science stories that caught your attention. What are they?

KWONG: How AI can help emergency room workers identify patients who may develop PTSD.

BARBER: A new technique that helps us find really faint planets beyond our solar system.

DETROW: You know how I feel about space.

BARBER: That’s right.

KWONG: We put it in there for you.

BARBER: That’s why I picked it.

KWONG: (Inaudible) put this story in there for you (laughter).

BARBER: Yeah.

KWONG: And why daylight saving time could be bad for cybersecurity.

DETROW: All right. Let’s start with – this is interesting. There’s been so many stories about AI identifying healthcare trends that might be hard for doctors to spot. Let’s talk about this. Identifying…

KWONG: Yeah.

DETROW: …PTSD in emergency departments.

KWONG: This is very interesting. So researchers took trauma patients answers’ to a routine screening question – what brought you to the emergency department today? They fed the answers to that question into an AI model that was designed to predict the patient’s risk of developing PTSD a month later.

DETROW: Huh. I mean, how good was the model making that prediction?

BARBER: So it got it right around 95% of the cases. A psychologist and lead researcher, Katharina Schultebraucks, says that these results are exciting because they’re using a patient’s story in their own words.

KATHARINA SCHULTEBRAUCKS: Some people talk, like, very elaborative of what happens. Some are in such high distress they may only say a few words and a few sentences. And we are looking not on PTSD symptoms right now, but we’re looking in what happens in the future.

BARBER: They published their findings in the journal Nature Communications.

DETROW: I’m curious, is this something that clinicians have been missing?

KWONG: Yeah, so when trained clinicians are screening for PTSD risk, they do tend to find it. The problem is that giving a PTSD risk assessment is an extra step that a lot of emergency departments just simply do not have time for. You know, they’re, like, tending to the emergency. There’s also the stigma around mental health that might clam up some patients. So using this model could streamline the process, and medical professionals can act sooner to prevent PTSD from developing.

BARBER: And Scott, what’s unique about this work is that the model gave its reasoning at each step of the analysis.

MUNMUN DE CHOUDHURY: What this work does, it says that the AI doesn’t have to be a black box.

BARBER: So this is Munmun De Choudhury, a computer scientist not involved in the work. She said this work has real potential to guide clinicians to the right kind of interventions.

DETROW: Perpetual question with this stuff – are there worries about privacy of sending a patient’s words to an AI model?

BARBER: Yeah.

KWONG: Good question. Katharina says that the model did not share the data on a cloud service and was HIPAA-compliant. But she adds that it is important any patient consents to having their information recorded and used by an AI model.

DETROW: So new topic.

BARBER: Yeah.

DETROW: Gina.

BARBER: Yeah.

DETROW: Exoplanet.

BARBER: (Laughter) Yeah.

DETROW: New exoplanet. Let’s – this is a – meaning a planet beyond our solar system.

BARBER: Yeah.

DETROW: What is this technique that scientists used to find it?

BARBER: So this exoplanet was actually already known. A team of researchers studied existing images of the star it orbits to see if light from 51 Eridani b could be found, just to kind of verify this technique. And this is important because often it’s hard to see exoplanets in images.

KWONG: And that’s because the light that exoplanets reflect from the closest star is often much fainter than the light the star gives off.

DETROW: Right.

KWONG: So the star, like, kind of hides the planet.

DETROW: I mean, so how did they uncover it, then? Because again, stars, very bright.

BARBER: Yeah.

DETROW: Planets, when they’re far away, not very bright.

BARBER: So they subtracted light from the image until they could clearly see the planet and not the stars. The paper was published in Nature last week. And this new technique corrects that issue and corrects for millions of more issues that can make it harder to find an exoplanet, like the telescope being unstable or instrumental errors or temperature fluctuations.

DETROW: Is this kind of like in a – in, like, a cops TV show where they’re looking at the security camera and they say, oh, enhance on that one image, and they zoom in?

BARBER: I mean, kind of (laughter). But Scott, what I’m actually surprised about is that they did all of this with just physics knowledge and algorithms and no AI.

DETROW: No AI.

BARBER: Yeah.

DETROW: All right.

KWONG: (Laughter).

DETROW: So what can these images tell us about this planet?

KWONG: So yeah, because these images were taken in infrared by the powerful James Webb Space Telescope, they can tell us more about what these exoplanets are made out of.

RODRIGO FERRER-CHAVEZ: With this, it is the first time that we can actually get this specific information-rich image of this planet that might tell us a lot about how it formed.

KWONG: That’s lead author Rodrigo Ferrer-Chavez, a Ph.D. candidate at Northwestern University. He also says that this tool can be used to find even more exoplanets hiding out in James Webb Space Telescope images.

BARBER: This technique can also be used to make future telescopes, that are already supposed to be more powerful, even better at finding exoplanets. But I do want to say that other astronomers I talked to said that this is just one cool technique of many that scientists are developing for future research in finding these faint, distant worlds.

DETROW: OK. All right, back here on Earth. Third story. It’s about sleep.

KWONG: Mm-hmm.

BARBER: Mm-hmm.

KWONG: And daylight saving.

BARBER: Yeah. So, Scott, did you know that nearly one-third of U.S. adults don’t get enough sleep?

DETROW: Honestly, I would assume the number is higher.

BARBER: (Laughter) Yeah.

DETROW: One-third actually sounds promising (laughter).

BARBER: Yeah.

KWONG: Oh, gosh.

BARBER: And we know that it’s bad for our health, right? It’s bad for decision-making.

KWONG: But just how bad is it for decision-making if you lose a modest amount of sleep? That’s what Eitan Schechtman, an assistant professor at the University of California, Irvine wanted to know. And he found a fun way to test this with phishing attacks.

BARBER: (Laughter).

KWONG: His university regularly sends fake phishing emails to staff to train them on cybersecurity. These emails contain suspicious links you are not supposed to click. But would they click if they were a little sleepy?

DETROW: Interesting. Did this become a full experiment?

BARBER: Yeah, so Eitan worked with the university’s security team to divide 11,000 participants into two groups. Some people got the fishing email the Monday before daylight saving, and some got it the Monday immediately after, you know, when we can assume some may have lost an hour of sleep. So people emailed after daylight saving were 13% more likely to fall for the phishing attack the following morning and click that bad link. The study was published in the Proceedings of National Academy of Sciences this week.

KWONG: I spoke to Muhammad Adeel Rishi, a sleep specialist at Indiana University School of Medicine who wasn’t involved in this study. He found the finding compelling and said it adds to growing evidence about the consequences of daylight saving. Spring forward has been linked to an increase in heart attacks and fatal traffic accidents.

BARBER: Yeah.

DETROW: And apparently cybersecurity problems. What’s the solution here?

BARBER: So Eitan wants cybersecurity companies to be aware of this finding so they can beef up their security for any moments where there’s widespread sleep loss. So think about holidays or major sporting events.

EITAN SCHECHTMAN: There might be more of these – more susceptibility to click these links or generally to make impulsive decisions, whether it’s on the road, in the stock market or cybersecurity-related decision.

DETROW: Some places have just, like, gotten rid of time changes, right?

BARBER: Yeah.

DETROW: Like most of Arizona and Hawaii.

BARBER: Yeah.

KWONG: Yes. This is a whole other debate. But just to put science’s finger on the scale, multiple studies have shown that standard time is healthier for our bodies. Standard time has more natural light in the morning, which better aligns to our bodies’ circadian rhythms.

DETROW: All right. Good thing we’re in standard time right now.

KWONG: (Laughter).

BARBER: That’s right.

DETROW: That is Emily Kwong and Regina Barber with NPR’s science podcast Short Wave, where you can listen for new discoveries and the science behind the headlines. Thanks to both of you.

KWONG: Thank you.

BARBER: Thank you, Scott.

(SOUNDBITE OF PHARRELL WILLIAMS SONG, “NUMBER ONE (FEAT. KANYE WEST)”)

Copyright © 2026 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

Accuracy and availability of transcripts for NPR programming may vary. Transcript text may also be revised to correct errors or match updates to audio. Audio may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio record.

Please note that NPR platforms and products also include some programming produced by third parties; NPR cannot guarantee transcripts for that content and some transcripts may be produced using automated or other tools.