AI could help hospitals diagnose PTSD in real time
Hosted by Emily Kwong and Regina Barber
- 12m
- Transcript
Summary
Written by a model from the transcript; not yet reviewed.
This news roundup explores three recent science stories spanning medicine, astronomy, and cybersecurity. First, the hosts examine a new AI model that analyzes emergency room patients' answers to a routine screening question to predict their risk of developing PTSD a month later. Second, they look at an image-processing technique developed using physics and algorithms rather than AI, which subtracts starlight from James Webb Space Telescope images to make hidden exoplanets visible. Finally, the discussion turns to a study investigating whether losing an hour of sleep during daylight saving time increases vulnerability to phishing attacks.
- Emergency room screening answers can be fed into an AI model to predict PTSD risk with 95% accuracy. 6:34
- Researchers can uncover faint exoplanets in telescope images by subtracting interfering starlight using physics and algorithms. 9:43
- Employees who received a phishing email the Monday after daylight saving time were 13% more likely to click the malicious link. 12:10
Chapters
- 0:00Intro and Sponsor Messages
Hosts introduce the episode and play NPR sponsor messages.
- 1:21PTSD Risk Prediction Story Intro
Hosts explain how AI can flag trauma patients at risk of PTSD using a single question.
- 1:57Mock Patient Examples
Hosts present two fabricated patient responses to test the AI model’s predictions.
- 2:49Exoplanet Detection Technique
Hosts transition to astronomy, explaining a new image‑processing method to find faint exoplanets.
- 6:15Model Accuracy and Methodology
Researchers reveal the AI model’s 95% accuracy and describe how it uses patient narratives.
- 8:39Privacy and Consent Concerns
Discussion of HIPAA compliance, data handling, and patient consent for AI use.
- 11:31Daylight Saving and Cybersecurity Study
Hosts cover a study linking sleep loss from daylight saving to higher phishing click rates.
Highlights
- 1:21
AI Predicts PTSD Riskinsight
Researchers developed an AI model that predicts a trauma patient's risk of developing PTSD by analyzing their answer to a single emergency room question.
- 7:33
AI Reasoning Transparencystrategy
The AI model for PTSD prediction provides its reasoning at every step, identifying factors like dissociation, loss of control, and protective factors.
- 9:08
Exoplanet Imaging Techniquestory
Scientists studied existing images of star system 51 Aradani B to test a new technique that corrects for instrumental errors and unstable telescope conditions.
- 10:10
No AI Used for Exoplanetsinsight
Researchers successfully uncovered the exoplanet using only physics knowledge and algorithms, completely without artificial intelligence.
- 12:10
Daylight Saving and Phishingdata
A study of 11,000 university employees found that those tested the Monday after daylight saving time were 13% more likely to click a fake phishing link.
Quotes
“They fed those answers into this AI model and the model predicted that patient's risk of developing PTSD a month later.”
Emily Kwong · 6:23 “We are looking not on PTSD symptoms right now, but we're looking at what happens in the future.”
Unnamed speaker · 6:56 “The study was published in the Proceedings of National Academy of Sciences this week.”
Regina Barber · 12:18 “People emailed after daylight saving were 13% more likely to fall for the phishing attack the following morning and click that bad link.”
Regina Barber · 12:10
Sponsored segments
- Vuori 0:00–0:16
- Google Workspace 4:24–4:45
- Planet Visionaries 4:42–5:12
- Angi 5:12–5:45
- Google Workspace 14:04–14:38
- Bayer 14:40–14:53
About this episode
Every year, millions of people in the United States develop PTSD from a traumatic event—like a car crash, an assault or freak accident. A lot of those people end up in the ER answering the same question: “What brought you to the emergency department today?” Now, a research team in New York has an AI model that can use patients’ answers and demographics to spit out one number that tells medical staff how likely a patient is to develop PTSD a month later. The work – published in the journal Nature Communications https://www.nature.com/articles/s41467-026-78768-3 – could help an often-overloaded emergency department team prevent mental health disorders before they start. Interested in more science in the news? Email us your question at shortwave@npr.org mailto:shortwave@npr.org. Support public media with NPR+ and enjoy perks for over 25 podcasts like this one. This show’s perks include sponsor-free listening. Learn more at plus.npr.org http://plus.npr.org. See pcm.adswizz.com https://pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences. NPR Privacy Policy https://www.npr.org/about-npr/179878450/privacy-policy
Transcript
2,656 words · 8 speakers
Speaker names marked are inferred from speaking time and the show’s known hosts — they are not voice-verified.
Topics
- AI
- Minority Report
- Nature
- James Webb Space Telescope
- Nature Communications
- Proceedings of National Academy of Sciences
- PTSD risk model
- exoplanet imaging technique
- daylight saving time
- phishing attacks
Page updated · transcript and summary are machine-generated