Subscribe
Share
Share
Embed
Truth Seekers: Where Data Meets Reality
Tired of sensational headlines and conflicting health advice? Join Alex Barrett and Bill Morrison as they cut through the noise to uncover what scientific research actually says about the claims flooding your social media feed.
Each week, Alex and Bill tackle a different health, nutrition, or wellness claim that everyone's talking about. From "blue light ruins your sleep" to "seed oils are toxic," they dig into the actual studies, examine the methodologies, and translate the data into plain English.
No agenda. No sponsors to please. No credentials to fake. Just two people committed to finding out what's really true by going straight to the source—the research itself.
Perfect for anyone who's skeptical of influencer health advice but doesn't have time to read every scientific study themselves. New episodes drop regularly, delivering clarity in a world full of clickbait.
Question everything. Verify with data. Find the truth.
Disclaimer: Truth Seekers provides educational content based on published research. Nothing in this podcast should be considered medical, financial, or professional advice. Always consult qualified professionals for decisions affecting your health and wellbeing.
**What's Your Sleep Type? The Five Categories That Don't Actually Exist**
Alex: So I've got to admit, when I first saw this headline about scientists identifying five sleep types linked to your mental health, I actually tried to figure out which one I was.
Bill: You did?
Alex: I did. I'm reading through these descriptions—poor sleep with anxiety, short sleep with cognitive issues, sleep medication users—and I'm thinking, "Oh, that's me, I'm definitely that one."
Bill: Which one did you pick?
Alex: I haven't a clue. That's the thing—I tried three different ones and they all sort of fit depending on the week. But then I actually looked at the study.
Bill: Right.
Alex: And the study is basically screaming "please don't do what the headlines are telling you to do."
Bill: Yeah, the coverage basically invites you to self-diagnose, which is kind of the problem. This story's been everywhere.
Alex: It's quite something. The researchers from Concordia University and the National University of Singapore published this in PLOS Biology. They looked at 770 healthy young adults, analyzed over a hundred different psychological, physical, and cognitive measures, did brain imaging, and found these five statistical patterns linking sleep to mental health and brain activity.
Bill: Mmm.
Alex: And the media coverage presents this like, "Here are five sleep types, figure out which one you are, and this'll help you understand your sleep problems." CNN even had a section asking how someone can figure out what type they are.
Bill: Which is brilliant, except for one small detail.
Alex: Which is?
Bill: You can't. The researchers themselves explicitly said that people cannot replicate this on their own. You'd need access to those 118 specialized biopsychosocial measures plus brain imaging plus sophisticated statistical analysis.
Alex: So the headline promises something the study literally cannot deliver. But let's back up, because I want to understand what they actually found before we get into what they didn't find.
Bill: Okay, so they used this technique called canonical correlation analysis to identify five profiles. Profile one was poor sleep combined with general mental health issues—depression, anxiety, daytime impairment. That accounted for... hang on, let me get this right... 88% of the statistical signal.
Alex: 88%?
Bill: Yeah.
Alex: That's nearly all of it.
Bill: Right. Then profile two was people who had mental health symptoms but didn't report sleep complaints—they called this "sleep resilience." That was 4% of the variance. Profile three was sleep medication use with positive social relationships—3%. Profile four was short sleep duration with cognitive impairment—2%. And profile five was sleep disturbances with anxiety and substance issues—1%.
Alex: Hang on, so the first profile is 88%, and the other four combined are just 12%?
Bill: Exactly.
Alex: Huh.
Bill: And here's where it gets interesting. The researchers did something called cross-validation, which is basically a test to see if these patterns hold up when you look at new data. In their own control analyses, profiles three, four, and five—that's three out of the five sleep types—did not survive cross-validation.
Alex: Wait, say that again?
Bill: Three of the five sleep types that are all over the headlines didn't even replicate when the researchers tested them. The researchers' own words: these profiles "might not be as robust and generalizable."
Alex: So three of the five sleep types didn't even replicate when the researchers tested them?
Bill: That's exactly what happened. Only profiles one and two showed they could generalize to new data. And even profile two barely squeaked by with a correlation of 0.19.
Alex: But the coverage treats all five equally. I saw NBC presenting all five categories as these distinct patterns that affect your brain activity in different ways.
Bill: Which is technically true—they did find different brain connectivity patterns associated with each profile in this specific sample. But finding something in one sample of 770 people doesn't mean it's a generalizable phenomenon you can use to categorize yourself or anyone else.
Alex: Especially when your own quality checks show three of them fall apart.
Bill: Right.
Alex: What does it actually mean when something doesn't survive cross-validation? Because I'm guessing most people reading the headlines don't know what that is.
Bill: So cross-validation is where you split your data into different subsets and see if the pattern you found in one subset shows up in the others. It's one of the basic ways we check if we've found something real or if we've just found noise that's specific to our particular sample. When I was doing A/B testing, this was like the first thing you'd check—does this hold up when you look at different slices of your data?
Alex: And if it doesn't show up in the other subsets?
Bill: Then you probably found noise. Or at minimum, you found something that's not robust enough to generalize beyond your specific sample.
Alex: Okay, but here's what I'm struggling with. You said profile one is 88% of the signal and it did survive cross-validation, right?
Bill: Yeah.
Alex: So isn't that the real finding? Poor sleep and mental health issues go together. Don't the other four profiles just... not matter?
Bill: I mean, that's basically what the data's saying.
Alex: But you said earlier it was interesting that they found five profiles. I'm not seeing what's interesting about profiles that don't replicate and explain barely any variance.
Bill: Okay, that's fair. I think I was giving them credit for the methodology more than the actual findings. The brain connectivity stuff is scientifically interesting even if—
Alex: Is it though? If three of the profiles don't replicate, why would we trust the brain connectivity patterns associated with them?
Bill: That's... actually a really good point.
Alex: Because it feels like we're back to profile one—poor sleep and mental health are connected—which we already knew.
Bill: Yeah. Yeah, you're right. The study's actual contribution is more that they used rigorous methods to confirm something well-documented and found that the other patterns they looked for didn't hold up. Which is less interesting than I made it sound.
Alex: Okay.
Bill: Though I will say, 770 participants with brain imaging is a substantial effort. The methodology was rigorous—they used permutation testing, bootstrap confidence intervals, proper statistical controls.
Alex: They did the science right.
Bill: They did the science right, and they were honest about the limitations. The problem is entirely in how it got translated for the public.
Alex: This feels familiar. Didn't we—hang on, didn't we just do another sleep study like this?
Bill: The AI prediction thing?
Alex: Yes! The Stanford model that was supposed to predict diseases from your sleep. Same problem, wasn't it? Headlines saying you can use sleep patterns to diagnose yourself when the study absolutely did not support that.
Bill: And same self-reported sleep data. The Pittsburgh Sleep Quality Index.
Alex: That's the one they used here too, isn't it?
Bill: Yeah, exactly. Which is self-reported and only covers the previous month.
Alex: One month. That's a snapshot.
Bill: Right. And one of the sleep experts who wasn't involved in the study—J. Todd Arnedt from University of Michigan—told NBC that if you looked at these same people a year later, they might fall into completely different categories.
Alex: So not only can you not identify which category you're in, but even if you could, it might be different next month.
Bill: Or next week, depending on your stress levels or what's going on in your life.
Alex: Which brings us to that sample. Because this wasn't just any 770 people.
Bill: Right. These were healthy young adults, ages 22 to 36, average age about 29. 78% white. About 71% employed full-time. And critically—no one with diagnosed psychiatric conditions.
Alex: So this is a study of healthy people that's being marketed as applicable to anyone with sleep problems.
Bill: Yeah, and when you're trying to understand sleep problems, studying only healthy people is... limiting. You're missing older adults, you're missing people with actual sleep disorders, you're missing psychiatric patients.
Alex: I'm thinking about someone reading this coverage who has actual insomnia or sleep apnea, trying to figure out which category they fit into, when they weren't even represented in the study population.
Bill: Right.
Alex: This is what I saw all the time in journalism. You've got researchers being appropriately cautious, and then the headline is "Figure Out Your Sleep Type!" because that's what gets clicks. The science gets done right, and then the translation just... mangles it.
Bill: And the researchers even anticipated this problem. They explicitly warned that their findings "would benefit from including a more diverse sample" and that sleep and biopsychosocial outcomes "were mostly self-reported through questionnaires" which can be unreliable. They put all the caveats right there in the paper.
Alex: And then the coverage just ignores them.
Bill: Well, some coverage acknowledged limitations. CNN actually quoted the researchers saying people can't self-identify these categories. But that's buried in the article after the headline has already promised you can learn about your sleep type.
Alex: The retraction never gets the same reach as the original claim. Or in this case, the caveat never gets the same emphasis as the hook.
Bill: Yeah.
Alex: So what should people actually take from this study? Because I don't think it's worthless. There's got to be something real here.
Bill: There is. Profile one—the poor sleep combined with mental health issues—that's well-documented, it survived cross-validation, and it's consistent with a lot of previous research. The connection between sleep problems and mental health is real.
Alex: Right, but we already knew that.
Bill: We did. I think what the study actually shows is that sleep isn't just one thing—sleep quality, sleep duration, sleep disturbances, medication use are distinct dimensions that don't always go together. Someone can have short sleep duration but not feel impaired. Someone can have what looks like decent sleep on paper but still have mental health struggles.
Alex: So the insight is that sleep is multidimensional, not that there are five neat categories everyone fits into.
Bill: Exactly. But that's a much less satisfying headline than "Find Your Sleep Type."
Alex: "Sleep Is Complicated and We Need More Research" doesn't exactly go viral.
Bill: No.
Alex: Here's what worries me, though. People are actually trying to identify their sleep type based on these articles. And if you self-diagnose as one of these categories, you might think you understand your sleep problem when you actually need to see a doctor.
Bill: Or you might have undiagnosed sleep apnea and think you fit the "sleep disturbance" profile, missing the actual medical condition.
Alex: Or you accept that your sleep issues are just your "type" rather than something that could potentially be treated.
Bill: And it normalizes this idea that complex health patterns can be reduced to five boxes you can self-select, which isn't how any of this works.
Alex: Because here's the other thing—it found correlations, right? Statistical associations between sleep patterns and mental health outcomes. But correlations don't tell you causation.
Bill: Right. Does poor sleep cause anxiety, or does anxiety cause poor sleep?
Alex: Exactly.
Bill: Or are both caused by something else entirely—stress, genetics, socioeconomic factors? The study can't answer that, and the researchers themselves explicitly say their findings "do not inform on the directionality or causality of these effects."
Alex: But the media framing definitely leans toward "your sleep type affects your mental health," which implies a direction that isn't supported.
Bill: Yeah. And that's frustrating because I actually think the researchers did solid work here, they were honest about what they found and what they didn't find, and then the whole thing gets turned into this self-diagnosis tool that they explicitly said it's not.
Alex: So if someone's listening to this and they did try to figure out their sleep type, what should they actually do?
Bill: If you're having sleep problems, talk to a doctor. Not because you need to figure out which of five categories you fit into, but because sleep issues can have lots of different causes that need different approaches.
Alex: And if you're reading health coverage that promises you can identify your type or category based on a quiz or a description, check if the actual study supports that.
Bill: Look for the limitations section. That's where researchers are usually most honest about what their work can't tell you.
Alex: In this case, the researchers literally said you cannot self-identify these profiles, that three of them might not be generalizable, that they studied only healthy young adults, and that the findings don't show causation. All of that's in the paper.
Bill: But almost none of it made it into the headlines.
Alex: Right. The study's real insight—that sleep is multidimensional and personalized approaches might help—that's valuable. The idea that you can sort yourself into one of five stable categories is not supported by the evidence.
Bill: And knowing the difference between those two things is what actually helps you make better decisions about your health.
Alex: Not to mention making you much harder to mislead next time someone promises a simple answer to a complicated question.
Bill: Which, judging by how often we see this, is going to be pretty soon.
Alex: Yeah. Probably another sleep study.