Sleep Deprivation Does Not Treat Depression

in #science6 years ago (edited)

Recent claims that sleep deprivation is an effective treatment for depression demonstrate a common problem with press releases in science journalism.


Noon – Rest from Work (after Millet) by Vincent van Gogh, 1890. Location: Musée d'Orsay, Paris.

Misleading press can hurt science

Press releases are a common source of scientific information for news publications, and it's essential that they relay information accurately. Unfortunately this isn't always the case. To reach the widest audience possible, institutions that distribute press releases are incentivized to make new science sensational, which can harm accuracy and integrity in science news.

In 2009, an analysis of press releases published by academic medical centers showed that 58% failed to mention relevant caveats that would affect a study's interpretation, and 29% exaggerated the importance of a study's findings.

For example, Penn Medicine News of the University of Pennsylvania (or Penn, where I am a student) recently published a press release titled "Sleep Deprivation is an Effective Anti-depressant for Nearly Half of Depressed Patients." While this makes for a sexy headline, it is an absurd and dangerous claim to publicize. After reading the original paper that the release is based on, it's not hard to see why.

Sleep Deprivation and Depression

Penn's press release simply states that "sleep deprivation is an effective anti-depressant for nearly half of depressed patients" and suggests that the findings of this study can offer relief for the 16 million adults who experience an episode of depression in a given year. I wish that were true, because it be an incredible benefit to many people. Yet like many press releases, this one both exaggerates its study's findings and fails to mention caveats that would dramatically affect its interpretation.

The heart of the issue here is the use of the phrase "effective anti-depressant." It's worth noting that since anti-depressants (or just "antidepressants") commonly refer to drugs that treat depression in humans, it is a bit unconventional to refer to sleep deprivation as an antidepressant. Current antidepressant drugs are by no means perfect (they may have side effects, take several weeks to work, and often people do not respond to them), but they are shown to be an effective long term treatment for depression in around 50% of people. Sleep deprivation, however, has not been shown to be so useful.


Seroxat (paroxetine), an antidepressant drug (photo by Jussi Mononen)

Problem: there are few control groups

It is a hallmark of scientific experiments to test a control group against an experimental group. This protects results from researcher bias and false conclusions. In the case of sleep deprivation, the experimental group would consist of people diagnosed with depression who experienced sleep deprivation, while a control group might consist of people diagnosed with depression who were not sleep deprived.

However, 91% of the data sets in the paper came from studies that "did not use control groups." This makes it much harder, if not impossible, to establish a causal link between sleep deprivation and effective depression treatment. This huge limitation was not mentioned by Penn Medicine News.

Problem: most people relapse from treatment

If someone sold me an fridge that spoiled my food after a few days, I'd be pretty mad about it. If something is branded as an "effective anti-depressant" it had better be work long enough to help people. But that's not what sleep deprivation does:

"The disadvantage of SD [sleep deprivation] is that the effect is not stable. In more than 80% of responders to SD, a relapse into depression occurred after the recovery night." (Henmeter et al., 2010)

Penn's press release failed to recognize this crucial point. However, the authors of the study (who used data most often collected within a day of patients' sleep deprivation) took note:

"The availability of an antidepressant treatment that has rapid effects in 50% of patients would mark a radical improvement in clinical practice, if we can find ways to maintain the effects over time." (Boland et al., 2017)

That "if" here is crucial. While sleep deprivation is remarkable for its temporary antidepressive effects, it offers no long term solutions to treat depression. It is an absolute misrepresentation of the study to say that it is an effective antidepressant.

Irresponsible Press Has Ripple Effects

News spreads fast, and a story can change as its whispered down the lane. The student newspaper at the University of Pennsylvania recently regurgitated this press release, but also added that "medical professionals are recommending partial sleep deprivation" to people seeking treatment for depression. If your doctor suggests this, find a new one. By spreading misrepresented information, a bad press release can end up leading to bad medicine and public misinformation.

Screen Shot 2017-10-04 at 3.32.34 AM.png

Screenshot of a Google News search of the press release's headline

In many ways, it is not fair to pick on this particular press release, because this kind of thing happens all the time. But since I'm a student at the university it came from, a university where sleep deprivation and depression are both serious issues, it struck a nerve. It's a pet peeve to see information misrepresented, but it feels much worse to see a "depression cure" sensationalized.

America's universities are seriously affected by both depression and sleep deprivation, as shown by a 2014 survey of over 79,000 college students. At least once in the the previous 12 months, 32.6% of students had felt too depressed to function, and 21% of students reported that sleep difficulties had adversely impacted their academic performance. 41.3% of students said that sleepiness was "more than a little problem" during daytime activities. At an upsetting extreme of the college mental health crisis, 14 Penn students have committed suicide since February 2013.

For too many people, college is a mentally challenging place with overwhelmingly large workloads and hopelessly uncertain futures. With all this constantly in mind, it is hard to tolerate an institution whose environment fosters these health problems spreading misleading information about their treatment. Whether you are a member of a community that's impacted by depression, a press release writer, or a scientist, it's important to represent information accurately, or else we risk losing sight of the reason this kind research is done in the first place: to help people.


American College Health Association. (2014). American College Health Association-National College Health Assessment II: Reference Group Executive Summary Spring 2014. Hanover, MD: American College Health Association. Retrieved from
Antidepressant. (2017, October 3). In Wikipedia. Retrieved from
Boland, E. M., Rao, H., Dinges, D. F., Smith, R. V., Goel, N., Detre, J. A., … Gehrman, P. R. (2017). Meta-Analysis of the Antidepressant Effects of Acute Sleep Deprivation. The Journal of Clinical Psychiatry.
Hemmeter, U.-M., Hemmeter-Spernal, J., & Krieg, J.-C. (2010). Sleep deprivation in depression. Expert Review of Neurotherapeutics, 10(7), 1101–1115.
McFadden, J. (2017, September 27). Penn researchers have found that sleep deprivation can improve depression. Retrieved October 4, 2017, from
Sleep Deprivation is an Effective Anti-depressant for Nearly Half of Depressed Patients. (2017, September 19). Retrieved October 4, 2017, from
Spinelli, D. (2017, September 2). Sigma Alpha Mu acknowledges death of College senior Nicholas Moya in Facebook post. Retrieved October 4, 2017, from
Thomas Insel. (2011, December 6). Antidepressants: A complicated picture. Retrieved October 4, 2017, from
Woloshin, S., Schwartz, L. M., Casella, S. L., Kennedy, A. T., & Larson, R. J. (2009). Press Releases by Academic Medical Centers: Not So Academic? Annals of Internal Medicine, 150(9), 613.

Until next time

This post is under the CC BY 4.0 License. (?)


Nice post. A side note: How do you feel about crediting the institutional owner of the Van Gogh? I think it is Musée d'Orsay, Paris, France.

Not so fast @benfink! Technically I didn't have to credit this work at all, not even to Van Gogh himself. While the physical copy resides in Musée d'Orsay (correctly identified by @kenfinkel), the work itself is public domain in both its country of origin and the United States. From the image's online source:

This work is in the public domain in its country of origin and other countries and areas where the copyright term is the author's life plus 100 years or less.

This work is in the public domain in the United States because it was published (or registered with the U.S. Copyright Office) before January 1, 1923.

Agreed, but setting the bar low and leaving off relevant and useful information - while not a crediting crime - is a flaw. In such cases, there's a hierarchy of information. The artist and location of the original are essential, in my view. The date, medium and size of the original are useful, though less important. I know it's a burden for the writer, but there are at least two versions of this painting in different museums. It may be only an illustration in this case, but it's also an authentic artifact that should be accompanied by at least some of it's critical metadata.

Thanks for the feedback @kenfinkel. I am unaccustomed to citation conventions for works of art, and I have edited the post to reflect the this piece's location.

Ah, much better... Thank you.

im glad my students can't wiggle their way out of citations as well as you @somethingburger "Technically" my butt :P

Researchers have focused on sustaining the often ephemeral effects of bee stings, either with multiple administrations of whole-hive bee attacks or just a few bee stings for example; however, more work needs to be done to bring the field to more of a consensus as to how best to apply insect inoculations. As Hemmeter et al note in their review of studies of repeated administrations of buzzing stingers, how a depressed individual responds clinically to a single buzz butt is not often predictive of how that individual will respond to subsequent sharp butts. Indeed, temporal trends either have not been observed or have been contradictory, demonstrating both increased as well as decreased response to subsequent bee attacks.

OK that was kind of a silly re-write, by my point is that when the authors of the original study have few ideas or evidence about why sleep deprivation could actually work as a treatment, the mechanism. If absence of understanding this, and in the presence of only short-term effects, why not see the effects of exercise? Or dunking your head in ice water? Or getting stung by some goddamn bees.

The potential that it "resets of the body’s circadian rhythms" seems really interesting, but also with potential dangerous effects of its own? I dunno, seems like for something to be really valid for future study, long term or at least medium term effects should be compulsory.

I have no real issues with the original study, or its authors! This was responsible research: it combined existing data without running new experiments on people, and it fully addressed its limitations. My issue is with the press release and its loose use of the phrase "effective anti-depressant" and neglect to mention a few caveats, which would only take a few sentences to include.

While the article was useful research (as would be your hypothetical on bee stings), its press release was misleading (as it would be to call bee stings "effective anti-depressants" based on your paragraph).

Helpful reply! But two things make me uncomfortable with this kind of meta-study in addition to the "no mechanism" part.

  1. Short term effects, why do they matter? Is physiologically or psychologically masking depression an interesting or useful thing? Maybe the former, probably not the later. That's partially what my bee sting thing was about. I'm sure there are many ways to mask depression, and the weaker your a priori mechanistic justification needs to be, the longer or more ridiculous the list will be.
  2. Meta-studies like this one might really suffer from publication bias. They aren't sampling a random pool of studies. Rather, because positive results are more likely to be published, they are more likely to be included in this study. Therefore, such a meta-study would overestimate any effect, even more so than most individual papers. I notice only a few of the studies they sampled had effect sizes that overlapped with zero (Fig. 2 below). Including only randomized samples could help with this!

Just some thoughts! Criticizing is way easier than building up and the authors did impressive work.

Great, great points. There are some more nuanced things that I didn't mention as well: many of the studies used a modified version of the Hamilton Depression Rating Scale that excluded symptoms concerning sleep. This affects results because it changes the study's definition of depression, excluding symptoms that might be a cornerstone. Kind of a problem, but not one there's an easy solve for.

So basic academia 101.
Data massaging is a common practice.

Exercise is the best anti-depressant. None other needed.

that doesn't seem fair @granunic0rn

Exercise can be amazing for depression. It can also be terrific at lowering cancer risk. But, if you get cancer, nobody will tell you "Just keep exercising." Just because depression is a mental disorder, doesn't mean it doesn't have a real physiological basis. Some people need drugs or special treatments to help their brain chemistry get back to healthy levels.

Cancer is the result of a fixed set of physiological conditions that arise primarily from inadequate diet that cause excess glycation of telomerase, and obesity which causes excess cell multiplication (more fat = more tissue = more cell multiplication = more risk of cancer) so cancer I'd address primarily with diet. Of course certain activities, such as drinking and smoking also cause cancer because they bring toxins into your system (ethanol = destroys mitochondria, smoking = a bunch of shit). Diet can also affect mood, but that's somewhat besides the point.

Depression can also be addressed with cognitive behavioral therapy in severe cases so I agree, exercise is not the only possible solution, but if you exercise regularly the likelihood that you will develop depression is much lower.

The SSRI's etc. are super useless in my experience. They don't work for many patients and need to be cycled if they work, so in short not a solution.

Exercise on the other hand is an actual solution because it has several beneficial effects on multiple brain systems that are involved in the development of depression (e.g.: the hypothalamus, which does not atrophy when you exercise regularly, in fact, it appears as if exercise facilitates neurogenisis in the hypothalamus).

Hey @granunic0rn, please cite your sources! I'm glad that you're discussing some points on the molecular and systems level, but these are some serious claims made without any background cited.

Like, access google scholar, type in: effects of glycation on genome expression. Read abstracts of results, form opinion or find the next best review that discusses the issue at length to form an adequate opinion on the consensus of the field.

In Biology they all pretty much agree that glycation and adiposity is bad for health because it destroys the genome. They have numerous reasons for that believe (it's a theory, but it has plenty of evidence: I refer to the DNA = broken you'll get cancer theory).

I personally read a summary by Amen (2016), but that one is privately published so IDK if you want to give me a 10 minute mail so that I can send the PDF to you.

Alternatively, if you really really really want to know I'll provide you with a list of references that Amen used in his paper.

It's mostly rodent, mammal and bird studies though. Obviously cannot lock humans into cages and feed them diets that are suspected to cause cancer or vice verse study effect of diets that prevent cancer (human lifespan is a serious limitation for those, rodents are much better because you can have a few generations in a decade and publish results). However, the AMEN paper mostly addresses diet and the physiological mechanisms that destroy the body when you are on a 'bad' diet.

Nonetheless, you can extrapolate some of the research findings to bio-physiological processes and then speculate that they would also affect the psyche in a detrimental way. Speculate because there is currently no consensus in neuroscience.

There is consensus on certain things, like facilitation of neurogenisis in the hypothalamus of rats that are exercised with some evidence supporting this claim in humans (e.g.: elderly patients in a nursing home after 12 week light exercise program), but of course you cannot claim that this can HEAL a disorder.

It can merely reverse some of the physiological deterioration that occurs in old age and with the (somewhat) limited restoration of the system there is a possibility of the reversal of symptoms (more or less).

I recently wrote a paper on the topic of exercise and self-regulation. Exercise seems to improve self-regulation because it improves blood flow which increases blood flow to the PFC which is the decision making center. The PFC is furthest away from the heart and thus more likely 'just because of the distance' to be cut off from blood flow compared to say the brain stem which is actually connected to the two main arteries going into the brain which basically means its never starved for blood. On the other hand the PFC is only connected to one artery and thus could be more easily starved from blood if the heart muscle is not able to move sufficient volume of blood to the brain. Exercise makes this less likely because exercise grows the heart muscle and improves blood circulation (sheer volume increases per pump, larger heart muscle pumps more blood per beat, larger heart muscle can pump at higher frequency if necessary, etc.)

You can start by reading the paper from Reynard, Gervitz, Berlow et. al (2011) in Psychophysiological Biofeedback. The article is called Heart Rate Variability as a Marker of Self-Regulation.

The article itself summarizes a lot of the finding, BUT it obviously links to a long list of articles who explain the biological, physiological and psychological background at length. Please don't ask me to link all the articles I have on that topic individually and research it for yourself a little. The articles cited in this article are a good start, but there are much better biological papers I have to admit.

My offer for the Amen article and its content list still stand. I'll have to find the email with his article but I know I still have it. However, it's like 2 AM here already and I think I won't look that up tonight. So get back to me tomorrow and I'll see what I can find for you.

If you are interested in diet effect mood there are really interesting studies with parasites, like the mildew mushroom. It can produce human hormones like estrogen or melatonin. So the parasite can induce sleep or sleep disorders. It also likes to make hormones that increase hunger so you keep feeding it before it moves to the next target.

Of course, in this case this would be an effect of diet in conjunction with a parasite, but diet in general can have effect on mood too.

We can argue about the magnitude of the effect on the human psyche, nonetheless the positive or negative effects of diet on the physiological function of the human body are undeniable.

The better the foundation, the better your likelihood to stay healthy (physically and mentally). You are what you eat is a maxim that holds true, however, I won't claim that you'll get depression just because you eat a certain type of food. No one can proof that beyond a doubt. We can just show that consumption of certain macro (or micro-nutrients) correlates with certain symptoms of certain disorders.

Good luck actually finding the causal relations. That's a different beast to slay.

Coin Marketplace

STEEM 0.17
TRX 0.08
JST 0.023
BTC 26576.78
ETH 1594.56
USDT 1.00
SBD 2.18