Sleep Deprivation Alters Brain Connectivity

By Neuroskeptic | January 1, 2016 5:47 am

Missing out on a night’s sleep causes “robust alterations” in the functional connectivity of the brain, according to a new paper just out in Neuroimage.

Neuroscientists Tobias Kaufmann et al. of the University of Oslo acquired resting state fMRI scans from 60 male volunteers. The participants were scanned three times each – once in the morning, again the same evening, and then finally the next morning. But only some of the volunteers were allowed to sleep in the night before the final scan. The others were kept awake, experiencing sleep deprivation.

The results showed that sleep-deprivation altered connectivity, compared to the well-rested condition. The evening condition (i.e. at the end of day spent awake) was intermediate between the two, suggesting that wakefulness causes progressive changes that build up over time.


Sleep deprivation was associated with altered connectivity across diverse brain regions. Generally, positive connectivity between brain regions was increased by lack of sleep. Of the 17 connectivity “edges” that were significantly affected by sleep status, 13 became more positive, while only 4 became more negative.

This result is a reminder that positive connectivity isn’t always “good”. Intuitively we might expect brain connectivity to be a sign of “information processing” and “integration”; but Kaufmann et al.’s data suggest that the brain gets more connected when it’s exhausted.

On the other hand, are there more prosaic interpretations of these data? When I first saw this paper, I thought: we know being asleep alters brain connectivity, so what if the sleep deprived participants fell asleep during their final scans? But the authors were alert to this problem, and they used an fMRI sleep classification analysis to detect periods of sleep, based on brain activity. The results were not changed after excluding these periods.

Likewise, the sleep deprived volunteers had more head motion, another potential fMRI confound, but the results held even after controlling for this.

But I spotted another potential problem. Kaufmann et al. asked their volunteers to keep their eyes open during the resting state scan. This is common practice, but we all know that it can be difficult to keep your eyes open when you’re sleep deprived. Closing your eyes has pronounced effects on brain activity as measured with EEG. Maybe it also affects fMRI? Kaufmann et al. had no way of knowing whether people’s eyes were open or closed, and couldn’t control for it.

Still, the authors note that wakefulness (and the time of day when scans take place) is a potential confound in fMRI studies. If some of the participants are scanned in the morning and others have their scans in the evening, the latter will be more tired, and their brain activity may differ:

By suggesting an important role of time-of-day on brain connectivity, [our data] have potentially important implications for a range of applications, including a potential bias in scientific studies when group assignments are not balanced for time-of-day.

ResearchBlogging.orgKaufmann T, Elvsåshagen T, Alnæs D, Zak N, Pedersen PØ, Norbom LB, Quraishi SH, Tagliazucchi E, Laufs H, Bjørnerud A, Malt UF, Andreassen OA, Roussos E, Duff EP, Smith SM, Groote IR, & Westlye LT (2015). The brain functional connectome is robustly altered by lack of sleep. NeuroImage PMID: 26712339

  • Pingback: Sleep Deprivation in Teens Largely Due to 'All-Night Te…()

  • Bernard Carroll

    It would be helpful to see results of follow-up scans after ‘recovery’ from the night of sleep deprivation. Should these now seem no different from the baseline scans – which is what we could expect – then the inference would be that connectivity differences are reversible. In turn, that would be informative for studies in disease states.

    • Neuroskeptic

      Good point. The changes probably are reversible. The behavioral effects of sleep deprivation certainly are.

      • Bernard Carroll

        Actually, sleep deprivation is a well known experimental antidepressant procedure. I have witnessed it cause a switch from bipolar depression to hypomania overnight. Pity that it doesn’t seem to last, though.

        • reed1v

          Just observe medical interns pulling their 24-48 hour shifts. Every hospital is a mini lab.

    • youni

      Scrupulous studies of human psychomotor performance suggest that on average humans can recover the equivalent of one hour “sleep loss” by sleeping a full 8-9 hours on each subsequent day. The effect of insufficient sleep on performance is cumulative. In principle – sleeping 6 hours (instead of 7 hours) on 5 consecutive workdays leaves the individual with a deficit of 5 hours. But sleeping longer (>8 hours) on two weekend days will only recover “two” hours. The neurological implications of chronic sleep deficit is not well understood. In animal studies, researchers have just recently started to uncover evidence of brain cell injury/loss in locus coeruleus and hippocampus that is not reversed. Another finding is that clearance of brain toxins (e.g. beta-amyloid) is heavily dependent on “glymphatic” flow; this clearance doubles during sleep. If replicated in humans, the implications would seem potentially profound?

      • Craig_Knaak

        who can sleep 9 hours?

  • Pingback: Sleep Deprivation Alters Brain Connectivity()

  • Pingback: Sleep Deprivation Alters Brain Connectivity()

  • Jim Costich

    In the 90’s I was traumatized in the middle of the night. After that I couldn’t sleep. I’d wake up at midnight (time of trauma) with my heart pounding, gasping for air and dumping a full adrenal load of pure panic. After a few weeks of this I was a mess. My doctor prescribed sleeping pills and slowly I recovered. It took about 6 months before I could sleep normally again and although drugged sleep was better than no sleep I was far from normal until I could finally sleep without it and no panic attack. It is no surprise to me that war “survivors” go years, decades or even never recover. Connectivity breaks may be reversible sometimes but I think we already know through experience that in many instances they’re just not.

    • Steven Astravas

      The brain is a strange thing. Five years ago, I was traumatized in the middle of the night from a bad dream that triggered the beginning of a “Lucid dream” where I was fully conscious inside the dream, walking around a neighborhood for what seemed about 10 minutes and where I saw an event that came true 6 months later. I saw another event that only now I’m beginning to see that it may come true. I never had a dream like that again. The brain can do miraculous things, so all good life is sacred in my opinion.

    • reed1v

      I get that way thinking about Hillary as President.

  • practiCalfMRI

    “Likewise, the sleep deprived volunteers had more head motion, another potential fMRI confound, but the results held even after controlling for this.”

    There will be residual motion in the data and if there is a group-wise difference in residual motion then it may drive the effects. The motion profiles should be rendered similar across the groups, e.g. by more rigid head restraint. Post hoc fixes don’t work, and based on some sims we did a few years ago, even perfect (yes, perfect) post hoc motion “correction” leaves spurious correlations when using an array receive coil. (This study used an 8-ch coil.) Fore more info:

    As a field we must move towards motion elimination. Let 2016 be the year we recognize that post hoc motion “correction” isn’t sufficient!

    • Aziz Poonawalla

      Obviously your focus is on fMRI so the statement about post-hoc correction not being sufficient has an implied qualifier. But I do think it important to acknowledge that post-hoc correction for other types of scans, including diffusion, is perfectly valid.

      • practiCalfMRI

        The best approaches for diffusion, which is motion sensitive by intent, are to build robustness into the acquisition, e.g. navigator echoes, re-acquisition of data, etc. So I would wager that a post hoc correction if diffusion data isn’t sufficient either.

    • reed1v

      Well, you could chop off their heads and place them on a common grid so identical environments are facilitated. Then altering sleep patterns will provide better controlled responses.

      • Bill C

        this dovetails nicely with a suggestion I made on another blog in the past few days that (previously determined) executions should be carried out inside of an fMRI machine, to study the effects of trauma and answer interesting questions like how long does the brain live once the head is cut off, etc.

        • reed1v

          Based on some medical evidence along those lines, the brain continues to fire away its synapses for about 10 minutes after the heart stops or blood flows to the brain stop. The optic nerve continues to send, albeit at a diminishing level, impulses to the brain for about half this time.
          During WW2, German “doctors” would kill healthy subjects after asking them to consult with God and report back somehow what was revealed.
          Evidently it turned into a long wait for those reports.

  • Jeff Olney

    More positive connections for sleep deprivation? Could the brain be doing this due to lack of synapse repair in connections already open? Could the brain be consuming more calories to do the same amount of work? I wonder if that’s what we’re seeing here.

    • Ivan Dimkovic

      If I recall correctly, there was an article describing how an serendipitous observation of increased likelihood of epileptic seizures after sleep deprivation led to an experiment confirming slightly increased excitability as an response to TMS.

      One hypothesis would be that brain performs homeostatic “synaptic scaling” during the sleep, maintaining memory of the information collected the previous day but preventing runaway excitation. If there is a biological basis for this hypothesis, it could explain why sleep deprivation increases risk of epileptic activity in vulnerable patients and why it also interferes with the acquisition of new memories.

  • Mo Bakht

    I expected they have reported that Sleep Deprevation leads to decresing brain connectome while they have reported that it increases brain connectome! So, for brain disorders which are as a result of decresing brain conectom Sleep Deprevation Therapy is recomended!

    • Ivan Dimkovic

      “Connectome” is just an artificial measure based on an abstract model, it does not necessarily correlate positively with an ability to form memories, perform executive function, etc.

      In fact, it could also be exactly anti-correlated in some cases, where increased connectivity leads to adverse effects. For example, there are theories confounding increased connectivity with autism spectrum disorders.

      Thankfully, in this particular case, if there is an underlying “increase in connectivity”, it seems to be perfectly reversible by… (sound of drums)… sleep.

  • nik

    The conclusion?
    Lack of sleep leaves one tired.

  • Pingback: Morsels For The Mind – 08/01/2016 › Six Incredible Things Before Breakfast()

  • Sandy Gibbs

    All male volunteers? I guess women don’t need to know how sleep deprivation affects them.

  • Ray Franklin

    The graphic shows diagrams of a sort I have not encountered. They are clearly something unique to neuroscience. Got a link to an explanatory article?

  • Pingback: Lectuur op zaterdag: slaap, Belle en het beest, angst en mythes bij de politie (en meer) | X, Y of Einstein?()



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar