Does L-theanine work for me? Effect on sleep and cognition

Many supplements are said to improve sleep, but how can I know if they work for my individual case if the effects are not obvious? To answer this, I ran a self-experiment with L-theanine using this platform.

L-theanine, commonly described as an aid to relaxation, is a non-protein amino acid that can most easily be found in green tea. It is credited with a wide range of positive effects, such as:

1. Increasing relaxation under resting conditions [1]
2. Reducing “acute stress and anxiety in people experiencing stressful situations” [2]
3. Raising sleep quality in individuals with ADHD and schizophrenia [3]
4. Improving learning ability and cognition when taken with caffeine [4]
5. Boosting immune function in certain cases when taken with cysteine [5]

Of course, the issue with relying on studies is that the response to a supplement varies across individuals. I may be a high-responder or I may get no benefit at all. Which will I be for L-theanine? Let’s see what the data says.

Over the past year I used two different L-theanine brands: one which came in 500mg capsules and the other in 200mg. To give you an idea of how much L-theanine that is, 250 mL of brewed green tea yields approximately 8 to 30 mg of L-theanine; so around 2.5 litres of brewed green tea would be needed to get 200mg of L-theanine. You can see the pattern of usage in the graph below; I tracked this using the Journal feature on this website. The blank in the middle is when I stopped taking it for a few months.

Sleep data was collected first using a Fitbit Charge 5 and then a Garmin Forerunner 255. The data collected by these watches on sleep, exercise, and heart rate was then synced directly into this website using the import widget.

In the second period, I tried to vary usage day to day. The earlier period was marked by more streaks, which will be useful in order to have data to tell me if the effect wears off: something for a future analysis when I improve the analytics further.

This study was not blinded so there could be a placebo effect. This study should therefore be seen as somewhat preliminary—if there is no effect or just a tiny effect, then we have evidence to stop taking the supplement. And if there is an effect, then a self-blinded experiment could establish whether what we see is a placebo effect, or something more.

The study uses regression analysis. This allows us to control for multiple factors that might affect sleep, such as weather, air quality, daily activity such as exercise, and other supplements.

Data on weather and air quality is imported by Metasophi after identifying the approximate location of residence on the Dashboard. I lived in three different places during the time of study, so the regressions also control for that.

The regressions analyse how well sleep quality variables, such as total sleep duration, deep sleep, REM sleep, and heart rate, are affected by L-theanine and all controls. The controls for sleep quality initially tested include:

• Melatonin
• Afternoon Intermittent Fast
• GABA
• Total Calories Spent
• Ashwagandha
• Creatine
• Wetbulb Temperature

After the first regression is run, the least significant variable with a p-value greater than 5pc is then dropped and the analysis is rerun. Data for other variables such as air quality (PM2.5) is available but is not used as it would reduce the sample size of the analysis by more than twenty percent. The standard errors of the regression are then adjusted depending on whether heteroskedasticity or autocorrelation is present.

The results generally indicate that L-theanine might lengthen REM sleep duration, but any other effects are close to zero. Below is a ranking of the variables that L-theanine possibly affects, taken directly from the Results tab on the dashboard. The raw regression output for the one (borderline) significant relationship is included in the Appendix.

So what do these results mean? A typical value of 200mg of L-theanine is correlated with Rem Sleep Duration increasing by around 4 minutes. There is a 6 percent chance that there is no effect i.e. L-theanine causes no change in REM sleep duration.

The grey bars indicate that the level of uncertainty is too great to make a conclusion, implying basically no effect. For example, a typical value of 200mg of L-theanine is correlated with Deep Sleep Duration increasing by just 2 minutes, with a 25.2 percent chance that there is no effect.

A key piece of data missing here is sleep latency i.e. how long it takes one to get to sleep. My wearables simply did not report it.

If L-theanine makes me more relaxed, it should lower my heart rate [6]. But here we see no effect. It's possible that a lower heart rate is masked by a higher amount of REM sleep; during this stage of sleep the heart rate is relatively high compared to deep or light sleep. However, another indicator for psychological stress is heart rate variability (HRV); where higher HRV is indicative of lower stress [7]. The above analysis tests the impact on Heart Rate Variability: if it did reduce stress, then we should see HRV increase. But in fact we see basically no effect. It is therefore likely that L-theanine has no effect on stress or relaxation in my case.

L-theanine usage seems to be associated with more REM sleep. The literature does not seem to be very insightful on whether or why L-theanine could increase REM sleep more than other types. The one exception to this is a paper looking at how L-theanine and GABA affect sleep in mice; while L-theanine boosted REM sleep, both L-theanine and GABA taken together implied substantially greater increases in REM sleep compared to other types of sleep [8]. GABA is controlled for in the regression above, but the analysis above does not explicitly test for any synergy between them. I will retain an examination of that for when I add interaction effects to the standard analysis.

How sizeable is this effect? Over the months of May and June 2024, the average amount of nightly REM sleep I had was 82.3 minutes. A single 200mg capsule seems to increase REM sleep by 4 minutes, or by around 5 percent (after correcting for the presumed effect of L-theanine, not taken 26 of the last 60 days). Sixty 200mg capsules cost around €26, so 4 additional hours of REM sleep cost €26. And of course, there is a 6 percent chance that I effectively get 0 extra hours of REM sleep, meaning a 6 percent chance the spend yields essentially no benefit.

The question then is whether it is worth taking L-theanine, even at the optimistic range of the estimates. What benefit does a marginal amount of REM sleep actually have? In order to answer this question, I looked at the information on the effect of REM sleep on my exercise and cognitive performance:

What we see in the above graph is that REM sleep really only seems to affect one variable: the average time it takes to answer a card in Anki, an app I use to learn languages. The faster I answer, the better my cognitive performance. An extra minute of REM sleep is correlated with my answering the card 0.01 seconds faster. There is a 4.4 percent chance that there is no effect.

This result is very small but makes sense, as REM sleep is implicated in memory consolidation [9]: it should help one perform better on a memorisation activity. But this implies that a 200mg supplement makes me 0.04 seconds faster; a card takes around 4 seconds to review, so we are talking about a 1 percent improvement in processing speed. In terms of time savings, over the last year I averaged 177 reviews per day; so the supplement saves me around 7 seconds per day on this activity.

On the basis of these numbers, while I do benefit somewhat from using L-theanine, the effects are very small. That’s not to say that L-theanine can never be of any value to anyone: my findings probably don’t apply to people who are highly stressed, or drink caffeine late at night [10]. It may also have an effect on a variable this website is not designed to evaluate, like immune function. Perhaps the cognitive tests could be better designed—currently there is a lot of noise in that data.

So at some stage this analysis will be revised. In the coming weeks, I will produce similar reports on how Ashwagandha, L-Arginine, Creatine, GABA, Magnesium Bisglycinate, and Coenzyme Q10 affect exercise and/or sleep. Of course, this analysis is not perfect, and in the future I would improve it in the following ways:

Brand-level results: It would be good to provide estimates for the two brands separately to the overall analysis: some brands presumably are higher quality than others, and it would be nice to identify which. Currently, this could be done in the system but the results would not then be aggregated.

Test L-theanine on different variables: If we want to lengthen sleep time, then we would probably want the supplement to still be in the system five or six hours into sleep. But L-theanine has an elimination half-life of around 60 minutes—so after 4 to 5 hours, only around 6 percent of the dose is still left in the body. It may make more sense to test L-theanine on something we do one or two hours after taking it, like a cognitive test. An example here could be to test the effect on cognition by taking it in the morning before revising a set of flashcards.

Do a self-blinded study: In the event of a cognitive test, some kind of control for placebo would be important. This would require a self-blinded study—something I will explore in the future.

Add interaction effects: REM sleep is homeostatically regulated: if we don’t get much REM sleep one night, the body will compensate the next night. In this sense, it’s possible that the supplement is just shifting REM sleep from one period to another: two extra minutes one night might followed by two less the next night. Related to the previous point is that our body may get used to L-theanine over time: initially it will have a large effect but then that will diminish over time while we keep taking it. Then, if we stop taking it, sleep quality will fall for a while before returning to normal. If this is the case, then there may be little point in taking it. But this is a dynamic that we could also test for using an interaction effect (i.e. to see how today’s dose interacts with the cumulative amount taken over the past seven days): that will be looked at in a future analysis.

Adding the capacity to test for interaction effects would also allow us to see if two supplements have some sort of synergistic effects with each other, for example GABA and L-theanine in the study mentioned above.

Maybe when I improve the analysis, I might have evidence that would lead me to take L-theanine once again. For now, no improvement on relaxation and only a tiny increase in sleep means I'll be saving my money to buy other supplements.

If you would like to follow these developments and see the results of future studies, then be sure to subscribe to the newsletter:

[1] Lu, K., Gray, M.A., Oliver, C., Liley, D.T., Harrison, B.J., Bartholomeusz, C.F., Phan, K.L. and Nathan, P.J. (2004), The acute effects of L-theanine in comparison with alprazolam on anticipatory anxiety in humans. Hum. Psychopharmacol. Clin. Exp., 19: 457-465. https://doi.org/10.1002/hup.611

[2] Everett, Julian & Gunathilake, Dilanka & Dufficy, Lisa & Roach, Paul & Thomas, Jackson & Upton, Dominic & Naumovski, Nenad. (2015). Theanine consumption, stress and anxiety in human clinical trials: A systematic review. Journal of Nutrition & Intermediary Metabolism. https://doi.org/10.1016/j.jnim.2015.12.308

[3] “Boys who consumed L-theanine obtained significantly higher sleep percentage and sleep efficiency scores, along with a non-significant trend for less activity during sleep (defined as less time awake after sleep onset) compared to those in the placebo group.” Source: Lyon MR, Kapoor MP, Juneja LR. The effects of L-theanine (Suntheanine®) on objective sleep quality in boys with attention deficit hyperactivity disorder (ADHD): a randomized, double-blind, placebo-controlled clinical trial. Altern Med Rev. 2011 Dec;16(4):348-54. PMID: 22214254. https://pubmed.ncbi.nlm.nih.gov/22214254/

[4] According to one randomised-controlled trial, combined intake of L-theanine and caffeine improves cognitive performance and subjective alertness. See: https://pubmed.ncbi.nlm.nih.gov/21040626/

[5] “Co-administration of L-THE (280 mg) and cysteine (700 mg) before vaccination enhanced immune responses to influenza vaccine in elderly subjects with low serum total protein or haemoglobin… Co-treatment of L-THE (280 mg) and cysteine (700 mg) for 2 weeks restored the attenuation of natural killer cell activity in well trained men… Co-administration of L-THE (70 mg) and cysteine (175 mg) in 176 subjects correlated with a lower incidence in development of the common cold” See Williams, Jackson, Jane Kellett, Paul Daniel Roach, Andrew McKune, Duane Mellor, Jackson Thomas, and Nenad Naumovski. 2016. "l-Theanine as a Functional Food Additive: Its Role in Disease Prevention and Health Promotion" Beverages 2, no. 2: 13. https://doi.org/10.3390/beverages2020013

[6] “L-Theanine intake resulted in a reduction in the heart rate (HR) and salivary immunoglobulin A (s-IgA) responses to an acute stress task relative to the placebo control condition.” See: https://pubmed.ncbi.nlm.nih.gov/16930802/

[7] See https://www.webmd.com/heart/what-is-heart-rate-variability

[8] “Single l-theanine administration (20 mg/kg) significantly increased REM time, when compared to controls (88.6%) (p < 0.01), but there was no significant difference in NREM sleep time compared to the control group… NREM (20.7%) and REM (99.6%) were also significantly increased than control levels, after GABA/l-theanine mixture was orally administered (Figure 3(C,D), p < 0.05 and p < 0.001, respectively). The oral administration of GABA/l-theanine mixture improved sleep time and quality, as compared to GABA or l-theanine alone.” See: Kim S, Jo K, Hong KB, Han SH, Suh HJ. GABA and l-theanine mixture decreases sleep latency and improves NREM sleep. Pharm Biol. 2019 Dec;57(1):65-73. doi: https://doi.org/10.1080/13880209.2018.1557698. PMID: 30707852; PMCID: PMC6366437.

[9] Boyce R, Glasgow SD, Williams S, Adamantidis A. Causal evidence for the role of REM sleep theta rhythm in contextual memory consolidation. Science. 2016 May 13;352(6287):812-6. doi: https://doi.org/10.1126/science.aad5252. PMID: 27174984.

[10] If you drink a lot of coffee during the day or drink coffee late in the day, L-theanine may help neutralise the negative effects caffeine can have on sleep efficiency – this was indicated in a study on nine young women See: https://pubs.rsc.org/en/content/articlehtml/2023/fo/d3fo01247f