In this post I will describe my use of NLP (Natural language processing, not neuro-linguistic programming. Natural language processing is cool, while neuro-linguistic programming is some pseudoscience stuff) in the application of summarizing articles from the internet. Specifically, I chose the topic of l-theanine and psychiatry, as previously I have already summarized the Nootropics subreddit discussions on l-theanine. The next step, therefore, is to summarize existing articles on this topic.
Summarizing experience with green tea from the Nootropics subreddit
The first step was to perform an automated Google search for a specific term. I chose the term “l-theanine psychiatry” and set the number of unique urls to be 15. Some of the resulting urls are listed below:
Can L-Theanine Help Treat Symptoms of Bipolar Disorder?
Effects of L-Theanine Administration on Stress-Related Symptoms and Cognitive Functions in Healthy Adults: A Randomized Controlled Trial
How does the tea L-theanine buffer stress and anxiety
It can be seen that the article titles are quite relevant to our topic. The next step is formatting the text and summarizing the information.
The idea behind the summarization technique is calculating word frequencies for each word in the combined text of all articles (after stop words removal), and then selecting words in the top 10% of frequencies. These words will be the ones used in scoring each sentence. More frequent words will be given more importance, as they are deemed more relevant to the chosen topic, therefore sentences containing those words will receive higher scores. This is not a machine learning approach, but a basic frequency count method. In total, 148 words were used for sentence scoring. Some of the most frequent words (from all articles combined) are listed below:
Theanine, administration, effects, placebo, weeks, study, four, sleep, scores, cognitive, may, stress, function, fluency, studies, related, symptoms, participants, bacs, anxiety
BACS was one of the top frequent words, it stands for the Brief Assessment of Cognition in Schizophrenia. Once each sentence was scores, 15 highest scoring sentences were selected in order to create a summary. The summary of the articles is presented below. From the summary we can infer that l-theanine was studied for its effects on cognition, anxiety, and stress. Some studies had positive results, indicating that l-theanine performed significantly better than placebo in regards to positive cognitive effects such as improved verbal fluency and executive function. Studies also noted significant improvements in stress reduction with the use of l-theanine. Other studies did not find any significant differences between l-theanine and placebo.
Second, only about 20% of symptoms (the PSQI subscales) and cognitive functions (the BACS verbal fluency, especially letter fluency and executive function) scores showed significant changes after L- theanine administration compared to the placebo administration, suggesting that the effects are not large on daily function of the participants.
Although psychotropic effects were observed in the current study, four weeks L-theanine administration had no significant effect on cortisol or immunoglobulin A levels in the saliva or serum, which was inconsistent with previous studies reporting that salivary cortisol  and immunoglobulin A  levels were reduced after acute L-theanine administration.
Considering the comparison to the placebo administration, the current study suggests that the score for the BACS verbal fluency, especially letter fluency, but not the Trail Making Test, Stroop test, or other BACS parameters, significantly changes in response to the 4 weeks effects of L-theanine.
The BACS verbal fluency, especially letter fluency (p = 0.001), and executive function scores were significantly increased after L-theanine administration (p = 0.001 and 0.031, respectively; ), while the Trail Making Test A and B scores were significantly improved after placebo administration (p = 0.042 and 0.038, respectively).
When score reductions in the stress-related symptoms were compared between L-theanine and placebo administrations, changes in the PSQI sleep latency, sleep disturbance, and use of sleep medication subscales were significantly greater (p = 0.0499, 0.046, and 0.047, respectively), while those in the SDS and PSQI scores showed a non-statistically significant trend towards greater improvement (p = 0.084 and 0.073, respectively), during the L-theanine period compared to placebo.
Stratified analyses revealed that scores for verbal fluency (p = 0.002), especially letter fluency (p = 0.002), increased after L-theanine administration, compared to the placebo administration, in individuals who were sub-grouped into the lower half by the median split based on the mean pretreatment scores.
Discussion In this placebo-controlled study, stress-related symptoms assessed with SDS, STAI-T, and PSQI scores decreased, while BACS verbal fluency and executive function scores improved following four weeks L-theanine administration.
The present study aimed to examine the effects of four weeks L-theanine administration (200 mg/day, four weeks) in a healthy population, i.e., individuals without any major psychiatric disorder.
The PSQI subscale scores for sleep latency, sleep disturbance, and use of sleep medication reduced after L-theanine administration, compared to the placebo administration (all p < 0.05).
The effects on stress-related symptoms were broad among the symptom indices presented in the study, although a comparison to the placebo administration somewhat limits the efficacy of L-theanine administration for some sleep disturbance measurements.
For cognitive functions, BACS verbal fluency and executive function scores improved after four weeks L-theanine administration.
PMID: 31623400 This randomized, placebo-controlled, crossover, and double-blind trial aimed to examine the possible effects of four weeks L-theanine administration on stress-related symptoms and cognitive functions in healthy adults.
The anti-stress effects of L-theanine (200 mg/day) have been observed following once- [ 33 , 34 ] and twice daily [ 35 ] administration, while its attention-improving effects have been observed in response to treatment of 100 mg/day on four separate days [ 36 ] and 200 mg/day single administration [ 37 ], which was further supported by decreased responses in functional magnetic resonance imaging [ 38 ].
These results suggest that four weeks L-theanine administration has positive effects on stress-related symptoms and cognitive function in a healthy population.