TL;DR. Some claims about what reading does to your mind and body are solid, some are overstated and some cannot be verified at all. In this update we rebuilt five changes from primary research. In a 19-person pilot study at Emory University, people who read a novel over nine evenings showed increased brain connectivity the following mornings, but most of that increase faded quickly once the novel was finished. In a controlled experiment with 12 people at Brigham and Women’s Hospital (Harvard Medical School), reading from a light-emitting tablet before bed meant falling asleep about 10 minutes later than reading a printed book, with evening melatonin suppressed by 55%. The idea that fiction improves social cognition found support in a meta-analysis of 14 studies, but with a small effect (g = 0.15-0.16), and the most famous single study failed to replicate with 792 participants. The link between how much you read and your comprehension and language skills grows stronger with age, according to a meta-analysis of 99 studies. And frequent mental activity in later life is associated with slower cognitive decline, even after accounting for disease markers in the brain. We could not find a peer-reviewed publication behind the old claim that “six minutes of reading reduces stress by 68%”, so we no longer use that number.
What did we change in this update?
The 2020 version of this article supported five claims with links to news websites. In 2026 we traced every claim back to its source and applied three rules: use only numbers from primary publications we actually opened and read, state what each study really measured, and be explicit about the difference between correlation and causation.
As a result, three claims were removed:
- “Six minutes of reading reduces stress by 68%.” Secondary sources attribute this figure to a 2009 report by Mindlab International. We found no peer-reviewed publication containing the claim in PubMed or Europe PMC. We do not present a number as fact when its method, sample and statistics cannot be examined.
- “Keen readers have a 50% larger vocabulary.” We could not open a primary text containing this figure. We replaced it with a meta-analysis of 99 studies that answers the same question.
- “Fiction readers are more flexible and creative.” We could not access the abstract or sample details of the 2013 study cited for this, so we removed it as well.
The effect of reading on lifespan and the comprehension gap between paper and screen are covered in a separate CEOtudent article, so this piece focuses on measurable changes in the mind and body.
Change 1: Does a novel change the connections in your brain?
In 2013, Berns and colleagues at Emory University published an intriguing experiment in Brain Connectivity. Participants had a resting-state functional MRI scan every morning for 19 consecutive days. The first 5 days provided a baseline only. For the next 9 days, participants read roughly one ninth of a novel each evening (Robert Harris’s historical novel Pompeii) and were scanned the next morning. For the final 5 days, measurements continued without reading.
The findings came in two layers. On the mornings after reading, connectivity rose in regions associated with perspective taking and story comprehension (the left angular/supramarginal gyrus and right posterior temporal areas), but the increase faded quickly after the novel ended. In the somatosensory cortex, which is linked to bodily sensation, changes persisted for several days after the reading was over. The authors interpreted this as a possible mechanism linked to readers mentally simulating a character’s bodily experience.
Limits: 21 people took part, 2 were excluded from the analysis, leaving 19. Participants were aged 19-27. The authors explicitly describe the work as a “pilot study”, and there is no separate control group; each participant was compared with their own measurements before and after reading. So the conclusion is not “reading a novel permanently changes your brain” but “in a small group, a single novel left a measurable and mostly temporary trace”.
Change 2: How does reading before bed affect your sleep?
The most clearly measured change in the body has less to do with reading itself than with what you read on. In a study published in PNAS in 2015, Chang and colleagues kept 12 healthy young adults in a laboratory for 14 days. For five evenings, participants spent roughly 4 hours before bed reading on a light-emitting tablet in dim light; for five other evenings, they read printed books. The order of the conditions was randomised.
The measured differences:
- Tablet readers fell asleep after an average of 25.65 minutes, printed-book readers after 15.75 minutes. The difference is about 10 minutes (CEOtudent calculation: 9.9 minutes).
- In the tablet condition, evening melatonin was suppressed by 55%; in the printed-book condition there was no suppression.
- The onset of melatonin secretion, a marker of the body clock, came 90 minutes later in the tablet condition (22:31 versus 21:01).
- REM sleep was shorter in the tablet condition (109.04 versus 120.86 minutes), and participants felt less alert the next morning.
Limits: The sample was 12 people, all young and healthy, the reading session (about 4 hours) was longer than a typical evening, and room light was deliberately kept very dim. The result cannot be generalised directly to every current device and screen setting. Still, the mechanism (exposure to short-wavelength light in the evening) is well established, and the practical lesson is simple: reserving bedtime reading for paper or a non-light-emitting screen is a sensible default. If sleep problems persist, seeing a sleep specialist matters more than changing your reading habit.
Change 3: Does reading fiction really increase empathy?
This is the most debated topic in reading research, and its story is instructive.
The claim: In 2013, Kidd and Castano published five experiments in Science. Participants who read literary fiction performed better on theory-of-mind tests (inferring other people’s feelings and thoughts) than those who read popular fiction, nonfiction or nothing at all. The authors also noted that the effect was “temporary”.
The replication: In 2016, Panero and colleagues repeated the experiment in the Journal of Personality and Social Psychology, with three independent research groups and 792 participants. Literary fiction showed no significant advantage over any of the other conditions. One notable finding held up, however: scores on the Author Recognition Test, which measures lifetime exposure to fiction, predicted theory-of-mind scores across all conditions. The authors wrote that the most plausible explanation is that people with strong theory of mind are drawn to fiction, or that a lifetime of reading slowly strengthens the skill, or both; other variables such as verbal ability may also play a role.
The meta-analysis: In 2018, Dodell-Feder and Tamir combined 53 effect sizes from 14 studies in the Journal of Experimental Psychology: General. Compared with reading nonfiction or not reading at all, reading fiction produced a small but statistically significant improvement in social-cognitive performance (g = 0.15-0.16). The effect held up across sensitivity analyses.
How to read this: The evidence that a single passage will make you more empathetic overnight is weak. That heavy fiction reading over the long run goes hand in hand with better social cognition is a consistent finding, but how much of it is caused by reading is still unclear.
Change 4: Does reading grow your vocabulary and comprehension?
In place of the old “50% larger vocabulary” claim, we used the study that answers this question with the broadest data. Mol and Bus’s 2011 meta-analysis in Psychological Bulletin combined 99 studies with a total of 7,669 participants on leisure-time reading (print exposure) from preschool to university.
It found moderate to strong correlations between print exposure and reading comprehension, technical reading and spelling. More striking is that the relationship strengthens with age: the share of differences in oral language skills explained by print exposure rises from 12% in preschool to 34% at university. The authors describe this as an “upward spiral”: good readers read more, and those who read more get better every year. An important extra finding is that poor readers also appear to benefit from independent leisure-time reading.
Limits: These are correlational data. The relationship runs both ways, and the effect of reading alone cannot be isolated. Still, the picture is consistent: language skill behaves like capital that builds up over the years alongside how much you read.
Change 5: Does reading protect the ageing brain?
In 2013, Wilson and colleagues examined data from 294 people in the Rush Memory and Aging Project in the Chicago area, published in Neurology. On enrolment, participants reported how often they took part in mentally stimulating activities such as reading books, visiting a library and writing letters. They completed cognitive tests every year for an average of 5.8 years, and their brains were examined at autopsy after death. The average age at death was 89.3.
Even after accounting for plaques, tangles, infarcts and Lewy bodies in the brain, more frequent mental activity in later life was associated with slower cognitive decline. Compared with average activity (50th percentile), decline was 48.4% faster in the low-activity group (10th percentile) and 32.3% slower in the high-activity group (90th percentile). Early-life and late-life activity together explained 14% of the variability in decline not accounted for by disease markers. The authors found this consistent with the “cognitive reserve” hypothesis.
Limits: What was measured is not book reading alone but a composite of mental activities that includes reading. Activity frequency was self-reported, and early-life activity was recalled retrospectively. Because the study is observational, alternative explanations, such as people who are already cognitively healthier reading more, cannot be fully ruled out. If you have a personal concern about cognitive decline, the first step is a neurologist or geriatric specialist, not a reading plan.
The evidence scorecard: how far can you trust each claim?
Table 1. Reading claims and strength of evidence (CEOtudent editorial framework). Each row is our editorial team’s assessment based on what the relevant primary publication actually measured.
| Popular claim | What the study actually measured | Design and sample | Strength of evidence | Verdict |
|---|---|---|---|---|
| “Reading a novel changes your brain” | Resting-state brain connectivity on the mornings after reading | Pilot, no control group, 19 people | Weak to moderate | A measurable but mostly temporary trace |
| “Reading before bed helps you sleep” | Differences in sleep, melatonin and body clock between tablet and printed book | Randomised crossover experiment, 12 people | Moderate | Paper or a non-light-emitting screen is the safer choice |
| “Reading fiction increases empathy” | Theory-of-mind tests right after reading | Meta-analysis of 14 studies; the famous study’s 792-person replication failed | Moderate (small effect) | Small effect; no miracle from a single passage |
| “Heavy readers have a bigger vocabulary” | Link between print exposure and comprehension, language and spelling | 99 studies, 7,669 people, correlational | Strong association, partial causation | Capital that builds up over years |
| “Reading protects the brain against ageing” | Frequency of mental activity and rate of cognitive decline | Cohort with autopsy, 294 people, observational | Moderate | Consistent association; reading is not the only factor |
| “Six minutes of reading cuts stress by 68%” | Unknown; no peer-reviewed publication found | Method cannot be examined | Unverifiable | Do not use the number |
The verified numbers in one table
Table 2. Measured values behind the five changes (verified data). Sources: Berns et al. (2013), Chang et al. (2015), Dodell-Feder and Tamir (2018), Panero et al. (2016), Mol and Bus (2011), Wilson et al. (2013). Difference columns are CEOtudent calculations.
| Study | Measure | Value |
|---|---|---|
| Berns et al., 2013 | Participants analysed / scan days / reading evenings | 19 / 19 / 9 |
| Chang et al., 2015 | Time to fall asleep, tablet vs printed book | 25.65 min vs 15.75 min (difference 9.9 min) |
| Chang et al., 2015 | Evening melatonin suppression, tablet | 55% |
| Chang et al., 2015 | Melatonin onset, tablet vs printed book | 22:31 vs 21:01 (difference 90 min) |
| Chang et al., 2015 | REM sleep, tablet vs printed book | 109.04 min vs 120.86 min (difference 11.8 min) |
| Dodell-Feder and Tamir, 2018 | Effect of fiction reading on social cognition | g = 0.15-0.16 (14 studies, 53 effect sizes) |
| Panero et al., 2016 | Literary fiction advantage in the replication | Not significant (792 participants) |
| Mol and Bus, 2011 | Share of oral language differences explained by print exposure | Preschool 12%, university 34% |
| Wilson et al., 2013 | Rate of decline relative to average | 48.4% faster with low activity, 32.3% slower with high activity |
In the AI era: does reading a summary produce the same changes?
In 2026, many readers read a few AI-generated paragraphs instead of a whole book or report. What the changes above have in common is time and effort: a novel read over nine evenings in the Berns study, reading accumulated over years in Mol and Bus, lifelong mental activity in Wilson. We found no peer-reviewed study that directly tests whether a short summary can substitute for these processes.
The closest data come from a preprint. Kosmyna and colleagues at the MIT Media Lab used EEG to monitor 54 participants writing essays in three groups: with ChatGPT, with a search engine, or with no tools. The no-tools group showed the strongest and most widely distributed brain connectivity; the large language model group showed the weakest and struggled to quote accurately from their own essays. The study is about writing rather than reading, its sample is small, and it is a preprint that has not been peer reviewed, so we treat it as a signal worth noting rather than as evidence.
Our editorial takeaway: a summary is a good filter for deciding whether a text deserves your time; deep reading is where the changes above actually happen. Think like a CEO and a summary manages the flow of information, while deep reading builds judgement. Think like a student and if you cannot restate what you learned from a summary in your own words, you have not learned it yet.
How do you turn this into a weekly system?
A practical framework that follows from the evidence scorecard:
- Move evening reading to paper. For bedtime reading, choose a printed book or a non-light-emitting e-ink reader. This is where the clearest bodily difference was measured.
- Finish at least one long text a week. The connectivity and vocabulary findings come from sustained reading, not fragmented content.
- Keep fiction on the list, but do not expect miracles. Its effect on social cognition is small; its real value lies in what accumulates over years.
- Do not make reading your only mental activity. The Wilson study measured reading together with activities such as writing and learning; combining them makes more sense.
- Use AI summaries as a gateway. Read the texts you find important in full after the summary, and write the main idea yourself in three sentences.
Frequently asked questions
Does reading books really reduce stress?
Many people find reading relaxing, but we could not find a peer-reviewed source for the widely shared “68% in six minutes” figure. So we do not give an exact number for how much reading reduces stress. Persistent stress that affects daily life calls for a mental health professional.
Does reading a novel permanently change the brain?
The evidence does not say that. In the 19-person pilot study by Berns and colleagues, connectivity increased after reading, but most of it faded quickly once the novel was finished. Only some changes in regions linked to bodily sensation lasted several days.
Is reading on a phone or tablet before bed harmful?
In the experiment by Chang and colleagues, reading on a light-emitting tablet delayed sleep onset by about 10 minutes compared with a printed book and suppressed melatonin by 55%. The study was small and the conditions extreme, but paper or a non-light-emitting screen is a safer default for evening reading.
Will reading literary fiction make me more empathetic?
Probably a little, but not instantly. A meta-analysis of 14 studies found a small effect (g = 0.15-0.16); the most famous single study failed to replicate with 792 participants. The link between long-term fiction reading and social cognition is more consistent.
Does reading in later life prevent dementia?
No, there is no evidence that says so. The study by Wilson and colleagues showed that frequent mental activity, including reading, was associated with slower cognitive decline independent of disease markers in the brain. That is an association, not proof of prevention. If you are concerned, see a doctor.
Can AI summaries replace reading books?
We found no peer-reviewed study that tests this directly. All of the changes above are linked to sustained, effortful reading. Using a summary to decide what to read in depth is a reasonable middle path.
Sources
- Berns, G. S., Blaine, K., Prietula, M. J., Pye, B. E. (2013). Short- and long-term effects of a novel on connectivity in the brain. Brain Connectivity, 3(6), 590-600.
- Chang, A.-M., Aeschbach, D., Duffy, J. F., Czeisler, C. A. (2015). Evening use of light-emitting eReaders negatively affects sleep, circadian timing, and next-morning alertness. Proceedings of the National Academy of Sciences, 112(4), 1232-1237.
- Kidd, D. C., Castano, E. (2013). Reading literary fiction improves theory of mind. Science, 342(6156), 377-380.
- Panero, M. E., Weisberg, D. S., Black, J., Goldstein, T. R., Barnes, J. L., Brownell, H., Winner, E. (2016). Does reading a single passage of literary fiction really improve theory of mind? An attempt at replication. Journal of Personality and Social Psychology, 111(5), e46-e54.
- Dodell-Feder, D., Tamir, D. I. (2018). Fiction reading has a small positive impact on social cognition: A meta-analysis. Journal of Experimental Psychology: General, 147(11), 1713-1727.
- Mol, S. E., Bus, A. G. (2011). To read or not to read: A meta-analysis of print exposure from infancy to early adulthood. Psychological Bulletin, 137(2), 267-296.
- Wilson, R. S., Boyle, P. A., Yu, L., Barnes, L. L., Schneider, J. A., Bennett, D. A. (2013). Life-span cognitive activity, neuropathologic burden, and cognitive aging. Neurology, 81(4), 314-321.
- Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I., Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. arXiv preprint 2506.08872 (not peer reviewed).
This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.
This post is also available in:
){kind=link}
&description=&image=https://ceotudent.com/wp-content/uploads/2021/07/vb602n.jpeg){kind=link}