Methodology
How we write our neuroscience content
This blog follows the methodology below in the interest of reproducible science and accurate medical information.
1. Research source standards
Every neuroscience claim must be backed by at least one of the following sources:
- Peer-reviewed journals: Nature, Cell, Neuron, Brain, Journal of Neuroscience, etc.
- Meta-analyses / systematic reviews: Cochrane, Campbell Collaboration
- Clinical guidelines: AAN (American Academy of Neurology), APA (American Psychiatric Association)
- Recognized health agencies: NIH, WHO, KDCA, FDA
- Large cohort studies: UK Biobank, Framingham Heart Study, ABCD Study
2. Research quality criteria
Before citing a study, we assess its quality using the following criteria:
- Sample size: animal N ≥ 10, human N ≥ 30 (observational), adequate statistical power for clinical trials
- Study design: single case reports are used only for generating hypotheses; RCTs and large cohorts are stronger evidence
- Reproducibility: results replicated by independent labs take priority
- Standing of the publishing journal: we consider impact factor and the journal's authority in its field
- Conflict-of-interest disclosure: whether industry funding and author conflicts of interest are disclosed
3. Stating the limits of research findings
The main limitations of neuroscience research are listed below, and this blog states them in the body of each post:
| Type of limitation | Example |
|---|---|
| Limits of extrapolating from animal models to humans | Findings in mice may not apply to humans |
| Single-time-point cross-sectional studies | Cannot support causal inference; they only show correlation |
| Resolution limits of fMRI / EEG | A trade-off between temporal and spatial resolution |
| Replication crisis | Reports that about 50% of neuroscience studies fail to replicate |
| Publication bias | Positive results get published while negative results go missing |
| Low statistical power | Risk of false discoveries in small samples |
| Selection bias | Clinical trial participants may differ from typical patients |
4. Statistical reporting standards
When we report statistics from a cited study, we also give:
- Effect size: Cohen's d, odds ratio, hazard ratio, etc. (a p-value alone is not enough)
- Confidence intervals: the 95% CI alongside the estimate
- Sample size: the N
- Multiple-testing correction: how voxel-level tests were corrected in brain imaging studies
- Control group / comparison: exactly what the effect was compared against
5. Language Guidelines
To convey medical information accurately, we follow these wording standards:
Recommended wording (accurate)
- "may", "is associated with"
- "One study observed..."
- "This effect was reported in N people under condition X"
- "suggests"
Prohibited wording (inaccurate / dangerous)
- "cures"
- "definitely"
- "no side effects"
- "without a prescription"
- "miracle", "revolutionary"
6. Additional standards for mental health content
When writing about depression, anxiety, ADHD or suicide risk, we:
- Follow safe suicide reporting guidelines (Journalists Association of Korea and Korea Suicide Prevention Center)
- Never describe methods of self-harm or suicide in detail
- List crisis hotline numbers (1393, 1577-0199) at the end of the post
- Include messages of hope, such as "depression is treatable"
- Avoid stigmatizing language (words such as "crazy" or "psycho" are banned)
7. Content review schedule
- Treatment guides: reviewed every 6 to 12 months
- Clinical trial information: updated as soon as new results are published
- Drug information: updated as soon as a safety warning is issued
- General neuroscience: reviewed once a year
8. Limitations and disclaimer
All content on this blog reflects the scientific consensus at the time of publication. Neuroscience moves quickly, and some information may be updated later in light of new evidence.
For our medical disclaimer, see the Medical Disclaimer.