cogneuro-paper-reader

Cognitive Neuroscience Paper Reader

Read a cognitive-neuroscience / experimental-psychology paper (PDF or plain text) and produce a structured, reproducible summary. Handles empirical studies, reviews, and methods/theory papers.

When to use

• The user shares a .pdf (or extracted text) of a paper on attention, working memory, sustained attention, memory, decision-making, or neural population coding and asks you to read, summarize, explain, or review it.

Workflow

1. Extract text. If given a PDF, extract text first (for example pdftotext -layout file.pdf file.txt), then read the text. If extraction returns no text (a scanned PDF), say so and ask for text or OCR — do not fabricate content.
2. Identify the paper type — empirical study, review/meta-analysis, or methods/theory paper — and tailor the emphasis of the output.
3. Fill the output template below. Pull exact numbers from the text; never invent statistics. Write "not reported" for anything the paper omits.
4. Define domain terms using references/domain-glossary.md so a lay reader can follow the summary.
5. Add a critical-reading section — confounds, alternative explanations, limitations the authors admit, and what you would want to verify.

Output template

1. Citation — authors, year, journal, DOI.
2. Plain-language summary — one paragraph: what they did and what they found.
3. Research question & hypotheses.
4. Methods — participants (n, population), task/paradigm, conditions, key manipulations, and measures (e.g. RT, accuracy, K, A′, response error).
5. Analyses & statistics — tests used (t-test, Spearman r, bootstrap, Bayes factor, Cohen's d) with exact values where reported.
6. Key results — numbered, each with its statistic.
7. Conclusions & implications.
8. Limitations & confounds.
9. Key terms — a short glossary of the domain terms used.
10. Critical reading — open questions, alternative explanations, what to check.

See references/statistics-guide.md for how to interpret the statistics, and references/domain-glossary.md for definitions.

Notes

• Prefer exact values; report both effect size and uncertainty (CI) when present.
• Note when a test is directional (one-tailed) or bootstrapped.
• Do not invent references or numbers; if absent, write "not reported".