Dataset for Ethical Governance Research on Brain–Computer Interfaces in China (2008–2025): A Systematic Mapping Review Based on 208 Chinese Publications

Reference tracking and snowball searching were also conducted. After deduplication, title and abstract screening, full-text review, and application of predefined inclusion and exclusion criteria, 208 Chinese academic publications focusing on the ethical, legal, governance, policy, or social implications of human brain–computer interfaces were finally included.The dataset contains the list of included publications, search strategy and screening records, coding tables, binary wide-format data, and aggregated statistical data.

The list of included publications records basic bibliographic information such as title, author, publication year, journal, geographic region, and database source. The search strategy and screening record document database sources, search terms, time range, inclusion and exclusion criteria, and the screening process. The coding table records the coded results for each publication across six dimensions: publication year, geographic region, application scenario, disciplinary background, key ethical themes, and content type.

The binary wide-format dataset converts multi-label variables into 0–1 dummy variables for descriptive statistics, cross-tabulation, ethical theme co-occurrence analysis, and exploratory Pearson φ correlation analysis. The aggregated statistical data include frequency counts, cross-tabulations, theme co-occurrence matrices, and correlation analysis results. The main units of analysis are publication counts and binary variable values, where “1” indicates that a publication involves a given category and “0” indicates that it does not.To improve data quality and reproducibility, literature screening and variable coding were conducted independently by two researchers.

Disagreements were resolved through full-text review and discussion. During data processing, synonymous or closely related terms were standardized. For example, terms such as “neural privacy,” “brain privacy,” and “mental privacy” were coded under the theme of privacy, while terms such as “autonomy,” “free will,” “agency,” and “behavioral control” were coded under the theme of autonomy and free will.

Multi-label variables were split and transformed into binary dummy variables to generate a wide-format dataset suitable for statistical analysis. Data cleaning included standardizing category labels, removing redundant spaces, merging synonymous expressions, and checking duplicate records and abnormal codes. Since this study is based on publicly available academic publications, the dataset does not contain personal private information or raw participant data.

For copyright reasons, full texts of the included publications are not provided; only bibliographic metadata, coding results, and data required for statistical analysis are included.This dataset can be used to reproduce the descriptive statistics, cross-tabulation analysis, ethical theme co-occurrence analysis, and exploratory Pearson φ correlation analysis reported in the article. It may also serve as a reference dataset for future research on BCI ethics and governance, neuroethics, technology ethics, science and technology policy, and bibliometric analysis.

Data organization and statistical analysis were mainly conducted using Microsoft Excel and Python, including multi-label variable processing, binary dummy-variable transformation, frequency statistics, cross-tabulation, co-occurrence matrix calculation, and Pearson φ correlation analysis.