NeuraDock: Open-source EEG Alpha workflow and agent tutorial

NeuraDock is an open-source EEG agent and tutorial that walks users from raw files to real-time visual cognitive-load prototypes. The paper, submitted on 25 Jun 2026 (arXiv:2606.26518), provides a reproducible, step-by-step workflow that enforces a quality-gated pipeline so Alpha and workload metrics are computed only after preprocessing and QC gating.

What does the tutorial include?

The tutorial provides a runnable, end-to-end walkthrough: installation, preprocessing and quality control, Alpha dynamics figures, within-subject Rest versus Task visual cognitive-load comparison, validation on a public mini-dataset, an online dashboard, a real-time API and an LLM interpretation layer for quality risks. The authors state the goal is practical: a reader should be able to install the agent and run the full chain from files to a dashboard and external API calls.

The tutorial emphasizes a "quality-gated workflow" rather than computing features directly from raw EEG. It includes code and data links in the paper's associated materials section and offers a reproducible validation using a mini-dataset provided with the article.

How was NeuraDock validated and what did it find?

The validation used a public mini-dataset and processed 18 recordings, producing 10 within-subject comparisons; the agent observed task-related posterior Alpha suppression in 7 of the 10 contrasts. The paper also reports an estimate of initial evidence for within-subject repeatability and that the authors benchmarked local online API latency, alongside the reference validation summary the tutorial provides for comparison.

Those concrete numbers come from the paper's included mini-dataset validation and are presented as part of the tutorial's reproducible walkthrough. The validation steps let users compare their runs against the reference validation summary the authors supply.

How does the system work technically?

NeuraDock sequences standard EEG acquisition, preprocessing, QC gating, Alpha feature extraction, and online endpoints so downstream metrics are gated on quality checks rather than raw signals. The tutorial documents how to run EEG preprocessing and quality control, generate Alpha dynamics figures, perform within-subject Rest/Task comparisons and call the real-time API from an external application.

The paper also layers an LLM interpretation component to explain quality risks encountered during preprocessing and QC. The walkthrough is explicitly designed to bridge the offline-to-online gap that often requires manually combining acquisition, QC, feature extraction and a web API.

Why it matters

NeuraDock offers a reproducible path for teams that need a transparent, quality-aware route from EEG files to real-time applications. The tutorial frames practical reproducibility as its objective and supplies a mini-dataset validation with concrete counts: 18 processed recordings, 10 within-subject contrasts and 7 contrasts showing posterior Alpha suppression. That combination of code, validation and an explicit quality gate addresses a common engineering gap between offline EEG toolkits and online, real-time prototypes.

What to watch

Check the paper's associated code and data links for runnable demos and the reference validation summary to reproduce the 18-recording run. Also watch for follow-up validations that report the benchmarked local online API latency the authors measured and for community replications of the 7-of-10 posterior Alpha suppression finding.

Authors and citation: Zhiyuan Xu, Yueqing Dai, Junling Li and Junwen Luo, "NeuraDock Visual Cognitive Load Agent Tutorial: A Quality-Gated Open-Source EEG Workflow for Alpha Dynamics and Real-Time Applications," arXiv:2606.26518, submitted 25 Jun 2026.