Signal Loom: offline synthetic EEG/ECG-like living score for performance art

Abstract

Signal Loom is an offline browser performance-art prototype by Spark AI NLP. Separate synthetic EEG-like and synthetic ECG-like streams become an interactive visual score: an EEG slow-envelope feature expands an abstract 2D field; ECG-derived synthetic R–R intervals set pulse rhythm; an adjustable β/τ injection draws a timing thread. Contents. Browser UI (index.html, app.js, styles.css), signal core (signal-model.js), mappings (mapping.js), held-out robustness and shuffled-code mapping-control checks (validation.js), and Node tests with no npm dependencies. Local static serve via Python http.server. Results framing. The β/τ and beat-alignment checks are a matched-model synthetic software check: generation and fitting use the same synthetic beat-template family, and fitting receives known synthetic clock-offset corrections. They are not a blind estimator, not neuroscience validation, and not evidence of real EEG–ECG physiology or generalization. Limits. Evidence: SYNTHETIC only. Not a medical device; not diagnostic; not for treatment, patient care or clinical decision-making; not a physical hologram; does not read thoughts or represent a real person's mind or heart. No live data, sensor/device connection, audio, camera, or network request.