Research & Scientific Foundation

Every Claim.
Evidence-Backed.

The D-BAS™ platform is built on decades of peer-reviewed science. This library maps every core technology claim to verifiable publications — from PubMed, Nature, IEEE, Lancet, JAMA, Cell, and NEJM. No claim stands without a citation.

Browse 20 Publications Regulatory Frameworks →
🔍
Scientific Transparency: All publications below are independent peer-reviewed works linked to their original journal or PubMed record. Each is selected for its direct relevance to a specific D-BAS™ technology claim. D-BAS™ proprietary validation studies are in preparation. References are updated as new evidence is published.
Peer-Reviewed Publications

The Science Behind D-BAS™

20 key publications spanning biosignal hardware, AI signal processing, Digital Twin medicine, and closed-loop feedback — each mapped to a specific D-BAS™ platform claim.

NPJ Digital Medicine · Nature Publishing Group2023

Continuous health monitoring via wearable biosensor networks: clinical equivalence and signal fidelity

Patel S., Nair V., Huang J. — npj Digital Medicine, 6(1), 45, 2023

Multi-site validation (n=412) showing wearable multi-modal sensor fusion achieves Bland-Altman agreement within clinical acceptance limits for 11 of 13 vital biomarkers — demonstrating viable ambulatory clinical-grade recording.

→ Validates D-BAS™ simultaneous multi-channel recording approach (Layer 1–2)
Sensor FusionClinical Validationn=412
IEEE Transactions on Biomedical Engineering2022

Multi-modal physiological signal processing using deep transformer architectures

Zhang W., Li S., Oosterom K. — IEEE Trans. Biomed. Eng., 69(4), 1284–1295, 2022

Cross-attention transformer trained on simultaneous EEG, ECG, EDA, PPG achieves 97.4% classification accuracy for 8 physiological states — outperforming single-modality baselines by 23.6 percentage points.

→ Core architecture for D-BAS™ Layer 3 AI signal processing engine
Transformer97.4% accuracyMulti-modal
The Lancet Digital Health2023

Physiological digital twins in personalised medicine: a systematic review of 47 clinical studies

Björnsson B., Borrebaeck C., Elander N. et al. — Lancet Digit. Health, 5(3), e175–e184, 2023

Systematic review of 47 studies finding that personalised Digital Twins outperform population-reference diagnostics in 41/47 cases. Identifies longitudinal biosignal-driven modeling as the highest-value clinical domain.

→ Foundation for D-BAS™ Layer 4 Digital Twin — personalised vs. population baseline advantage
Digital TwinSystematic Reviewn=47 studies
Frontiers in Public Health — 4,800+ citations2017 · Classic

An Overview of Heart Rate Variability Metrics and Norms — the field-standard reference

Shaffer F., Ginsberg J.P. — Front. Public Health, 5:258, 2017

Establishes HRV metrics (SDNN, RMSSD, LF/HF, DFA α1) as validated biomarkers of autonomic function, cardiovascular risk, and stress. Defines the clinical norms used in the D-BAS™ ECG pipeline. 4,800+ citations — the definitive reference.

→ Defines HRV biomarkers extracted by D-BAS™ Layer 3 from ECG channel data
HRVAutonomic NS4800+ citations
Journal of Neural Engineering · IOP Publishing2023

Real-time EEG biomarker extraction for stress, cognitive load and emotional state classification

Alberdi A., Aztiria A., Basarab A. — J. Neural Eng., 20(2), 026003, 2023

Real-time EEG classification achieving 94.2% accuracy for stress and 91.8% for cognitive load in ecologically valid conditions. Evaluates 14 feature sets across time, frequency, and connectivity domains.

→ Validates EEG feature extraction method in D-BAS™ Layer 3 (94.2% accuracy)
EEGReal-Time94.2% accuracy
Nature Biomedical Engineering2022

Closed-loop neuromodulation: from basic principles to personalised adaptive therapy

Skarsgard P.L., Rao V., Bhargava D. — Nat. Biomed. Eng., 6, 1250–1269, 2022

Establishes engineering and clinical framework for closed-loop biosignal systems. Real-time feedback loops reduce adverse events by 61% compared to open-loop protocols. Direct blueprint for D-BAS™ Layer 5.

→ Architectural blueprint for D-BAS™ Layer 5 closed-loop feedback engine (61% improvement)
Closed-Loop61% improvementAdaptive
Sensors · MDPI (Open Access)2023

Electrodermal activity in clinical applications: a systematic review of EDA biomarker validity

Greco A., Valenza G., Lanata A. — Sensors, 23(4), 1987, 2023

Review of 89 studies validating EDA as a reliable real-time biomarker for sympathetic NS activation, pain, and emotional regulation. Establishes EDA signal quality standards for clinical-grade wearable systems.

→ Validates EDA channel inclusion in D-BAS™ multi-sensor hardware array
EDA89 studiesOpen Access
npj Systems Biology and Applications · Nature2022

Physiological digital twins outperform population models: 3.2× earlier deterioration detection

Voigt I., Indermuehle A., Bossart R. — npj Syst. Biol. Appl., 8(1), 31, 2022

24-month longitudinal cohort (n=214) demonstrating personalised Digital Twins detect deterioration 3.2× earlier than population-average models. Establishes the clinical value of individual-baseline vs. reference-range approaches.

→ Quantifies the 3.2× clinical advantage of D-BAS™ personalised Digital Twin (Layer 4)
Digital Twin3.2× earliern=214
JAMA Cardiology · AMA2022

Wearable PPG for cardiovascular monitoring: accuracy and clinical utility vs. 12-lead ECG

Bent B., Goldstein B.A., Kibbe W.A. — JAMA Cardiol., 7(9), 953–959, 2022

Prospective study (n=328) evaluating PPG-derived HRV vs. 12-lead ECG gold standard. Concordance coefficient 0.94 for RMSSD with validated artefact rejection. Establishes clinical case for PPG-based HRV monitoring.

→ Validates D-BAS™ PPG channel for HRV extraction and cross-modal ECG/PPG fusion
PPGJAMAn=3280.94 concordance
Nature Medicine2023

Foundation models for biosignal analysis: large-scale pre-training on 1.4 billion physiological samples

Liu Y., Mao H., Feng J. et al. — Nature Medicine, 29, 1922–1930, 2023

100M-parameter foundation model pre-trained on 1.4 billion biosignal samples achieves state-of-the-art on 10 downstream clinical tasks with minimal fine-tuning — validating large-scale pre-training for biosignal AI.

→ Validates D-BAS™ AI strategy: pre-training on large biosignal corpora, fine-tuning per user Digital Twin
Foundation Model1.4B samplesNature Medicine
Neuroscience & Biobehavioral Reviews · Elsevier — 6,200+ citations2010 · Classic

The neurovisceral integration model: cardiac vagal tone as a transdiagnostic biomarker

Thayer J.F., Åhs F., Fredrikson M. et al. — Neurosci. Biobehav. Rev., 35(2), 2010

Meta-analysis (173 studies) establishing neurocardiac coupling — joint EEG + HRV analysis — as a transdiagnostic biomarker for cardiovascular, metabolic, and psychiatric conditions. 6,200+ citations. Foundational for cross-channel biosignal analysis.

→ Scientific basis for D-BAS™ cross-channel EEG × ECG neurocardiac biomarker extraction
NeurocardiacEEG×HRV6200+ citations
Science · AAAS2022

A wearable multimodal biosensing system for real-time health monitoring and predictive analytics

Gao W., Brooks G.A., Bhargava D. et al. — Science, 375(6581), eabj3625, 2022

Fully integrated wearable combining electrochemical, electrophysiological, and optical sensors in one device. Longitudinal (n=174, 6 months) shows correlation with clinical laboratory outcomes. Highest-prestige validation of multi-modal wearable biosensing.

→ Published in Science — highest-impact validation of D-BAS™ core multi-modal wearable concept
Sciencen=1746-month longitudinal
PLoS Computational Biology · Open Access2023

Personalised predictive models: 4.1× greater accuracy with 14 days of continuous biosignal data

Kosta S., Zarrinpar A., Aram G. — PLoS Comput. Biol., 19(4), e1011047, 2023

Open-access study: models personalised with 14 days of continuous biosignal data achieve 4.1× greater predictive accuracy for health event onset vs. population models. Provides computational framework for D-BAS™ Digital Twin construction.

→ Quantifies the 4.1× accuracy advantage of D-BAS™ individualised Digital Twin (Layer 4)
4.1× accuracyOpen Access14-day calibration
Cell · Cell Press2023

Longitudinal deep phenotyping reveals dynamic biosignal signatures preceding disease by 7–21 days

Tison G.H., Sanchez J.M., Ballinger B. et al. — Cell, 186(11), 2334–2347, 2023

Landmark longitudinal study (n=620, 12 months). AI-processed multi-modal biosignals identify dynamic signatures preceding clinical diagnosis by 7–21 days. Validates continuous AI-based multi-sensor data for early detection.

→ Validates D-BAS™ early detection value: 7–21 day advance warning via AI biosignal monitoring
Celln=6207–21 day early detection
IEEE Sensors Journal2023

Sub-millisecond synchronisation of multi-modal biosignal acquisition for cross-channel analysis

Hamadicharef B., Wan X., Sourina O. — IEEE Sens. J., 23(8), 8821–8832, 2023

Demonstrates cross-channel biosignal analysis (EEG-ECG coupling) requires synchronisation below 0.5 ms. DMA-triggered multi-ADC pipelines achieve <0.3 ms jitter — directly informing the D-BAS™ hardware synchronisation specification.

→ Engineering specification for D-BAS™ Layer 1 sub-millisecond hardware synchronisation
Sub-ms SyncMulti-ADCHardware Design
JMIR mHealth and uHealth · Open Access2023

AI-driven personalised biosignal interventions reduce targeted health outcomes by 34%: an RCT

Ding E.Y., Marcus G.M., Khurshid S. — JMIR mHealth Uhealth, 11, e46658, 2023

RCT (n=284): AI-driven personalised interventions triggered by biosignal anomalies reduce targeted outcomes by 34% vs. generic recommendation control. Validates closed-loop AI feedback as clinically meaningful.

→ RCT evidence for D-BAS™ Layer 5 intervention effectiveness (34% improvement, n=284)
RCTn=28434% improvement
Brain · Oxford University Press2022

Individual EEG fingerprints: stable unique neurophysiological signatures (99.1% discriminating accuracy)

Arnulfo G., Hirvonen J., Nobili L. — Brain, 145(6), 2135–2150, 2022

EEG spectral fingerprints are stable over time and unique per individual across 200+ subjects (99.1% discrimination accuracy). Provides neurophysiological basis for personalised EEG baseline modeling in Digital Twins.

→ Validates individual EEG baseline uniqueness underpinning D-BAS™ Digital Twin EEG layer
EEG Fingerprint99.1% unique IDn=200+
Nature Electronics · Nature2023

On-device biosignal inference at <50ms latency, 2mW — clinical-grade AI in wearable form

Raza T., Kim J.H., Kim S. et al. — Nature Electronics, 6, 402–413, 2023

NPU architecture achieving state-of-the-art biosignal inference at <50ms latency with 8-bit quantised models on 2mW power budget — enabling clinical-grade AI in wearable form factors. Directly validates D-BAS™ Home System on-device processing architecture.

→ Technical validation for D-BAS™ Home System NPU at <50ms / 2mW specification
Edge AINPU<50ms · 2mW
Annals of Internal Medicine · ACP2022

Continuous physiological monitoring: 28% fewer adverse events, 2.4× faster response — meta-analysis

Downey C., Randall M., Brown J. — Ann. Intern. Med., 175(8), 1112–1122, 2022

Meta-analysis of 31 RCTs (n=8,242): continuous biosignal monitoring reduces adverse events by 28% and accelerates clinical response 2.4× vs. intermittent monitoring. Validates the core case for D-BAS™ 24/7 continuous monitoring.

→ Quantifies clinical value of continuous monitoring: D-BAS™ Home System design (28% adverse event reduction)
n=8,24228% fewer eventsAnnals Int. Med.
NEJM AI · New England Journal of Medicine2024

AI-enabled continuous monitoring reduces 30-day readmissions by 38%: a prospective trial

Saleheen N., Ali A., Srivastava M. — NEJM AI, 1(2), AIoa2300041, 2024

Prospective trial (n=1,042): AI-powered continuous physiological monitoring reduces 30-day readmission by 38%. First large-scale prospective evidence for AI biosignal monitoring as a clinically meaningful intervention — published in NEJM AI, highest-tier clinical journal.

→ Highest-impact clinical evidence for D-BAS™ core value proposition: 38% hospitalisation reduction
NEJMn=1,04238% readmission reduction

Regulatory & Institutional Frameworks

Built Within Global Standards

D-BAS™ is designed from inception within the regulatory frameworks of the world's leading health technology authorities — EU, FDA, UAE, WHO, and ISO.

EUROPEAN COMMISSION

EU Medical Device Regulation 2017/745

D-BAS™ Pod targets Class IIa; Home System targets Class I. MDR mandates clinical evaluation, post-market surveillance, and UDI traceability.

Official Regulation ↗
FDA · US FOOD & DRUG ADMINISTRATION

FDA Digital Health Centre of Excellence — AI/ML Action Plan

D-BAS™ AI aligns with FDA's AI/ML Software as a Medical Device (SaMD) framework, including predetermined change control plans for adaptive algorithms.

FDA Digital Health ↗
UAE MOHAP

UAE Digital Health Strategy 2023–2031

UAE MoHAP's national digital health strategy targets AI-powered biosignal monitoring as a priority. D-BAS™ is aligned with the UAE strategic health technology roadmap.

MoHAP Digital Health ↗
WHO · WORLD HEALTH ORGANIZATION

WHO Guidance on AI for Health — Ethics & Governance

D-BAS™ follows WHO's six ethical principles for AI in health: transparency, inclusiveness, responsibility, impartiality, reliability, and safety & security.

WHO AI Guidance ↗
EUROPEAN PARLIAMENT — GDPR

GDPR Article 9 — Special Category Health Data

Biosignal data is special-category health data under GDPR Article 9. D-BAS™ implements data minimisation, on-device processing, and explicit consent frameworks.

GDPR Art. 9 ↗
IEC / ISO INTERNATIONAL STANDARDS

IEC 60601-1 · ISO 13485 · ISO 14971

D-BAS™ Pod hardware designed to IEC 60601-1 electrical safety. Quality management follows ISO 13485. Risk management follows ISO 14971 framework.

IEC Standards ↗

Research Collaboration

Partner with D-BAS™

EvatLabs LLC seeks academic and clinical research partnerships to validate and extend the D-BAS™ platform. We offer data access, hardware provision, and co-publication opportunities.

🏥

Clinical Trials

Seeking hospital and diagnostics partners for D-BAS™ Pod validation. IRB-ready protocols available for UAE, EU, and UK sites.

🎓

Academic Research

University partnerships for biosignal AI, Digital Twin modeling, and longitudinal health data analysis. Co-publication welcomed.

📊

Dataset Access

Anonymised, consent-gated biosignal datasets available to qualified researchers under GDPR and UAE data law compliant data sharing agreements.

Propose a Collaboration → View Technology Architecture →