Multi-Site EEG Harmonisation
Improve EEG Comparability Across Devices, Sites and Visits
Multi-site EEG harmonisation is the process of making recordings from different centres, devices and operators comparable enough to pool and analyse together. Without it, site becomes a confound and apparent differences may reflect equipment rather than biology.
Neuromap™ supports harmonisation through a study-specific acquisition specification and central quality control, so that comparability is designed in from the start rather than repaired afterwards.
Sources of Site Variability
- Different EEG systems, amplifiers and electrode types
- Montage and reference differences
- Sampling rate and filter settings
- Operator technique and impedance practice
- Recording environment and artefact profile
- Metadata completeness and file formatting
Harmonisation Strategy
Harmonisation combines upstream standardisation and downstream central processing. Upstream, a shared acquisition specification aligns montage, sampling, conditions and metadata. Downstream, central quality control and a fixed pipeline ensure every recording is checked and processed identically, with residual differences documented.
Central Quality Control
Central QC screens each recording for completeness, channel integrity, signal quality, artefact burden and protocol conformity against defined acceptance criteria. Recordings that fail are flagged and escalated through predefined routes rather than silently included.
Assessing Residual Differences
Even after harmonisation, some between-site variance remains. Neuromap™ can support reproducibility assessment and controlled comparison across devices, sites or cohorts so that residual effects are quantified and considered in interpretation.
A Practical Multi-Site EEG Harmonisation Plan
Effective multi-site EEG harmonisation follows a practical sequence. First, agree a shared acquisition specification covering montage, sampling, conditions and metadata. Second, train sites against that specification. Third, route all recordings through central quality control. Fourth, process every recording with one fixed pipeline. Fifth, quantify residual between-site variance. Neuromap™ can support each step of this multi-site EEG harmonisation plan.
- Shared acquisition specification issued to all sites
- Site set-up checks and test recordings
- Central quality control on every recording
- Single fixed analytical pipeline
- Quantification of residual site effects
Test Recordings and Site Qualification
Before a site contributes study data, a small number of test recordings can confirm that its equipment and technique meet the specification. This site-qualification step catches montage, reference and formatting problems early, when they are cheap to fix, and is a core part of robust multi-site EEG harmonisation.
Reporting Harmonisation in Publications
Journals increasingly ask how multi-site data were harmonised. Neuromap™ documents the acquisition specification, central quality control and residual-variance assessment so that the harmonisation methods can be reported clearly, strengthening the credibility of multi-site EEG research.
Multi-Site EEG Harmonisation for International Studies
International and multi-centre studies raise the stakes for multi-site EEG harmonisation, because sites may use different equipment, conventions and languages of documentation. A shared acquisition specification and central quality control become even more important. Neuromap™ supports harmonised EEG analysis across multi-centre and international studies so that data collected in different places remain comparable.
- Consistent acquisition specification across countries
- Central quality control regardless of site location
- One fixed analytical pipeline for all sites
- Documented, quantified residual site effects
Harmonisation and Reproducibility Go Together
Multi-site EEG harmonisation and reproducibility are closely related. A harmonised, centrally processed dataset can improve reproducibility when acquisition, preprocessing, software versions, exclusions and site effects are fixed or documented. Neuromap™ treats harmonisation as part of the study’s research infrastructure rather than as a one-off correction at the end.
What You Receive
Harmonisation engagements deliver a comparable, centrally processed dataset with documented residual between-site variance.
- Structured, machine-readable data exports keyed by participant and visit
- Clear tables and publication-oriented figures where requested
- A documented record of settings, definitions and any exclusions
- Study-level summaries at agreed milestones
- Optional specialist scientific review at the contracted service level
Every deliverable is prepared in professional British English and is designed to slot directly into the study team’s statistical, regulatory or manuscript workflow.
Getting Started
A shared specification and site checks are the practical starting point.
The usual first step is a short feasibility review. The study team shares the research outline, the expected data volume, the acquisition details and the deliverables required. Compatibility is then confirmed and a study-specific scope and quotation provided before any commitment.
Training Sites and Checking Compliance
Harmonisation is as much about people as about equipment. Even with a clear specification, sites vary in technique, impedance practice and documentation habits. Brief training against the specification, followed by review of a small number of test recordings, catches most problems before study data are collected. This up-front investment is far cheaper than discovering, after recruitment, that recordings from one centre are not comparable with the rest.
Ongoing checks matter too. Periodic review of incoming recordings against the specification keeps standards from drifting over a long study. When a site’s recordings begin to deviate, early detection allows a quick correction rather than the loss of weeks of data. Documenting these checks also strengthens the eventual publication, because reviewers increasingly expect a clear account of how comparability was established and maintained across every contributing centre.
Key Terms Explained
The following terms appear across this area of research and may help teams new to advanced electrophysiology.
Site effect
Variance attributable to the site rather than to biology.
Acquisition specification
A shared definition of how every site should record.
Site qualification
Confirming, via test recordings, that a site meets the specification.
Volume conduction
Spread of electrical activity that can confound sensor-level measures.
Residual variance
Between-site differences remaining after harmonisation, which are quantified.
Central processing
Applying one fixed pipeline to all recordings regardless of origin.
A Note on Responsible Use
All outputs are provided for research use. They consist of data and neutral observations, and a suitably qualified clinician or investigator retains responsibility for interpretation, clinical meaning and any decisions that follow.
Building Comparability Into the Study
For any programme drawing data from more than one centre, comparability is not something that can be added at the analysis stage; it has to be built into how recordings are captured and checked from the very first participant. A shared specification, brief site training, test recordings and consistent central processing together turn a collection of centre-specific datasets into one coherent whole. Teams that invest in this groundwork early tend to spend far less time later reconciling differences that could have been prevented, and produce results that reviewers accept with fewer questions about method.
Summary
Multi-site EEG harmonisation makes recordings from different centres comparable through upstream standardisation and downstream central processing. Neuromap™ supports harmonisation and central quality control so that site becomes a documented, quantified factor rather than a hidden confound.
Scope
Neuromap™ provides EEG and qEEG data analytics for research. Outputs are data and neutral observations only; a suitably qualified clinician or investigator completes all interpretation. Neuromap™ does not provide a clinical diagnosis and is not an accredited clinical diagnostic laboratory.
Frequently asked questions
Common questions about multi-site EEG harmonisation and central quality control.
Can different EEG machines be pooled?
Potentially, with a shared acquisition specification and central QC. Compatibility and residual site effects are assessed at feasibility.
What is central QC?
A consistent set of technical checks applied to every recording in a study before advanced analysis, so quality is comparable across sites.
Does harmonisation remove all site differences?
No. It reduces and documents them; residual variance should be assessed and considered in analysis.
Do all sites need identical equipment?
Not necessarily. A shared acquisition specification and central processing allow comparable data from different systems, with residual differences assessed.
Can you help write the acquisition specification?
Yes. A study-specific specification covering montage, sampling, conditions and metadata can be provided for site training.
Limitations and important notes
- Harmonisation cannot fully retrofit comparability onto poorly standardised acquisition.
- Residual site effects should be quantified, not assumed absent.
- Any published figure for analytical-output count, reference datasets, processing capacity or turnaround is conditional on the documented platform version and the requirements of the specific study.