☰

BrainCapture

Portable EEG at global scale

A medical device company delivering portable EEG worldwide, with software doing the heavy lifting

Industry
Medical Devices
Headquarters
Kongens Lyngby, Denmark
Public information as of
January 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of BrainCapture's published strategy and is not endorsed by, or produced in cooperation with, BrainCapture. Company website

Strategic priorities

BrainCapture is a Danish medical-technology company founded in 2019 that has built the BC-1, a portable low-cost electroencephalogram (EEG) headset designed for clinical use outside specialist neurology departments. The company holds a CE Mark under the European Union Medical Device Regulation (MDR) achieved in early 2024 and operates an 'EEG as a Service' revenue model priced at approximately $25 per scan, where the device, the cloud upload and a remote-neurologist interpretation are sold together rather than the hardware alone.

The growth plan sets a public target of reaching 2,000 healthcare facilities across 35 countries by 2030, with a deliberate focus on low- and middle-income countries where roughly 80 percent of people with epilepsy have no access to a neurologist. As of the most recent public figures the BC-1 is in production through a contract manufacturer in Krakow, Poland, the smartphone software is developed with a partner in Cluj-Napoca, Romania, and operational hubs are active in Kenya, the Philippines and Indonesia.

Capital sits behind the plan. BrainCapture has raised approximately $1.65 million in equity across seed and later-stage rounds, with additional funding from the European Innovation Council (around $2.26 million in 2022) and other non-dilutive European sources. The funding is being deployed to scale the EEG-as-a-Service operation rather than to fund new hardware programmes, which means each new facility, country or device in the field places a fresh weight on the same underlying software, data and regulatory systems.

The load-bearing gaps sit there, not in the headset. EEG data uploads to BrainCapture's cloud over Bluetooth 5.0 from non-expert operators in electrically noisy environments, the AI interpretation pipeline is described as 'planned enhancements' rather than a production system, and the company is pursuing FDA 510(k) clearance (a US premarket notification for moderate-risk devices) while continuing Post-Market Clinical Follow-up under MDR in parallel. Multi-jurisdiction data residency, real-time data-quality assurance at the point of recording, and software-managed regulatory evidence are the same capability called three different names.

Challenges we see

  • Operations Supply Chain

    Coordinating consumables across multi-country operations

    The BC-1 kit contains an amplifier, charger, cabling, and five cap sizes. Manufacturing sits with a contract partner in Krakow, Poland, while operational hubs run in Kenya, the Philippines and Indonesia through local distributors. In 2024, roughly one in three European medical-device shippers reported difficulty securing materials on time.

    Where the consumables that determine whether a kit is usable in clinic move through a chain of distributor relationships, the position that matters is field inventory rather than the upstream factory. Joining field readings, distributor stock and kit-level consumption into one view lets replenishment follow use rather than guess it.

  • Digital Integration

    Maintaining signal integrity across a distributed fleet

    The BC-1 amplifier connects to a smartphone over Bluetooth 5.0 and uploads EEG recordings to BrainCapture's cloud. Deployment plans describe growth from current Kenyan sites to a 2,000-facility fleet by 2030, with devices expected to operate in electrically noisy clinics where power quality and electromagnetic interference vary by site.

    Where many devices run in environments that differ from the lab bench, monitoring each device the way a fleet is monitored shifts attention from the individual reading to the population of readings, so calibration drift and signal-quality changes appear on a dashboard rather than in a later report.

  • Digital Data Science

    Training AI on the populations the device serves

    BrainCapture is developing AI algorithms for automated EEG interpretation. The training pipeline that takes raw recordings from Kenya, the Philippines and Indonesia, labels them and rebuilds models on new data is described in public communications as a 'planned enhancement' rather than a production operation. Public commentary from the founders emphasises that electrodes need to work with diverse hair types.

    Where the AI is meant to read recordings across hair types, ages and clinical contexts that were thin in the training data, the test of the pipeline becomes whether new data from a region changes model behaviour rather than whether the pipeline can produce a model at all.

  • Compliance Regulatory

    Producing regulatory evidence in parallel across jurisdictions

    BrainCapture achieved CE Mark under MDR in early 2024 and is pursuing FDA 510(k) clearance. Operating in Kenya, the Philippines and Indonesia brings local data-residency requirements alongside the General Data Protection Regulation (GDPR) in Europe and the Health Insurance Portability and Accountability Act (HIPAA) compliance posture expected for US sales.

    Where filings and follow-up evidence have to be assembled across more than one regulator at the same time, the difference between compiling the evidence by hand and having it generated as a side-effect of the device data becomes the difference between a quarter spent on paperwork and a quarter spent on the next market.

  • Operations Quality

    Keeping data quality consistent with non-expert operators

    The BC-1 is designed to be operated by nurses without specialist training. Public communications describe the training as 'minimal', and the device is deployed at scale across Africa and Southeast Asia. Inconsistent electrode placement or environmental noise can produce recordings that fail clinical review.

    Where recordings are made by staff who do not have a neurophysiology background, the question shifts from after-the-fact review to whether the operator can be guided during the recording, so an unusable scan never enters the clinical queue.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Reading fleet and signal-quality data from each BC-1 in the field

    Devices are deployed across multiple facilities in Kenya with plans to scale to a 2,000-site fleet by 2030. Device telemetry, signal-quality indicators and calibration state are not aggregated into one system.

    A fleet telemetry layer that pulls battery, usage and signal-quality data from each BC-1 lets calibration drift and downtime precursors be caught at the fleet level instead of in a single case report.

    • BrainCapture 2025 company profile (PitchBook / Tracxn, January 2026)
    • Conformance product entry — BrainCapture BC-1 (Consonance, 2024)

  2. Building the data pipeline behind EEG AI interpretation

    Automated EEG interpretation is described as a 'planned enhancement'. The pipeline that ingests raw recordings, labels them and retrains models on data from Kenya, the Philippines and Indonesia is not yet a production operation.

    An MLOps pipeline (a machine-learning operations workflow that handles data, training and deployment continuously) that handles ingestion, labelling and retraining on the populations the device actually serves makes model behaviour a measurable property of the deployment rather than a one-off result.

    • MIT Solve — BrainCapture BC-1 solution profile (2024)
    • Wolfpack Digital — BrainCapture project page (2024)

  3. Connecting BrainCapture's cloud to hospital record systems

    EEG recordings reach BrainCapture's cloud but often stay there, separated from the Electronic Medical Record (EMR) and Hospital Information System (HIS) used by the treating clinician.

    A standard HL7 FHIR (a healthcare data-exchange standard maintained by the standards body Health Level Seven) integration layer lets the cloud load its results into the EMR the hospital already runs, which makes the EEG report part of the patient's record.

    • Wolfpack Digital — BrainCapture project page (2024)
    • BrainCapture 2025 company profile (Tracxn)

  4. Guiding electrode placement during the recording

    Recording quality depends on consistent electrode placement, and operators are non-specialists. Without real-time feedback during acquisition, low-quality recordings can reach the reading queue.

    An on-device quality-gate that flags electrode impedance and noise during the recording lets the operator adjust before the scan is committed, so the clinical-review queue contains recordings that meet a documented quality threshold.

    • Conformance product entry — BrainCapture BC-1 (Consonance, 2024)
    • BrainCapture Academy — About (2024)

  5. Producing regulatory evidence as a by-product of the devices

    Post-Market Clinical Follow-up under MDR, the FDA 510(k) submission and data-residency requirements across the 35 target countries each need their own evidence documents. Manual assembly of that evidence is a recurring load on the clinical and quality team.

    A regulatory evidence pipeline that pulls Post-Market Clinical Follow-up and clearance-traceability material directly from device and clinical logs lets each quarter's filings be assembled from existing records.

    • Health Tech Hub Copenhagen — BrainCapture CE Mark announcement (2024)
    • BrainCapture 2024 Annual Update (MIT Solve profile)

What we'd propose

  • Digital CDMO

    Fleet telemetry and predictive maintenance for the BC-1

    A telemetry layer that streams battery, usage and signal-quality data from each BC-1 device to a central model, supports predictive maintenance and feeds the usage records the EEG-as-a-Service billing runs on.

    • Device telemetry acquisition

      Data off the device

      Connect the BC-1 amplifier and the smartphone client so battery state, recording count, signal-quality indicators and firmware version are published to a central time-series store, using either OPC UA (Open Platform Communications Unified Architecture) adapters or MQTT (a lightweight message protocol common on factory and IoT devices) where the device firmware supports it.

    • Predictive maintenance view

      Fleet before device

      Build a fleet-level view that highlights devices drifting on calibration or running down on battery across all 2,000 target sites, so the next maintenance trip is chosen by the data rather than by the calendar.

    • Usage-based billing feed

      From scan to invoice

      Forward confirmed scan events into a billing interface so the EEG-as-a-Service revenue model is read off the device rather than reconstructed from distributor reports.

    • Decisions about devices follow readings from the field rather than quarterly stock counts.
    • Each scan event becomes the trigger for a billing event, which is what the EEG-as-a-Service model rests on.
    • Calibration drift and battery health are caught at the fleet level instead of in a single case report.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    MLOps pipeline for EEG interpretation models

    A machine-learning operations pipeline that takes raw EEG recordings from each deployment region, labels them, trains and re-trains the interpretation models and validates them against documented accuracy thresholds before they ship.

    • Recording ingestion with provenance

      Recordings in, lineage kept

      Take raw EEG recordings from the cloud store together with site, demographic and operator metadata, so each training set is traceable back to the patient population it came from.

    • Label-assisted training workflow

      Specialist time used well

      Use active learning (a technique that picks the records the model is least sure about) to route the recordings the model is least sure about to a neurologist for labelling, so specialist time is spent on the cases that change the model the most.

    • Continuous retraining and validation

      Models renewed, not replaced

      Run model retraining on a defined cadence and validate the new model against a held-out test set covering each target demographic, so the move from old model to new model is a documented step rather than a side effect.

    • AI behaviour against each demographic is something the team can read, not something they trust by reputation.
    • Specialist labelling time is spent on the cases that change the model the most.
    • New recordings from new regions arrive as new training data rather than as exceptions to the existing model.

    We've done this before

    See how we do Digital Lab
  • Enterprise AI

    HL7 FHIR integration into hospital electronic medical records

    An HL7 FHIR (Health Level Seven Fast Healthcare Interoperability Resources, the healthcare data-exchange standard) integration layer that publishes BrainCapture's EEG reports into the Electronic Medical Record (EMR) and Hospital Information System (HIS) the hospital already operates, so results reach the treating clinician in the system they work in.

    • FHIR resource mapping

      Reports in the hospital's vocabulary

      Define the EEG report, observation and encounter as FHIR resources (the structured building blocks of the HL7 FHIR standard), so the report can be received into target EMR systems without per-vendor middleware.

    • EMR connection adapters

      Hospitals connected, not integrated

      Build per-hospital adapters for the EMR systems in use across Kenya, the Philippines and Indonesia, tested against each vendor's FHIR sandbox (a vendor-supplied test environment that mimics the live system).

    • Result delivery and audit trail

      A receipt the audit can read

      Record each delivery of an EEG report to a patient record with site, recipient and timestamp, so clinicians and auditors can read who received which report and when.

    • EEG results appear in the system the treating clinician already works in, not in a separate portal.
    • Each delivery of a result to an EMR is logged with who received it and when.
    • Adding a new hospital site is a configuration step rather than a project.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    On-device recording-quality guidance for non-expert operators

    A real-time recording-quality gate that runs on the BC-1 smartphone companion, watching electrode impedance, motion and ambient noise during the scan, and gives the operator a clear instruction while the recording is still in progress.

    • Electrode and signal-quality model

      Quality measured as it happens

      Run a compact, on-device model on the live EEG stream that estimates per-electrode impedance and overall signal quality in under a second, so the operator gets a readout before the recording is committed.

    • Operator-facing guidance

      Operator guided, not graded

      Show the operator which electrode to check and which kind of noise to look for, in the language of the clinical site, so a nurse in Mombasa or Manila sees the same quality gate as a nurse in Copenhagen.

    • Quality threshold at recording commit

      Unusable scans do not commit

      Set the threshold at which a recording is allowed to commit to the cloud, recorded against a documented criterion, so that low-quality recordings are caught at acquisition and not after reading.

    • Recording quality is checked at the moment of recording instead of during the next clinical read.
    • Operators are guided, not graded — the gate is part of the workflow, not a separate review.
    • A documented quality threshold makes the same standard measurable at every site.

    We've done this before

    See how we do Digital Lab
  • Digital Lab

    Regulatory evidence pipeline for MDR, FDA and data residency

    An evidence pipeline that takes Post-Market Clinical Follow-up inputs, design-history records and per-jurisdiction data-residency evidence directly from device and clinical systems, and assembles the submission packages each regulator requires.

    • PMCF evidence collection

      Real-world evidence as a by-product

      Pull the safety and performance inputs for Post-Market Clinical Follow-up from device logs and clinical outcomes, with each number carrying the source record it came from, so the periodic report is assembled rather than reconstructed.

    • Data-residency routing

      Records stay where the rule says

      Route device data to the storage location each jurisdiction's residency rule requires, with the rule and the routing recorded alongside each record for later audit.

    • Submission package assembly

      Filings assembled, not typed

      Compose the submission package for each regulator from the evidence pipeline's records, so the same source data serves the MDR package and the FDA 510(k) package without a separate documentation round.

    • Quarterly regulatory work runs off the existing record set rather than off a fresh compilation.
    • The same evidence serves the MDR package and the FDA submission, which shortens parallel timelines.
    • Data-residency rules are stated, routings are recorded and audits read the record, not the wiring.

    We've done this before

    See how we do Digital Lab

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what BrainCapture's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
IT/OT convergence 35 → 85
Devices stream data over Bluetooth 5.0 to a smartphone and on to the cloud. The fleet view that joins devices across the 2,000 target sites is not yet in place, which the roadmap to 35 countries will require.
Data architecture 40 → 90
Recordings reach the cloud and demographic metadata is captured, but the layer that joins raw EEG, clinical outcomes and per-jurisdiction data-residency routing is still being built.
AI/ML maturity 25 → 80
Automated EEG interpretation is described in public communications as a planned enhancement. The maturity shift is from producing a model once to a continuous-training operation across diverse populations.
Regulatory automation 30 → 75
MDR is held and FDA 510(k) is in progress. The work that has to be repeated every quarter across jurisdictions — Post-Market Clinical Follow-up and data-residency reporting — is still assembled by hand.
Quality assurance 45 → 85
A documented quality threshold at the moment of recording, and a quality gate that runs on the operator's device during acquisition, are the steps between nurse-led acquisition and nurse-led, evidence-graded acquisition.
Supply chain visibility 30 → 70
Manufacturing is concentrated in Poland and field operations run across Africa and Southeast Asia through distributor partners. Joining field readings to distributor stock to replenishment is what prevents a kit failing for want of a cap.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of January 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with BrainCapture, and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].