Clinical Intelligence OS

Connect the whole patient picture.

Phase 1 brings qEEG reports and approved Neuromodulation Studio templates into a clinician-reviewed workflow, supported by patients, scheduling, inbox and evidence.

Verified clinic beta. Each clinical report remains subject to its readiness, consent and release gates. The wider product map below includes preview capabilities.

Decision support—not autonomous diagnosis or treatment.

LIVING CLINICAL INTELLIGENCE

Mental Health &
Neuromodulation NERVE

Many signals. A connected clinical picture.

Current focus Future concepts
Zoom into Twins, analyzers and individual data fields

Explore the connections behind clinical decisions.

Illustrative architecture. Availability varies. Clinical review remains essential.

Evidence workflows connect to PubMedClinicalTrials.govCochrane Library Sources do not imply endorsement.
The evidence database

Ask what the neuromodulation research says.

Ask in plain language and get answers from 186,143 indexed papers, with every study linked to its source. Classified by Jev, our research AI, and always labelled as advisory.

Where the evidence is

papers per treatment × condition
TMStDCSDBSVNSSCS / DRGNeurofeedbackDepression3,3201,2321,043Parkinson’s disease9,164Epilepsy1,4572,111137OCD4911,171Essential tremor1,415Chronic pain1,160Stroke rehabilitation322ADHD376

Papers tagged to each curated protocol indication in the v6 corpus, captured 2026-09-25. Counts describe how much has been published, not how well a treatment works. A dot means no tagged protocol yet, not “no evidence”. Select a number to open its evidence map.

Indexed evidence corpus

Evidence, with its source in view.

Counts from the DeepSynaps evidence database, with the corpus date shown alongside them.

Backend snapshot
186,133Indexed papersRecords in the evidence corpus
2,297Clinical trial recordsIndexed ClinicalTrials.gov records
348FDA device recordsRegulatory records in the corpus
84,347Papers with abstracts45.3% of indexed papers

Snapshot totals are shown until the backend can be verified.

Explore the original product snapshot Product snapshot · 2026-04-24

Explore the product snapshot

These filters explore the bundled modality and condition matrix. They do not query the backend corpus.

Original snapshot totals and scope

These original product-snapshot totals use different scopes and have not been independently reconciled. The year series, modality totals and matrix are separate views; filters do not change these headline counts. They are not verified live corpus totals.

184,669Papers
1,409Clinical trials
3,920Meta-analyses
53Conditions
13Modality groups
12Protocol templates
13Devices
10Evidence sources

Showing all bundled registry evidence dimensions.

A.Modality snapshotSeparate modality subset
B.Condition × modality snapshot8 selected conditions · snapshot grades

Scroll sideways to see all five modalities →

rTMStDCStACStRNSNFMajor DepressiveTreatment-Resistant DepressionObsessive-CompulsivePost-Traumatic StressAnxietyChronic PainADHD (Adult)Alzheimer's
LowerHigherNo benefit signal
C.Snapshot over timePapers · 2020–2025
010K19K 2020: 8,400 papers20202021: 10,200 papers20212022: 12,800 papers20222023: 15,600 papers20232024: 19,400 papers20242025: 12,200 papers2025
Papers in bundled year distribution
Evidence inspectionRegistry matrix entry
Condition
Major Depressive Disorder
Modality
rTMS
Evidence grade
A
Studies
50
Effect context
0.35-0.55
Regulatory context
approved
Evidence note
95% dose ~34,773 pulses
Source provenance
Study-level citations are not attached to this snapshot entry.
Source directoryOpen the original research services

Directory links open each service. They are not citations for the selected matrix entry. The product snapshot lists 10 source families: PubMed, OpenAlex, Cochrane, ClinicalTrials.gov, EMBASE, Scopus, PsycINFO, IEEE Xplore, bioRxiv, medRxiv.

Selected published research · separate from registry totals

Go directly to the research.

Two independent publications relevant to the field. They are examples for exploration, not clinical validation of DeepSynaps or sources for the snapshot counts above.

Neuromodulation · 2007

Efficacy and safety of transcranial magnetic stimulation in the acute treatment of major depression: a multisite randomized controlled trial

O’Reardon et al. · Biological Psychiatry
Randomised trial

An original study comparing active TMS with a sham procedure. Open the publication to inspect the design, measures and limitations.

EEG & computational research · 2020

An electroencephalographic signature predicts antidepressant response in major depression

Wu et al. · Nature Biotechnology
EEG model development

Research using resting EEG and antidepressant trial data to investigate response patterns. This is a published research model, separate from DeepSynaps.

Connected modalitiesqEEGMRI / fMRILabsWearablesGenomicsClinical notesOutcomes

Snapshot grades and regulatory notes are historical product data, not current clinical guidance. Consult the original research and current guidance with the responsible clinician.

Evidence, not vanity numbers

A product map you can inspect.

Counts come directly from the public product and evidence registries. Surface availability can vary by deployment.

25
registered analyzer surfaces
22
registry surfaces marked available; not beta approvals
3
surfaces pending infrastructure
22
distinct analyzer modalities
4
connected clinical studios
3
publicly linked evidence sources
The operating model

One connected workflow. Three clear outcomes.

01

See the full picture

Bring brain, behaviour, physiology, labs, medication and outcomes into one reviewable context.

02

Explore possible paths

Use DeepTwin to compare governed scenarios while keeping assumptions and uncertainty visible.

03

Learn safely over time

Use NERVE to connect evidence, provenance, outcomes and clinician review.

See it work

From signal to governed action.

An interactive overview built from the workflows already represented in the product registries. No patient data is shown.

Multimodal intelligence

Turn complex inputs into reviewable clinical context.

Quality checks, modality-specific visualizations, longitudinal comparison and explicit governance stay together in one workflow.

01Data quality
02Analyzer result
03Clinician review
Four connected workspaces

Studios built around real clinical work.

Four connected areas for clinical work, with shared patient context and clinician review. Neuromodulation leads the clinical beta; the other Studios remain previews with access by review.

01
Clinical beta · clinician review

Neuromodulation Studio

Draft · Review · Sign off

Plan TMS, tDCS, tACS and neurofeedback workflows with evidence, safety checks and clinician sign-off.

Care focusTMS · tDCS · tACS · neurofeedbackWorkflowDrafts, safety checks and sign-off

Templates and access are reviewed for each clinic.

  • Dashboard
  • Plan
  • Protocols
  • Evidence & Handbooks
  • Records
02
Preview · access by review

Medication Studio

Review · Reconcile · Monitor

Review medication history, interactions, side effects and follow-up in one clinician-led workspace.

Care focusHistory · interactions · follow-upWorkflowMedication review and monitoring

PGx is available only when enabled and consented.

  • Dashboard
  • Plan
  • Evidence
  • Prescriptions
  • Twin
03
Preview · access by review

Mental Health Studio

Assess · Review · Support

Bring assessments, therapy plans, safety checks and progress reviews into one connected workspace.

Care focusAssessments · therapy · progressWorkflowSession planning and review

Medication and neuromodulation link to their Studios.

  • Dashboard
  • Plan
  • Evidence
  • Assessments
  • Twin
04
Preview · access by review

Rehabilitation Studio

Assess · Review · Progress

Connect functional assessments, exercise plans, home programs and session progress.

Care focusMovement · exercises · home programsWorkflowPlans, sessions and progress

Clinical plans remain subject to review and sign-off.

  • Dashboard
  • Plan
  • Evidence
  • Assessments
  • Programs
Explore every signal

See what powers every analysis.

Follow each route from inputs and APIs through analysis methods, evidence resources and the relevant Twin scenario.

25product surfaces4specialist Twins
Product surface available

qEEG Analyzer

qEEG decision-support workstation — band power, asymmetry, coherence, topographic maps, spectral markers, and longitudinal trends. Clinician review required.

⇄
qeeg-analysisREST backend
◫
5 inputsaccepted types
◉
5 viewsvisual outputs
↔
5 linkscross-modal links
Algorithms & analysis methods
MNE spectralSpecParam / Welch powerAge/sex normative z-score
Infrastructure pending

EEG Normative Analyzer

EEG normative analysis.

⇄
neuro-normativeREST backend
◫
6 inputsaccepted types
◉
3 viewsvisual outputs
↔
2 linkscross-modal links
Algorithms & analysis methods
Normative deviation mapsCohort comparisonFeature-level QC
Infrastructure pending

MRI Analyzer

AI-assisted imaging intelligence workstation — region summaries, volumetric trends, structural findings, and image quality. Not a radiology report; no autonomous radiology diagnosis.

⇄
mri-analysisREST backend
◫
4 inputsaccepted types
◉
4 viewsvisual outputs
↔
4 linkscross-modal links
Algorithms & analysis methods
Volumetric segmentationICV-normalised z-scoresMONAI model gated
Infrastructure pending

fMRI Analyzer

AI-assisted resting-state fMRI workstation — network connectivity, sgACC-DLPFC anticorrelation, ROI matrices, motion QC, and functional-neuroimaging sign-off. Requires expert review.

⇄
mri-analysisREST backend
◫
6 inputsaccepted types
◉
4 viewsvisual outputs
↔
4 linkscross-modal links
Algorithms & analysis methods
Resting-state connectivityROI correlation matricesNetwork deviation
Product surface available

fNIRS Analyzer

AI-assisted fNIRS analysis — HbO/HbR trends, channel quality, task activation. Pipeline partially validated; interpret with caution and clinician/research review.

⇄
fnirsREST backend
◫
2 inputsaccepted types
◉
3 viewsvisual outputs
↔
4 linkscross-modal links
Algorithms & analysis methods
HbO / HbR trendsTask activationHomer3 / MNE-NIRS gated
Product surface available

Video / Movement Analyzer

AI-assisted behavioural and motor intelligence workstation — movement markers, affect/attention markers, and behaviour timelines. Privacy and consent governance mandatory.

⇄
movement/analyzerREST backend
◫
3 inputsaccepted types
◉
4 viewsvisual outputs
↔
3 linkscross-modal links
Algorithms & analysis methods
OpenCV featuresGait-cycle kinematicsNormative gait z-score

qEEG reports are in the phase-one clinical beta. Other surfaces are previews or planned capabilities; expand a card for its technical status and access limits.

For developers & builders

Build clinical intelligence into your own product.

Explore integration capabilities for clinical applications, devices and research workflows. Developer access, environments and credentials are reviewed with each organisation during onboarding.

24
declared analyzer API bases
25
analyzer surfaces in the catalog
3/5
FHIR types mapped in the registry
3.1
OpenAPI document version
In onboarding

Read-only public API

Connect reference and evidence data through six read-only public endpoints. Developer keys do not grant access to patient records or analyzer execution.

  • OpenAPI 3.1 document
  • /api/public/v1
  • Six non-PHI read endpoints
In onboarding

Typed TypeScript client

OpenAPI-generated types support a maintained TypeScript wrapper, with CI drift checks. The client is part of the DeepSynaps codebase; package access is reviewed during onboarding.

  • @deepsynaps/api-client/public
  • OpenAPI-generated types
  • CI drift checks
In onboarding

FHIR & OMOP mapping

The clinician-authenticated integration accepts a resources array and returns an OMOP-oriented transform summary. Its resource registry distinguishes mapped types from planned support.

  • Clinician authentication
  • Resource mapping registry
  • Transform summary
In onboarding

Controlled API access

Public developer access is limited by plan entitlement and read scopes. Clinical workflows use separate permissions, consent and review requirements.

  • Plan entitlement
  • Narrow-only read scopes
  • Non-PHI public data
In onboarding

Scoped API credentials

Manage keys for the public reference and evidence API. Secrets are shown once, stored as hashes and support rotation; access depends on your organisation’s enabled entitlement.

  • reference:read · evidence:read
  • Shown once; hash-only storage
  • Scoped key rotation
In onboarding

Test keys & signed webhooks

Test-key credentials and signed HTTPS deliveries support developer events such as key changes and webhook registration. Pings and delivery logs help verify the connection.

  • Developer metadata events
  • HMAC-SHA256 signed deliveries
  • Test ping + delivery log
Public API request formatScoped developer credentials
curl https://api-eu.deepsynaps.com/api/public/v1/meta \
  -H "X-API-Key: $DEEPSYNAPS_API_KEY" \
  -H "Accept: application/json"

Developer keys access read-only reference and evidence services. Patient data, analyzer execution and clinical workflows require separate clinician authentication and permissions.

InteroperabilityFHIR resource coverage
  • PatientMapped
  • EncounterMapped
  • ObservationMapped
  • DevicePlanned
  • DiagnosticReportPlanned

Patient, Encounter and Observation are mapped in the product registry. Device and DiagnosticReport remain planned. Deployment support is confirmed during onboarding.

Digital health buildersAdd multimodal analysis and evidence context to an existing clinical product.
Clinic software teamsConnect DeepSynaps to the EHR, scheduling and records systems already in use.
Research groupsRun governed cohort and modality workflows against your own datasets.
Device manufacturersRoute device and wearable signal into a reviewable clinical context.

API access is reviewed per organisation and covered by a data-processing agreement. No package permits autonomous diagnosis or treatment.

The intelligence layer

See every connection. Keep every guardrail.

Select any node to trace how APIs, databases, evidence, analyzers and four specialist Twin routes converge under clinician control.

DeepTwin scenarios — simulation-only trajectory review with assumptions and uncertainty kept visible.
NERVE governance loop — outcomes are routed through explicit clinician review and sign-off before patient-facing use.
Evidence, safety and provenance

Trust should be inspectable.

01

Evidence stays attached

Sources, grades, assumptions and update state remain visible alongside decision support.

02

Human review is explicit

Clinical outputs are gated for responsible-clinician review rather than presented as autonomous decisions.

03

Provenance is part of the workflow

Inputs, transformations and approvals are designed to remain traceable over time.

Provenance visible Availability disclosed Clinician review required Evidence graded Uncertainty disclosed
What it is
  • Clinical decision support that organizes and interprets data
  • Evidence-linked interpretation with provenance and grades
  • A multimodal, longitudinal view for clinical reasoning
  • Honest about uncertainty, limitations, and missing data
  • Clinician-reviewed: sign-off required before patient-facing conclusions
What it is not
  • Not a source of guaranteed or predicted outcomes
  • Not a treatment-recommendation or prescribing tool
  • Not a whole-brain model or "digital twin" of the patient
  • Not a replacement for clinical judgment

Designed to be questioned: open any finding and trace it back to its measurement, its provenance, and the literature behind it.

Founder story · 2026-05-22

From Oxford's Translational Health Sciences programme to AI innovation in autism care

University of Oxford — Nuffield Department of Primary Care Health Sciences

Read the official Oxford story Text reference only. No University logo or endorsement is implied.
Phase 1 clinical beta

One verified trial. One clinic plan.

Plans and report allowances load from the governed clinic-plan catalog. Plans and credits are agreed with each clinic during onboarding. Clinical governance never changes by plan.

Loading the current launch offer…

No plan enables autonomous diagnosis or treatment.

Bring the intelligence to your clinic

See DeepSynaps with your workflow in mind.

Talk with the team, explore product fit and request reviewed professional access.

WhatsApp Dr. Ali
Early access

Join the beta waiting list.

Be the first to try new analyzers, studio modules, and governance features as we open them. No spam — just product updates.

My Portal