How the evidence is built
Readings are evaluated stage by stage against the DiMe V3 framework, and prospective collections extend the evidence across sites and countries.
Evidence organised by DiMe V3
The Digital Medicine Society's V3 framework names three stages a digital measure has to clear: verification, analytical validation and clinical validation. Each Cephalgo result sits in one of them, and is documented in a Cephalgo report or poster.
Verification
Does the capture chain record what it should?
How Cephalgo addresses itQuality check at capture, version-controlled task scripts, and every file carrying its device, OS, app and analysis versions.
Analytical validation
Is the measurement precise and reproducible, including when the capture chain changes?
How Cephalgo addresses itRepeated within-participant capture, and a fixed analysis configuration evaluated across varied capture conditions.
Clinical validation
Does the reading correspond to the clinical reference the study already uses?
How Cephalgo addresses itReadings evaluated against the clinical reference scores on held-out participants, one indication at a time.
TRIPOD+AI structures how we report prediction models: cohort and splits, comparison group, and how each model was developed and evaluated.
Posters & reportsCollecting now
Speech is being collected now in Alzheimer's, MCI, BPSD, Parkinson's and Huntington's, on the participant's own devices.
Alzheimer's and MCI
BPSD
Parkinson's disease
Huntington's disease
Where would speech sit in your next protocol?
We work with pharmaceutical and clinical research teams on recordings a study already holds, endpoint integration and prospective studies.
Get in touch →