Brain-health AI and applied-AI glossary

Plain-language definitions of the terms we use in neuroanalytics, clinical workflow automation and predictive AI — written for clinicians, public-sector buyers and partners, not just engineers.

Care coordinationOrganizing a patient's care across providers, programs and sites so that assessments, follow-ups and handoffs happen on time.
Clinician in the loopA design principle in which AI assists, drafts and flags, but a qualified clinician reviews and makes every clinical decision.
De-identificationRemoving or transforming personal identifiers in health data so that individuals cannot reasonably be identified, enabling research use.
Digital biomarkerAn objective, measurable feature derived from digital data — sensors, devices, software — that reflects a biological or clinical state.
Digital phenotypingCharacterizing a person's health and behaviour from data generated by everyday digital devices, such as phones and wearables.
Federated learningA way of training a machine-learning model across several organizations without moving their raw data to a central place.
Freezing of gaitA sudden, brief inability to start or continue walking despite the intention to move, common in advanced Parkinson's disease.
Innovative Solutions Canada — Testing StreamA Government of Canada program that procures late-stage Canadian innovations (TRL 7–9) and tests them in real federal operational environments.
Longitudinal dataRepeated measurements of the same individuals over time, as opposed to a single snapshot.
Mild cognitive impairment (MCI)A measurable decline in memory or thinking that is greater than expected for age but not severe enough to be dementia.
NeuroanalyticsThe use of data analysis and machine learning on clinical, behavioural and sensor signals to measure and predict brain health over time.
Parkinson's disease subtypesGroups of people with Parkinson's disease who share a pattern of symptoms, progression rate or underlying biology, used to tailor care and research.
Predictive maintenanceUsing data and machine learning to forecast when equipment is likely to fail, so maintenance happens before the failure instead of after it or on a fixed schedule.
Reassessment trackingAutomatically computing and monitoring when each patient is due for a scheduled clinical reassessment, and who is responsible.
Technology Readiness Level (TRL)A 1-to-9 scale that describes how mature a technology is, from basic principles (TRL 1) to a system proven in real operations (TRL 9).