Eden Health was started by a practising psychiatrist because of a daily frustration: in most of medicine, decisions lean on measurement; in psychiatry, they lean almost entirely on conversation and judgement. We build the measurement, and we hold it to publication-grade evidence.
A cardiologist reads an ECG. An endocrinologist reads a lab panel. A psychiatrist assessing how a patient is doing has an interview, a rating scale and experience. These are real tools, but they lean on self-report and judgement, applied at single moments in time.
Between appointments, and after a patient leaves the hospital, there is usually no objective measurement at all. The states that matter most in mental health are dynamic, and today they change largely unobserved.
Our scientific programme targets exactly this gap: objective, repeatable correlates of clinically meaningful states, measured with the devices people already carry and wear, and interpreted by clinicians, not algorithms.
Every project we run starts from a clinical question, is designed with the clinicians who will use the answer, and ends with the decision where it belongs: in professional hands.
Startups pitch. We publish. The science we stand on is in the open record, where anyone can check it.
Our team conceived, designed and co-authored the proof-of-concept study in which a multi-scale convolutional neural network analysing mel-spectrograms of patient speech distinguished inpatients admitted after a recent suicide attempt from patients with severe depression without documented suicidal ideation, in an 88-patient, two-hospital study with strict patient-level cross-validation.
The congress companion to our proof of concept: an application concept for monitoring vulnerable patients, presented to the European psychiatric community and published as a congress abstract in European Psychiatry.
Our research runs with hospitals that treat patients every day and with academic collaborators who have spent careers on these questions.
Psychiatric researcher with an extensive publication record in suicide research, and scientific collaborator on our clinical methodology.
Professor and researcher in psychiatry and sleep medicine, collaborating on the clinical design of our measurement research.
Romania's largest psychiatric hospital and a site of our published proof-of-concept study, with an established voice-research workflow.
Clinical partner and second site of the proof-of-concept study.
The wider network behind our published work spans a university of medicine, a technical university, a national microelectronics institute and partner companies. It is the same span of disciplines we carry in-house. We also pursue European collaborative research programmes with academic, clinical and industry partners across the continent.
A note on how we talk about this work. Our research touches suicide prevention, a field where words matter. We follow safe-reporting practice, avoid sensational claims, and state plainly what our systems do (detect current states for clinicians) and what they do not (predict, decide, or replace care).