Artificial intelligence · Precision health

Human health,
in full resolution.

Powered byMount Sinai

We build computational representations of human state from the clinical record, video, voice, and movement to make disease more measurable.

Explore our approach Explore our publications
See the problem with conventional biomedical AI

Biomedical AI is only as rich as the data it receives.

Today, much of that data is a compressed proxy: a billing code, an ordinal score, or a few words in a note. Once human variation is discarded, no model can recover it.

Diagnosis code
ICD-10 F32.2 Major depressive disorder
Ordinal scale
MDS-UPDRS III: 24 A single total score
Clinical note
“Flat affect” Two words stand in for behavior
What is left behind?
This prototype requires WebGL, available in current Safari, Chrome, Firefox, and Edge browsers.
Human state
0108

A different starting point

People are high-dimensional.

Every interaction contains dynamic signals of motor, cognitive, affective, linguistic, and physiological state.

Research Areas

A compact view of the lab’s methods, modeling priorities, and paths from richer data to better care.

Highlighted Research

01 / 07
Mental health audiovisual AI research collage
Audiovisual AI · Mental health

Measuring mental state through clinical interaction

Selected Publications

View all publications

The Lab

Collaborate

Interested in
building with us?

Glicksberg Lab

Human health, in full resolution.

Research profiles

PubMedGoogle ScholarMount Sinai Scholars

Address

3 East 101st Street
New York, NY 10029

Explore

Research areasProjectsPublicationsNews