Digital characterization of movement to detect and monitor disorders

Inventors

Demiralp, Emre

Assignees

Verily Life Sciences LLC

Abstract

Introduced here are techniques for digitally characterizing the movement of a subject in order to detect the presence of a disorder or monitor the progression of the disorder. More specifically, one or more angular features can be identified that define how certain part(s) of the human body move relative to other part(s) of the human body. These angular feature(s) can be used, for example, to affirmatively diagnose instances of a disorder, eliminate a disorder as the source of symptoms experienced by a subject, generate confidence scores that can be used to assist in diagnosing a subject, monitor disorder progression due to treatment or lack thereof, etc.

Core Innovation

The invention digitally characterizes human movement by deriving angular features from motion data captured from anatomical locations of a person performing an activity over a period of time. An angular feature is defined by a spatial relationship between a first anatomical location, a second anatomical location, and a third anatomical location of the person, or by spatial relationships between multiple anatomical locations, and each value in a computed first set of values corresponds to a different time within the period of time.

The invention establishes either an affirmative diagnosis or a negative diagnosis for the person with respect to a disorder by computing statistical similarity between the first set of values and a second set of values. The second set of values is representative of one or more performances of the activity by one or more persons who are known to represent confirmed cases of the disorder, excluding the person for whom the diagnosis is established. The diagnosis is based on whether the statistical similarity exceeds a threshold.

The invention documents diagnostic outputs in a digital profile associated with the individual, including storing the first set of values and an indication of a diagnosis. It also includes producing a metric indicative of similarity and using the similarity metric to determine whether to affirmatively diagnose an individual as having a disorder. In addition, the invention can identify an appropriate program for treatment of the disorder and can estimate progression of the disorder by comparing first values to prior digital profile values.

Claims Coverage

The independent claims (three) cover angular-feature computation from sensor or visual tracking data, and a diagnosis decision based on statistical similarity to values representative of confirmed cases, with results stored in a digital profile. Across the independent claims, the inventive features are: angular feature value computation, statistical similarity comparison to a reference set from confirmed cases with thresholding, and diagnosis documentation in a digital profile; one independent claim additionally emphasizes wireless inertial-sensor acquisition and notification to a medical professional, and another emphasizes visually tracked movement with treatment program identification.

The independent claims collectively cover computing time-indexed or visually tracked angular-feature values from movement data and establishing disorder diagnosis decisions by comparing statistical similarity (or a similarity-indicative metric) to values representative of confirmed cases and applying a threshold. The claims also require recording diagnosis-related outputs in a digital profile associated with the individual, and one independent claim further includes identifying an appropriate program for treatment.

Stated Advantages

Documented Applications