Speech analysis devices and methods for identifying migraine attacks

Inventors

Berisha, Visar • PEPLINSKI, Jacob • SCHWEDT, Todd

Assignees

Linus Health Inc • Arizona State University Downtown Phoenix campus • Mayo Clinic in Florida

Abstract

Speech analysis devices and methods for identifying migraine attacks are provided. Migraine sufferers can experience changes in speech patterns both during a migraine attack and in a pre-attack phase (e.g., a time period before the migraine attack can be recognized by the migraine sufferer). Embodiments identify or predict migraine attacks during the pre-attack phase and/or the attack phase (such as early stages of a migraine attack) by comparing speech features from one or more speech samples provided by a user against baseline data. The speech features are indicative and/or predictive of migraine onset, and can be personalized to a user and/or based on normative data.

Core Innovation

A migraine identification system uses a mobile speech analysis approach in which speech samples are received as an input signal indicative of speech provided by a user. Speech samples are converted into an instantaneous multidimensional statistical signature of speech production abilities across perceptual dimensions, and the system extracts speech features that reflect speaking rate, articulation, prosody-related variability, and phonatory-related characteristics.

The instantaneous multidimensional statistical signature is compared against one or more baseline statistical signatures of speech production ability derived or obtained from the user. Baseline statistical signatures may be derived from normative acoustic data stored in a database, and baseline selection or adjustment can optionally use additional data. The comparison is used to assess changes in speech patterns associated with a migraine and to produce a migraine identification signal based on the multidimensional statistical signature comparison.

The system provides user notifications based on the migraine identification signal via a notification element. A longitudinal observational study described in the patent collects speech samples across interictal, pre-attack (0–12h), and attack phases and shows statistically significant differences in speech features, with the most consistent changes including speaking rate, articulation precision/entropy, pause rate, and phonatory duration.

Claims Coverage

The independent claim set covers three main claim classes that share the same inventive sequence: receiving user speech as an input signal, extracting an instantaneous or extracted multi-dimensional statistical signature of speech production abilities, comparing the signature to one or more baseline statistical signatures to assess changes associated with migraine, and providing a migraine identification signal, with device claims adding notification. Across the independent families, the inventive features focus on multi-dimensional statistical signatures and baseline comparisons for migraine identification.

Across the independent device, method, and medium claims, the core coverage is the extraction or generation of a multi-dimensional statistical signature of speech production abilities from user speech, comparison against one or more baseline statistical signatures to assess change associated with migraine, and output of a migraine identification signal, with the device claim further requiring a notification element to present notifications to the user.

Stated Advantages

Documented Applications