Back to results
US 10,440,939 B2Granted
Device and method of correlating rodent vocalizations with rodent behavior
AssigneeVium, Inc.
InventorsJonathan Betts-Lacroix, Laura Schaevitz
Life of this patent
Estimating the term…
Abstract
The field of this invention is classifying animal behaviors. In particular the fields of this invention include using animals in vivariums, such as rodents, particularly mice. Ultrasonic vocalizations are compared to times-shifted video behaviors of the same mice to identify vocalizations that occur prior to, during, and after particular video behaviors. The comparisons generate correlations weights that may then be used to apply known classifications of the video behaviors to the vocalizations.
Claims8
A device for automatically correlating animal vocalizations with animal behaviors comprising:
- a study animal type, wherein the study animal type is non-human;
- a vivarium comprising a first study animal in a vivarium home cage;
- at least one ultrasonic audio sensor proximal to the vivarium home cage, adapted to receive from the vivarium home cage audio vocalizations and then communicate the audio vocalizations to a first, second and third processor;
- at least one video camera proximal to the vivarium home cage, adapted to detect and communicate one or more video behaviors of the first study animal;
- the first processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a first audio vocalization, of the audio vocalizations, occurs prior to a first video behavior, of the one or more video behaviors;
- the second processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a second audio vocalization, of the audio vocalizations, occurs overlapped in time to a second video behavior of the one or more video behaviors;
- the third processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a third audio vocalization, of the audio vocalizations, occurs after to a third video behavior of the one or more video behaviors;
- wherein outputs of the first, second and third processors comprise correlation weights of respective comparisons of the first, second, and third processors;
- and
- wherein the first, second and third processor may be a same processor or two processors;
- wherein the correlation weights are selected to optimize an expected probability of a valid association of an audio vocalization with the one more video behaviors to which it is compared.
8 more elements
A method of classifying study animal behaviors using a device comprising:
- a study animal type, wherein the study animal type is non-human;
- a vivarium comprising a first study animal in a vivarium home cage;
- at least one ultrasonic audio sensor proximal to the vivarium home cage, adapted to receive from the vivarium home cage audio vocalizations and then communicate the audio vocalizations to a first, second and third processor;
- at least one video camera proximal to the vivarium home cage, adapted to detect and communicate one or more video behaviors of the first study animal;
- the first processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a first audio vocalization, of the audio vocalizations, occurs prior to a first video behavior, of the one or more video behaviors;
- the second processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a second audio vocalization, of the audio vocalizations, occurs overlapped in time to a second video behavior of the one or more video behaviors;
- the third processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a third audio vocalization, of the audio vocalizations, occurs after to a third video behavior of the one or more video behaviors;
- wherein outputs of the first, second and third processors comprise correlation weights of respective comparisons of the first, second, and third processors;
- and
- wherein the first, second and third processor may be a same processor or two processors;
- wherein the correlation weights are selected to optimize an expected probability of a valid association of an audio vocalization with the one more video behaviors to which it is compared.
8 more elements
A system of classifying study animal behaviors using a device comprising:
- a study animal type, wherein the study animal type is non-human;
- a vivarium comprising a first study animal in a vivarium home cage;
- at least one ultrasonic audio sensor proximal to the vivarium home cage, adapted to receive from the vivarium home cage audio vocalizations and then communicate the audio vocalizations to a first, second and third processor;
- at least one video camera proximal to the vivarium home cage, adapted to detect and communicate one or more video behaviors of the first study animal;
- the first processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a first audio vocalization, of the audio vocalizations, occurs prior to a first video behavior, of the one or more video behaviors;
- the second processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a second audio vocalization, of the audio vocalizations, occurs overlapped in time to a second video behavior of the one or more video behaviors;
- the third processor adapted to compare the audio vocalizations with the one or more video behaviors, wherein a third audio vocalization, of the audio vocalizations, occurs after to a third video behavior of the one or more video behaviors;
- wherein outputs of the first, second and third processors comprise correlation weights of respective comparisons of the first, second, and third processors;
- and
- wherein the first, second and third processor may be a same processor or two processors;
- wherein the correlation weights are selected to optimize an expected probability of a valid association of an audio vocalization with the one more video behaviors to which it is compared.
8 more elements
Dependencies are inferred from the claim text.