Method for predicting treatment classes using animal behavior informatics
View Patent ↗A system and method used to assess animal behavior includes a module having sensors that collects a variety of physical and biological data from a test subject. Interpretation of the data is provided to assess the test subjects behavior, neurology, biochemistry and physiology. The module is useful in observing the effects of a drug on the test animal and providing information on the drug's signature. Another advantage is the module's portability that allows it to be used in standard laboratory cages. This portability allows the animal to be tested in its own habitat, that can reduce any erroneous data due to stressing the animal when removed to a test cage. Additionally, the module's design allows for parallel data collection and interpretation from several laboratory animals undergoing different experiments. Multi-dimensional modeling of the test subject based the system's interpretation of the data allows pattern recognition of the drug signature, and predictive drug analysis.
1. A method of generating an electronic behavioral database of a plurality of stored treatment class signatures wherein the signatures are characteristic of at least one treatment class comprising
(i) collecting behavioral data for a plurality of known treatments from animals administered one or more of said known treatments and wherein said known treatments are members of at least one class of treatments;
(ii) using a computer to define said signatures for said treatment classes by segregating said behavioral data and by correlating features of said behavioral data with said known treatment class members; and
(iii) storing said signature information in a computer.
2. The method of claim 1 wherein said features are pre-determined.
3. The method of claim 1 wherein said features are not pre-determined.
4. The method of claim 1 wherein the computer defines said signatures by segregating the behavioral data based on supervised learning.
5. The method of claim 1 wherein the computer defines said signatures by segregating the behavioral data in part based on supervised learning, and in part based on unsupervised learning.
6. The method of claim 1 wherein said features comprise transitions among behaviors.
7. The method of claim 1 wherein said features comprise a temporal structure of behavior.
8. The method of claim 1 wherein the database is continuously updated with behavioral data from additional treatments.
9. The method of claim 1 wherein the database is continuously updated with additional behavioral data from treatments already stored in the database.
10. A method of generating an electronic behavioral database of a plurality of stored treatment class signatures wherein the signatures are characteristic of at least one treatment class comprising
(i) collecting behavioral data for a plurality of unknown treatments from animals administered one or more of said unknown treatments and wherein said unknown treatments are members of at least one class of treatments;
(ii) using a computer to define said signatures for said unknown treatments by segregating said behavioral data and by correlating features of said behavioral data with said unknown treatment class members; and
(iii) storing said signature information in a computer.
11. The method of claim 10 wherein said features are pre-determined.
12. The method of claim 10 wherein said features are not pre-determined.
13. The method of claim 10 wherein the computer defines said signatures by segregating the behavioral data based on supervised learning.
14. The method of claim 10 wherein the computer defines said signatures by segregating the behavioral data in part based on supervised learning, and in part based on unsupervised learning.
15. The method of claim 10 wherein said features comprise transitions among behaviors.
16. The method of claim 10 wherein said features comprise a temporal structure of behavior.
17. The method of claim 10 wherein the database is continuously updated with behavioral data from additional treatments.
18. The method of claim 10 wherein the database is continuously updated with additional behavioral data from treatments already stored in the database.
19. An electronic behavioral database of a plurality of stored treatment class signatures wherein the signatures are characteristic of at least one treatment class where said behavioral database is constructed by the steps comprising
(i) collecting behavioral data for a plurality of known treatments from animals administered one or more of said known treatments and wherein said known treatments are members of at least one class of treatments;
(ii) using a computer to define said signatures for said treatment classes by segregating said behavioral data and by correlating features of said behavioral data with said known treatment class members; and
(iii) storing said signature information in a computer.
20. An electronic behavioral database of a plurality of stored treatment class signatures wherein the signatures are characteristic of at least one treatment class where said behavioral database is constructed by the steps comprising
(i) collecting behavioral data for a plurality of unknown treatments from animals administered one or more of said unknown treatments and wherein said unknown treatments are members of at least one class of treatments;
(ii) using a computer to define said signatures for said unknown treatments by segregating said behavioral data and by correlating features of said behavioral data with said unknown treatment class members;
(iii) storing said signature information in a computer; and
(iv) outputting a result from (ii) that is indicative of at least one treatment class.
21. A system for identifying treatment class signatures based on behavior, said system comprising
(a) a computer for processing behavioral data wherein said data is obtained by collecting behavioral data for a plurality of known treatments from animals administered one or more of said known treatments and wherein said known treatments are members of at least one class of treatments; and
(b) a database of treatment class signatures wherein said signatures are obtained by said computer defining said signatures for said treatment classes by segregating said behavioral data and by correlating features of said behavioral data with said known treatment class members.