The goal of this project is to develop a knowledge-based (expert) system which identifies the cause of a patient's gait dysfunction from an analysis of basic gait data. The intent is to assist the clinician by providing the primary and secondary diagnosis, with an explanation of the underlying reasoning process. Specifically this involves: 1) design and development of an automated Data Analysis Expert System (DA/ES), or preprocessor, which will analyze and evaluate the raw data obtained from instruments and yield judgments relating the specific data set to the parameters of normal walking; 2) design and evaluation of a Gait Pathology Expert System (GP/ES), which uses the outputs from the DA/ES and a gait analysis knowledge base to identify the muscular dysfunction responsible for the observed deviations from normal gait; and 3) making preliminary plans for the establishment of a National Resource Center to disseminate the results of this project. Difficulties in walking experienced by cerebral palsy, head trauma, and stroke related to abnormal muscle control. Surgery can significantly improve function if the abnormal muscle activity can be isolated from normal performance. Clinical testing is woefully imprecise; comprehensive gait analysis is a critical factor in making an appropriate surgical plan. Problems in interpreting the data have limited the number of clinical gait laboratories, in spite of the enormous clinical need (in CP alone, the existing population approximates 240,000, with about 6,000 new patients each year). The ultimate goal of developing a National Resource Center to disseminate the results of this project will greatly expand the number of clinical facilities able to provide automated gait interpretation. The strong clinical focus of this effort will be based upon the knowledge of, and verified by, well-known experts in gait analysis. Gait analysis knowledge is uniquely suited for organization into a """"""""frame""""""""-based expert system. The team assembled for this development is well-qualified for work in this area of artificial intelligence. The proposed knowledge-based system is unique because it 1) has a strong clinical basis, 2) automatically interprets the four most significant types of data (foot-switch, EMG, Motion, and Force) to yield the most probable diagnosis of the patient's gait dysfunction, and 3) provides a thorough explanation of its reasoning process.
