AUDITORY BRAINSTEM RESPONSE (ABR) ANALYSIS AND INTERPRETATION EXPERT SYSTEM

听觉脑干反应（ABR）分析和解释专家系统

• 批准号: 3874836
• 负责人: J M DELEO
• 金额: --
• 依托单位: CENTER FOR INFORMATION TECHNOLOGY
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/3874836
• 关键词: artificial intelligence; auditory stimulus; brain stem; computer assisted medical decision making; computer system design /evaluation; electrophysiology

The objective of this project has been to build an expert system to analyze and interpret the Auditory Brainstem Response (ABR), an electrophysiological response of the human brainstem to acoustic stimuli. Expert reading of the ABR is essential in clinical decisions concerning retrocochlear disorders. We developed a rule-based expert system that performs neurodiagnostic interpretation of the auditory brainstem response. The system was developed on the IBM PC with the commercial system shell GURU and has been operational in the NIH Audiology Clinic for over one year. ABR interpretation strategies differ considerably among various clinics. Our system is a tool for rapidly modelling any ABR interpretive paradigm. Paradigms are easily added, modified, deleted, and selectively engaged at run time. Paradigm rules are boolean logic expressions containing discrete, continuous, and fuzzy variables representing ABR, subject, and stimulus parameters. The system models human ABR expertise in explicit representational form, handles routine cases well, reduce diagnostic errors of omission, and facilitates understanding of ABR clinical decision making. It can be used as a methodology for testing and validating any ABR interpretative strategy, a repository for meaningful ABR interpretative expertise, and a medium for sharing ABR knowledge.

该项目的目标是建立一个专家系统来分析 并解释听觉脑干反应（ABR） 人类脑干对声刺激的电生理反应。 ABR 的专家解读对于相关临床决策至关重要 耳蜗后疾病。 我们开发了一个基于规则的专家系统，可以执行神经诊断 听觉脑干反应的解释。 该系统是 在 IBM PC 上使用商业系统 shell GURU 开发并已 在 NIH 听力学诊所工作了一年多。 不同诊所的 ABR 解释策略差异很大。 我们的系统是一个快速建模任何 ABR 解释范式的工具。 范式可以轻松添加、修改、删除和有选择地参与 运行时间。 范式规则是布尔逻辑表达式，包含 代表 ABR、主题和的离散变量、连续变量和模糊变量 刺激参数。 该系统以明确的方式模拟人类 ABR 专业知识 代表性形式，处理好常规案件，减少诊断错误 遗漏，并有助于理解 ABR 临床决策。 它可以用作测试和验证任何 ABR 的方法 解释策略，有意义的 ABR 解释的存储库 专业知识，以及共享 ABR 知识的媒介。