任务(项目管理)
计算机科学
口语流利性测试
判决
语义学(计算机科学)
认知
语音识别
痴呆
图层(电子)
人工智能
疾病
自然语言处理
流利
心理学
医学
神经心理学
神经科学
化学
数学教育
管理
有机化学
病理
经济
程序设计语言
作者
Zigao Peng,Jiali Yang,Nuo Lei,Yang Jing-yu,Mingying Lan,Li Gao
标识
DOI:10.1109/iscsic57216.2022.00028
摘要
Research related to automatic detection of Alzheimer’s disease (AD) is important in which progression can be prevented or delayed by early diagnosis because it can help simplify the diagnosis method. Since Alzheimer's disease is known to alter the semantics and acoustics of spontaneous speech, linguistic biomarkers offer hope for non-invasive diagnosis of the disease. Nevertheless, most of the earlier studies only used one language task. Consequently, their findings may not generalize to other areas of cognitive ability. In this paper, we propose a two-layer model combining data from three connected speech tasks: picture description (PD), semantic fluency (SF) and sentence repetition (SR). In the first stage, the PD task is initially used to detect dementia. In the second stage, the SF and SR task are used to classify healthy controls (HC) and mild cognitive impairment (MCI). We also compare the performance of the proposed model with a widely used one-layer model using a pre-fusion approach, and this study shows that the proposed model performs better in the AD detection.
科研通智能强力驱动
Strongly Powered by AbleSci AI