A Preliminary Investigation into How Free Speech Tasks Help Detect People Who Are at Risk of Mild Cognitive Impairment
Сохранить в:
| Опубликовано в:: | Alzheimer's & Dementia vol. 21 (Dec 1, 2025) |
|---|---|
| Главный автор: | |
| Другие авторы: | , , , , , , , |
| Опубликовано: |
John Wiley & Sons, Inc.
|
| Предметы: | |
| Online-ссылка: | Citation/Abstract Full Text Full Text - PDF |
| Метки: |
Нет меток, Требуется 1-ая метка записи!
|
MARC
| LEADER | 00000nab a2200000uu 4500 | ||
|---|---|---|---|
| 001 | 3286889503 | ||
| 003 | UK-CbPIL | ||
| 022 | |a 1552-5260 | ||
| 022 | |a 1552-5279 | ||
| 024 | 7 | |a 10.1002/alz70859_102772 |2 doi | |
| 035 | |a 3286889503 | ||
| 045 | 0 | |b d20251201 | |
| 100 | 1 | |a Park, So Yoon |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 245 | 1 | |a A Preliminary Investigation into How Free Speech Tasks Help Detect People Who Are at Risk of Mild Cognitive Impairment | |
| 260 | |b John Wiley & Sons, Inc. |c Dec 1, 2025 | ||
| 513 | |a Journal Article | ||
| 520 | 3 | |a Background Voice biomarkers can effectively indicate early Mild Cognitive Impairment (MCI)1‐10. This study evaluates how each speech type performs in screening through active tasks. We also examine which combinations of these tasks best detect MCI. Method We designed three tasks to that end: Scripted Reading (Task 1) and Picture‐Based Question and Answer (Task 2) for structured speech, and Spontaneous Speech‐Based Storytelling (Task 3) for semi‐structured speech (Figure 1). We collected 129 speech samples from 21 participants. Using Recursive Feature Elimination with Cross‐Validation (RFECV), we selected 32 key features from over 1,700 acoustic and linguistic ones for classification (Figure 2). We framed the evaluation process as a combinatorial problem where y=f(Task1 Task 2 Task 3). Here, y indicates whether a person is at risk of MCI, and f represents the predictive model that we trained with the speech samples. Result The decoding analysis revealed that the combination of Task_1, 2 and 3 achieved the highest AUC performance (AUC 0.963; 100% ratio). Relative to this maximum performance, the combination of Task_2 + Task_3 achieved 0.869 (90.2%), followed by Task_3 0.822 (85.4%), and Task_1 + Task_3 0.817 (84.8%; Figure 3A). Among these, using all tasks resulted in the best classification Results an AUC of 0.963, specificity of 0.633, and sensitivity of 0.977 (Figure 3B). The feature importance analysis revealed that 1st quantile regression coefficient of MFCC[14] and 50% upper level time of MFCC delta[9] were the most significant features contributing to the classification model (Figure 3C). Lastly, the acoustic features derived from Task_3, pcm_zcr_sma_de_lpc4 (representing the fourth coefficient of linear predictive coding; LPC) and pcm_zcr_sma_de_lpgain (reflecting the signal‐to‐noise ratio based on energy distribution), showed clear differences between normal and patient groups (Figure 3D). Conclusion Among various task combinations, the combination of Task_1+Task_2+Task_3 consistently achieved the best results, with Task 3 being included in all high‐performance combinations. Feature importance analysis and target variable distribution further emphasized the greater contribution of acoustic features compared to linguistic features in classification performance. | |
| 653 | |a At risk populations | ||
| 653 | |a Spontaneous speech | ||
| 653 | |a Storytelling | ||
| 653 | |a Recursion | ||
| 653 | |a Classification | ||
| 653 | |a Signal to noise ratio | ||
| 653 | |a Cognition | ||
| 653 | |a Cognitive ability | ||
| 653 | |a Freedom of speech | ||
| 653 | |a Cognitive impairment | ||
| 653 | |a Acoustics | ||
| 653 | |a Elimination | ||
| 653 | |a Prediction models | ||
| 653 | |a Biological markers | ||
| 653 | |a Medical screening | ||
| 653 | |a Decoding | ||
| 653 | |a Speech | ||
| 700 | 1 | |a Lee, Ju Hyun |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Kim, Whani |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Ko, Hyun Jeong |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Yun, Byung Hun |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Kim, Jin Sung |u Graduate School of AI and Informatics, Sangmyung University, Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Lee, Jee Hang |u Graduate School of AI and Informatics, Sangmyung University, Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Kim, Geon Ha |u Ewha Medical Research Institute, Ewha Womans University, Seoul, Korea, Republic of (South) | |
| 700 | 1 | |a Kim, Jinwoo |u HAII Inc., Seoul, Korea, Republic of (South) | |
| 773 | 0 | |t Alzheimer's & Dementia |g vol. 21 (Dec 1, 2025) | |
| 786 | 0 | |d ProQuest |t Consumer Health Database | |
| 856 | 4 | 1 | |3 Citation/Abstract |u https://www.proquest.com/docview/3286889503/abstract/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch |
| 856 | 4 | 0 | |3 Full Text |u https://www.proquest.com/docview/3286889503/fulltext/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch |
| 856 | 4 | 0 | |3 Full Text - PDF |u https://www.proquest.com/docview/3286889503/fulltextPDF/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch |