WEKO3
アイテム
パーキンソン病患者のウェアリング・オフ推定における手首装着型フィットネストラッカーの非運動データセットとスマートフォンアプリケーションの使用可能性について
https://doi.org/10.18997/00008917
https://doi.org/10.18997/00008917448c5ddf-9a06-4803-a241-49fe57db24dd
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
sei_k_440.pdf (2.0 MB)
|
|
| アイテムタイプ | 学位論文 = Thesis or Dissertation(1) | |||||||
|---|---|---|---|---|---|---|---|---|
| 公開日 | 2022-06-20 | |||||||
| 資源タイプ | ||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_db06 | |||||||
| 資源タイプ | doctoral thesis | |||||||
| タイトル | ||||||||
| タイトル | The Feasibility of Using Wrist-Worn Fitness Tracker's Non-Motor Datasets and Smartphone Application in Estimating Wearing-Off among Parkinson's Disease Patients | |||||||
| 言語 | en | |||||||
| タイトル | ||||||||
| タイトル | パーキンソン病患者のウェアリング・オフ推定における手首装着型フィットネストラッカーの非運動データセットとスマートフォンアプリケーションの使用可能性について | |||||||
| 言語 | ja | |||||||
| 言語 | ||||||||
| 言語 | eng | |||||||
| 著者 |
Victorino John Noel, Cruz
× Victorino John Noel, Cruz
|
|||||||
| 抄録 | ||||||||
| 内容記述タイプ | Abstract | |||||||
| 内容記述 | Parkinson’s disease （PD） patients mainly suffer from motor symptoms such as tremor, bradykinesia, rigidity, and postural instability. Aside from motor-related symptoms, PD also affects non-motor functions such as sleep, speech, and other mental functions. Then, doctors prescribe the Levodopa treatment （L-dopa） to alleviate these symptoms. However, the “wearing-off” phenomenon occurs when symptoms reappear. Eventually, the prolonged use of L-dopa reduces its effective time, prompting the symptoms to fluctuate frequently from relief （“on” state） to discomfort （“off” state）. These sudden changes and re-appearance of symptoms affect the daily life of the patient and the quality of life （QoL）. Thus, PD patients need to monitor and report their experience to their clinicians for a proper and customized treatment plan. Monitoring and assessment have been focused on motor-related symptoms since PD primarily affects a patient’s motor function. Currently, clinicians routinely assess patients in-clinic to evaluate the L-dopa treatment plan. In addition, self-report measurement tools and patient diaries complement this routine check-up. In contrast to subjective monitoring, wearable devices have been used to collect motion data from hands, legs, or waists to objectively assess PD symptoms and wearing-off periods. The combination of subjective and objective assessment of motor-related symptoms allows clinicians to evaluate the treatment plan. However, PD also affects the patients’ non-motor aspects, such as sleep disturbances, mood changes, and other mental distress. Thus, it is important to include the non-motor aspects in understanding the patients’ PD experience and the wearing-off periods. This dissertation proposes using non-motor datasets found in commercially available fitness trackers to estimate wearing-off periods among PD patients. Hence, the materials and methodology were extensively discussed to investigate the feasibility of using non-motor datasets from commercially available wrist-worn fitness trackers. First, a data collection framework has been established to gather self-reported wearing-off periods utilizing a smartphone application while wearing the fitness tracker. Then, the data processing was explained on how the different datasets were combined. Finally, the model development pipelines were presented, along with their components. As a result, the developed models for each participant produced a weighted accuracy from 68.43% to 76.9% under different considerations. The first consideration was on the number of collection days, with the first round of data collection required 30 days and the second round taking 6 to 11 days. The next consideration was the chosen learning algorithm during the model selection phase. Gradient Boosting algorithm produced the highest average performance across 10 participants. But, since individual models were built for each participant, each participant had different best learning algorithms. Eight of the ten participants had tree-based learning algorithms over the linear algorithms. Finally, the heart rate feature was one of the top important features, aside from the drug intake time. Similarly, each participant had different sets of features, which contributed to the estimation of wearing-off. These considerations led to the feasibility of using commercially available wrist-worn fitness trackers as a source of non-motor datasets in estimating wearing-off. Ultimately, a holistic approach in detecting wearing-off will be considered both motor and non-motor aspects while improving the data collection process for PD patients. | |||||||
| 言語 | en | |||||||
| 目次 | ||||||||
| 内容記述タイプ | TableOfContents | |||||||
| 内容記述 | 1 Introduction| 2 Related Work| 3 Materials and Methodology| 4 Results and Discussion| 5 Conclusion | |||||||
| 備考 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 九州工業大学博士学位論文 学位記番号: 生工博甲第440号 学位授与年月日: 令和4年3月25日 | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Parkinson’s Disease | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Wearing-Off | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Wearable Devices | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Prediction Models | |||||||
| キーワード | ||||||||
| 主題Scheme | Other | |||||||
| 主題 | Machine Learning | |||||||
| アドバイザー | ||||||||
| 柴田, 智広 | ||||||||
| 学位授与番号 | ||||||||
| 学位授与番号 | 甲第440号 | |||||||
| 学位名 | ||||||||
| 学位名 | 博士（情報工学） | |||||||
| 学位授与年月日 | ||||||||
| 学位授与年月日 | 2022-03-25 | |||||||
| 学位授与機関 | ||||||||
| 学位授与機関識別子Scheme | kakenhi | |||||||
| 学位授与機関識別子 | 17104 | |||||||
| 学位授与機関名 | 九州工業大学 | |||||||
| 学位授与年度 | ||||||||
| 内容記述タイプ | Other | |||||||
| 内容記述 | 令和3年度 | |||||||
| 出版タイプ | ||||||||
| 出版タイプ | VoR | |||||||
| 出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 | |||||||
| アクセス権 | ||||||||
| アクセス権 | open access | |||||||
| アクセス権URI | http://purl.org/coar/access_right/c_abf2 | |||||||
| ID登録 | ||||||||
| ID登録 | 10.18997/00008917 | |||||||
| ID登録タイプ | JaLC | |||||||
