2017 Fiscal Year Final Research Report
Development of a Technology to Quantify Psychiatric Symptoms Utilizing Behavioral and Physiological Data
| Project/Area Number |
15K15434
|
|---|
| Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Psychiatric science
|
|---|
| Research Institution | Keio University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
三村 將 慶應義塾大学, 医学部(信濃町), 教授 (00190728)
江口 洋子 慶應義塾大学, 医学部(信濃町), 研究員 (70649524)
藤田 卓仙 名古屋大学, 経済学研究科, 寄附講座准教授 (80627646)
|
|---|
| Research Collaborator |
YOSHIMURA Michitaka
KITAZAWA Momoko
LIANG Kuo-ching
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Keywords | うつ病 / 機械学習 / 行動 / 生理 |
|---|
| Outline of Final Research Achievements |
Since there is no established biomarkers that reflect the severity of psychiatric disorders, there are problems such as low diagnosis agreement, difficulty in understanding the therapeutic effect in psychiatry. In this study, we examined whether it is possible to develop an machine learning algorithm for quantifying psychiatric symptoms using voice and images analyses. We interviewed patients with major depressive disorder/bipolar disorder and videotaped it. Machine learning was performed based on the collected voice/facial image data, and it was found that the presence or absence of depressive symptoms, its severity could be estimated with a certain degree of accuracy.
|
| Free Research Field |
精神神経科学
|
| Academic Significance and Societal Importance of the Research Achievements |
音声や画像データに基づく重症度評価が可能になれば、時間がかかる、負担が大きいなどで臨床で省略されがちな重症度評価が簡便に行えるようになり、治療効果の定量を通じて、より効果的な治療が選択されることにつながったり、ひいては新薬の開発が容易になったりする可能性がある。さらには、発病前の段階でのセルフケアツールとしても活用可能になれば、社会費用が甚大なうつ病の予防効果も期待できるようになる。
|
