Kitazawa, Momoko



School of Medicine, Department of Neuropsychiatry (Shinanomachi)


Project Assistant Professor (Non-tenured)/Project Research Associate (Non-tenured)/Project Instructor (Non-tenured)


Papers 【 Display / hide

  • Evaluating the severity of depressive symptoms using upper body motion captured by RGB-depth sensors and machine learning in a clinical interview setting: A preliminary study

    Horigome T., Sumali B., Kitazawa M., Yoshimura M., Liang K.c., Tazawa Y., Fujita T., Mimura M., Kishimoto T.

    Comprehensive Psychiatry (Comprehensive Psychiatry)  98 2020.04

    ISSN  0010440X

     View Summary

    © 2020 The Authors Background: Mood disorders have long been known to affect motor function. While methods to objectively assess such symptoms have been used in experiments, those same methods have not yet been applied in clinical practice because the methods are time-consuming, labor-intensive, or invasive. Methods: We videotaped the upper body of each subject using a Red-Green-Blue-Depth (RGB-D) sensor during a clinical interview setting. We then examined the relationship between depressive symptoms and body motion by comparing the head motion of patients with major depressive disorders (MDD) and bipolar disorders (BD) to the motion of healthy controls (HC). Furthermore, we attempted to predict the severity of depressive symptoms by using machine learning. Results: A total of 47 participants (HC, n = 16; MDD, n = 17; BD, n = 14) participated in the study, contributing to 144 data sets. It was found that patients with depression move significantly slower compared to HC in the 5th percentile and 50th percentile of motion speed. In addition, Hamilton Depression Rating Scale (HAMD)-17 scores correlated with 5th percentile, 50th percentile, and mean speed of motion. Moreover, using machine learning, the presence and/or severity of depressive symptoms based on HAMD-17 scores were distinguished by a kappa coefficient of 0.37 to 0.43. Limitations: Limitations include the small number of subjects, especially the number of severe cases and young people. Conclusions: The RGB-D sensor captured some differences in upper body motion between depressed patients and controls. If much larger samples are accumulated, machine learning may be useful in identifying objective measures for depression in the future.

  • Evaluating depression with multimodal wristband-type wearable device: screening and assessing patient severity utilizing machine-learning

    Tazawa Y., Liang K.c., Yoshimura M., Kitazawa M., Kaise Y., Takamiya A., Kishi A., Horigome T., Mitsukura Y., Mimura M., Kishimoto T.

    Heliyon (Heliyon)  6 ( 2 )  2020.02

    ISSN  24058440

     View Summary

    © 2020 Objective: We aimed to develop a machine learning algorithm to screen for depression and assess severity based on data from wearable devices. Methods: We used a wearable device that calculates steps, energy expenditure, body movement, sleep time, heart rate, skin temperature, and ultraviolet light exposure. Depressed patients and healthy volunteers wore the device continuously for the study period. The modalities were compared hourly between patients and healthy volunteers. XGBoost was used to build machine learning models and 10-fold cross-validation was applied for the validation. Results: Forty-five depressed patients and 41 healthy controls participated, creating a combined 5,250 days' worth of data. Heart rate, steps, and sleep were significantly different between patients and healthy volunteers in some comparisons. Similar differences were also observed longitudinally when patients' symptoms improved. Based on seven days' data, the model identified symptomatic patients with 0.76 accuracy and predicted Hamilton Depression Rating Scale-17 scores with a 0.61 correlation coefficient. Skin temperature, sleep time-related features, and the correlation of those modalities were the most significant features in machine learning. Limitations: The small number of subjects who participated in this study may have weakened the statistical significance of the study. There are differences in the demographic data among groups although we performed a correction for multiple comparisons. Validation in independent datasets was not performed, although 10-fold cross validation with the internal data was conducted. Conclusion: The results indicated that utilizing wearable devices and machine learning may be useful in identifying depression as well as assessing severity.

  • Survey of the effects of internet usage on the happiness of Japanese university students

    Kitazawa M., Yoshimura M., Hitokoto H., Sato-Fujimoto Y., Murata M., Negishi K., Mimura M., Tsubota K., Kishimoto T.

    Health and Quality of Life Outcomes (Health and Quality of Life Outcomes)  17 ( 1 )  2019.10

     View Summary

    © 2019 The Author(s). Background: Besides research on psychiatric diseases related to problematic Internet use (PIU), a growing number of studies focus on the impact of Internet on subjective well-being (SWB). However, in previous studies on the relationship between PIU and SWB, there is little data for Japanese people specifically, and there is a lack of consideration for differences in perception of happiness due to cultural differences. Therefore, we aimed to clarify how happiness is interdependent on PIU measures, with a focus on how the concept of happiness is interpreted among Japanese people, and specifically among Japanese university students. Methods: A paper-based survey was conducted with 1258 Japanese university students. Respondents were asked to fill out self-report scales regarding their happiness using the Interdependent Happiness Scale (IHS). The relationship between IHS and Internet use (Japanese version of the Internet addiction test, JIAT), use of social networking services, as well as social function and sleep quality (Pittsburgh Sleep Quality Index, PSQI) were sought using multiple regression analyses. Results: Based on multiple regression analyses, the following factors related positively to IHS: female gender and the number of Twitter followers. Conversely, the following factors related negatively to IHS: poor sleep, high- PIU, and the number of times the subject skipped a whole day of school. Conclusions: It was shown that there was a significant negative correlation between Japanese youths' happiness and PIU. Since epidemiological research on happiness that reflects the cultural background is still scarce, we believe future studies shall accumulate similar evidence in this regard.

  • Actigraphy for evaluation of mood disorders: A systematic review and meta-analysis

    Tazawa Y., Wada M., Mitsukura Y., Takamiya A., Kitazawa M., Yoshimura M., Mimura M., Kishimoto T.

    Journal of Affective Disorders (Journal of Affective Disorders)  253   257 - 269 2019.06

    ISSN  01650327

     View Summary

    © 2019 Background: Actigraphy has enabled consecutive observation of individual health conditions such as sleep or daily activity. This study aimed to examine the usefulness of actigraphy in evaluating depressive and/or bipolar disorder symptoms. Method: A systematic review and meta-analysis was conducted. We selected studies that used actigraphy to compare either patients vs. healthy controls, or pre- vs. post-treatment data from the same patient group. Common actigraphy measurements, namely daily activity and sleep-related data, were extracted and synthesized. Results: Thirty-eight studies (n = 3,758) were included in the analysis. Compared with healthy controls, depressive patients were less active (standardized mean difference; SMD=1.27, 95%CI=[0.97, 1.57], P<0.001) and had longer wake after sleep onset (SMD= − 0.729, 95%CI=[− 1.20, − 0.25], p = 0.003). Total sleep time (SMD= − 0.33, 95%CI=[− 0.55, − 0.11], P = 0.004), sleep latency (SMD= − 0.22, 95%CI=[− 0.42, − 0.02], P = 0.032), and wake after sleep onset (SMD= − 0.22, 95%CI=[− 0.39, − 0.04], P = 0.015) were longer in euthymic/remitted patients compared to healthy controls. In pre- and post-treatment comparisons, sleep latency (SMD=− 0.85, 95%CI=[− 1.53, − 0.17], P = 0.015), wake after sleep onset (SMD= − 0.65, 95%CI=[− 1.20, − 0.10], P = 0.022), and sleep efficiency (SMD=0.77, 95%CI=[0.29, 1.24], P = 0.002) showed significant improvement. Limitation: The sample sizes for each outcome were small. The type of actigraphy devices and patients’ illness severity differed across studies. It is possible that hospitalizations and medication influenced the outcomes. Conclusion: We found significant differences between healthy controls and mood disorders patients for some actigraphy-measured modalities. Specific measurement patterns characterizing each mood disorder/status were also found. Additional actigraphy data linked to severity and/or treatment could enhance the clinical utility of actigraphy.

  • Utilization of Facial Image Analysis Technology for Blink Detection: A Validation Study

    Kitazawa M., Yoshimura M., Liang K.C., Wada S., Mimura M., Tsubota K., Kishimoto T.

    Eye &amp; contact lens (Eye &amp; contact lens)  44   S297 - S301 2018.11

     View Summary

    PURPOSE: The assessment of anterior eye diseases and the understanding of psychological functions of blinking can benefit greatly from a validated blinking detection technology. In this work, we proposed an algorithm based on facial recognition built on current video processing technologies to automatically filter and analyze blinking movements. We compared electrooculography (EOG), the gold standard of blinking measurement, with manual video tape recording counting (mVTRc) and our proposed automated video tape recording analysis (aVTRa) in both static and dynamic conditions to validate our aVTRa method. METHODS: We measured blinking in both static condition, where the subject was sitting still with chin fixed on the table, and dynamic condition, where the subject's face was not fixed and natural communication was taking place between the subject and interviewer. We defined concordance of blinks between measurement methods as having less than 50 ms difference between eyes opening and closing. RESULTS: The subjects consisted of seven healthy Japanese volunteers (3 male, four female) without significant eye disease with average age of 31.4±7.2. The concordance of EOG vs. aVTRa, EOG vs. mVTRc, and aVTRa vs. mVTRc (average±SD) were found to be 92.2±10.8%, 85.0±16.5%, and 99.6±1.0% in static conditions and 32.6±31.0%, 28.0±24.2%, and 98.5±2.7% in dynamic conditions, respectively. CONCLUSIONS: In static conditions, we have found a high blink concordance rate between the proposed aVTRa versus EOG, and confirmed the validity of aVTRa in both static and dynamic conditions.

display all >>