Image by bamenny from Pixabay
Short Summary: Large Language Models (LLMs) have the potential to revolutionize the healthcare field by offering a wide range of services, such as symptom evaluation, health advice, emotional support, and health education. However, the application of LLMs into healthcare presents several challenges, including the propagation of biases from their training data, a lack of empathy, hallucination, a lack of personalization, and a scarcity of comprehensive and diverse datasets for training. In this project, we are taking several steps to address these issues. First, we are developing and investigating a set of evaluation metrics tailored to healthcare chatbots. Next, our goal is to integrate multimodal health data and various health data analysis tools with LLM-based conversational agents. To achieve this, we are building an open-source LLM-powered framework (opencha.com) that empowers conversational health agents to provide critical thinking, facilitate knowledge acquisition, and enhance problem-solving abilities.
Key Person(s): Prof. Ramesh Jain
Duration of my involvement in the project: 2023 – Present
Selected publications:
- Abbasian, M., Khatibi, E., Azimi, I., Oniani, D., Abad, Z.S.H., Thieme, A., Sriram, R., Yang, Z., Wang, Y., Lin, B., Gevaert, O. and Li, L.J., 2024. Foundation Metrics: Foundation Metrics for Evaluating Effectiveness of Healthcare Conversations Powered by Generative AI. npj Digital Medicine, Nature.
- Abbasian, M., Azimi, I., Rahmani, A.M. and Jain, R., 2023. Conversational Health Agents: A Personalized LLM-Powered Agent Framework. arXiv preprint arXiv:2310.02374.
- Yang, Z., Khatibi, E., Nagesh, N., Abbasian, M., Azimi, I., Jain, R. and Rahmani, A.M., 2024. ChatDiet: Empowering Personalized Nutrition-Oriented Food Recommender Chatbots through an LLM-Augmented Framework. Elsevier Smart Health, IEEE/ACM CHASE.
Image by Chenspec via Pixabay
Short Summary: College students have been adversely affected by the COVID-19 pandemic, experiencing increased loneliness due to social isolation. To tackle this issue, a study is being conducted to explore the impact of social connectedness on college students’ health. Employing a digital and closed-loop approach, the study will provide participants with an app-based intervention named for remote interventions. The study’s outcomes will be measured after the four-week intervention and again at three months post-intervention, aiming to assess the effectiveness of the monitoring in reducing loneliness and fostering social connectedness among college students.
Key Person(s): Profs. Amir M. Rahmani and Jessica Borelli
Duration of my involvement in the project: 2022 – Present
Selected publications:
- Jafarlou, S., Azimi, I., Lai, J., Wang, Y., Labbaf, S., Nguyen, B., Qureshi, H., Marcotullio, C., Borelli, J.L., Dutt, N.D. and Rahmani, A.M., 2024. Objective Monitoring of Loneliness Levels using Smart Devices: A Multi-Device Approach for Mental Health Applications. PLOS One.
- Yang, Z., Azimi, I., Jafarlou, S., Labbaf, S., Borelli, J., Dutt, N. and Rahmani, A., 2023. Loneliness Forecasting Using Multi-modal Wearable and Mobile Sensing in Everyday Settings. IEEE-EMBS International Conference on Body Sensor Networks – Sensors and Systems for Digital Health (BSN’23).
- Sarhaddi, F., Azimi, I., Niela-Vil’en, H., Axelin, A., Liljeberg, P. and Rahmani, A., 2023. Maternal Social Loneliness Detection Using Passive Sensing: Continuous Monitoring in Everyday Settings, JMIR Formative Research.
Image by Mohamed_hassan via Pixabay
Short Summary: The traditional approach to mental healthcare has conventionally relied on episodic psychotherapy, where patients receive treatment plans over multiple visits with providers. However, recent advancements in wearable and mobile technology have sparked interest in digital mental healthcare, enabling individuals to address their episodic mental health symptoms more effectively. The Mental Health Navigator project aims to create a comprehensive platform capable of accurately measuring and estimating an individual’s mental health state. This project includes a therapist-in-the-loop, cybernetic goal-based system, integrating technology and therapist expertise to guide individuals in efficiently managing their mental health.
Key Person(s): Profs. Ramesh Jain and Amir M. Rahmani
Duration of my involvement in the project: 2022 – Present
Selected publications:
- Rahmani, A.M., Lai, J., Jafarlou, S., Azimi, I., Yunusova, A., Rivera, A., Labbaf, S., Anzanpour, A., Dutt, N., Jain, R. and Borelli, J.L., 2022. Personal mental health navigator: Harnessing the power of data, personal models, and health cybernetics to promote psychological well-being. Frontiers in Digital Health, 4, p.933587.
- Yang, Z., Wang, Y., Yamashita, K.S., Khatibi, E., Azimi, I., Dutt, N., Borelli, J.L. and Rahmani, A.M., 2024. Integrating Wearable Sensor Data and Self-reported Diaries for Personalized Affect Forecasting. Elsevier Smart Health, IEEE/ACM CHASE.
- Jafarlou, S., Lai, J., Azimi, I., Mousavi, Z., Labbaf, S., Jain, R.C., Dutt, N., Borelli, J.L. and Rahmani, A., 2023. Objective prediction of next-day’s affect using multimodal physiological and behavioral data: Algorithm development and validation study. JMIR Formative Research, 7(1), p.e39425.
- Khatibi, E., Abbasian, M., Azimi, I., Labbaf, S., Feli, M., Borelli, J., Dutt, N. and Rahmani, A., 2023. Impact of COVID-19 Pandemic on Sleep Including HRV and Physical Activity as Mediators: A Causal ML Approach. IEEE-EMBS International Conference on Body Sensor Networks – Sensors and Systems for Digital Health (BSN’23).
Image by GoldenViolinist via Pixabay
Short Summary: This project introduces a cutting-edge community engagement model that boasts intelligence, interconnectedness, and efficient coordination. By utilizing ubiquitous monitoring and lifelogging, this model brings together a diverse array of community members, such as mothers, families, care providers, and outreach resources. The objective is to empower women through personalized intervention and education, leading to enhanced self-management skills. The project champions a model that is scalable in size, portable across different ethnic communities, and promises improved outcomes through better self-management and community enhanced motivational factors. To assess its impact, the project conducts a controlled study involving underserved Orange County mothers, collaborating with non-profit agencies, hospitals, and local support organizations.
Principal Investigators: Profs. Nikil Dutt, Amir Rahmani, Marco Levorato, and Yuqing Guo
Duration of my involvement in the project: 2019 – Present
Selected publications:
- Feli, M., Azimi, I., Anzanpour, A., Rahmani, A.M. and Liljeberg, P., 2023. An energy-efficient semi-supervised approach for on-device photoplethysmogram signal quality assessment. Smart Health, 28, p.100390.
- Wang, Y., Azimi, I., Feli, M., Rahmani, A.M. and Liljeberg, P., 2023, March. Personalized Graph Attention Network for Multivariate Time-series Change Analysis: A Case Study on Long-term Maternal Monitoring. In Proceedings of the 38th ACM/SIGAPP Symposium on Applied Computing (pp. 593-598).
- Labbaf, S., Abbasian, M., Azimi, I., Dutt, N. and Rahmani, A.M., 2023. ZotCare: A Flexible, Personalizable, and Affordable mHealth Service Provider. Frontiers in Digital Health, 5, 1253087.
- Azimi, I., Oti, O., Labbaf, S., Niela-Vilen, H., Axelin, A., Dutt, N., Liljeberg, P. and Rahmani, A.M., 2019. Personalized maternal sleep quality assessment: An objective IoT-based longitudinal study. IEEE Access, 7, pp.93433-93447.
Image by lavnatalia via Pixabay
Short Summary: Pregnant women who are obese face an increase in the risk of experiencing gestational diabetes mellitus, depression, miscarriage, and preterm birth, among other complications that can have adverse effects on their unborn children. This global challenge demands immediate action. To address it, this project will create an intelligent monitoring system based on the Internet of Things. The primary goal is to detect and predict obesity-related pregnancy complications at the earliest possible stage. Leveraging the concept of cybernetic health, the project will integrate lifestyle and environmental data with medical bio-signals to develop a closed-loop system, revolutionizing maternity care to become more effective, dynamic, and tailored to individual needs. The platform will rely on portable devices and affordable wearable sensors, powered by a multimodal event modeling, activity recognition, and life-logging engine, ultimately providing an all-encompassing and widely accessible pregnancy monitoring service for end-users, mothers, and healthcare providers, enabling timely detection, prediction, assessment, and prevention of pregnancy events.
Principal Investigators: Profs. Amir Rahmani and Anna Axelin
Duration of my involvement in the project: 2018 – 2022
Selected publications:
- Saarikko, J., Axelin, A., Huvinen, E., Rahmani, A.M., Azimi, I., Pasanen, M. and Niela-Vilén, H., 2023. Supporting lifestyle change in obese pregnant mothers through the wearable internet-of-things (SLIM)-intervention for overweight pregnant women: Study protocol for a quasi-experimental trial. PLOS One, 18(1), p.e0279696.
- Niela-Vilen, H., Ekholm, E., Sarhaddi, F., Azimi, I., Rahmani, A.M., Liljeberg, P., Pasanen, M. and Axelin, A., 2023. Comparing prenatal and postpartum stress among women with previous adverse pregnancy outcomes and normal obstetric histories: A longitudinal cohort study. Sexual & Reproductive Healthcare, 35, p.100820.
- Azimi, I., Pahikkala, T., Rahmani, A.M., Niela-Vilén, H., Axelin, A. and Liljeberg, P., 2019. Missing data resilient decision-making for healthcare IoT through personalization: A case study on maternal health. Future Generation Computer Systems, 96, pp.297-308.
Image by sujins via Pixabay
Short Summary: Quality of Experience (QoE) is a crucial factor in ensuring the successful delivery of end-user services for IoT-enabled applications. However, maintaining consistent QoE for end-users presents formidable challenges due to limited resources and dynamic variations across multiple levels of the IoT system stack, including applications, networks, resources, and devices. To address this, we propose an innovative self-aware cognitive architecture called the Internet of Cognitive Things (IoCT). The IoCT intelligently adapts to changing infrastructural computing, communication, and resource requirements while learning from and adjusting to end-user behavior. This adaptability is achieved by leveraging edge computing architectures, specifically Fog computing, to introduce intelligence into integrated multi-scale IoT systems. Our objective is to efficiently manage information acquisition, communication, and processing across different scales of the IoT systems while simultaneously customizing services based on end-user behaviors. Our project includes a partnership with Turku University Hospital to demonstrate a personalized ubiquitous healthcare framework utilizing the Early Warning Score (EWS) system for human health monitoring. Given that healthcare spending accounts for a substantial portion of the GDP in the US, our exemplar application focusing on efficient early detection of life-threatening signs holds the potential to save lives by enhancing the quality of care and providing timely critical health indicators.
Principal Investigators: Profs. Nikil Dutt, Marco Levorato, and Amir Rahmani
Duration of my involvement in the project: 2017 – 2020
Selected publications:
- Anzanpour, A., Amiri, D., Azimi, I., Levorato, M., Dutt, N., Liljeberg, P. and Rahmani, A.M., 2020. Edge-assisted control for healthcare internet of things: A case study on ppg-based early warning score. ACM Transactions on Internet of Things, 2(1), pp.1-21.
- Azimi, I., Takalo-Mattila, J., Anzanpour, A., Rahmani, A.M., Soininen, J.P. and Liljeberg, P., 2018, September. Empowering healthcare IoT systems with hierarchical edge-based deep learning. In Proceedings of the 2018 IEEE/ACM International Conference on Connected Health: Applications, Systems and Engineering Technologies (pp. 63-68).
- Azimi, I., Anzanpour, A., Rahmani, A.M., Pahikkala, T., Levorato, M., Liljeberg, P. and Dutt, N., 2017. HiCH: Hierarchical fog-assisted computing architecture for healthcare IoT. ACM Transactions on Embedded Computing Systems (TECS), 16(5s), pp.1-20.
Image by BiancaVanDijk via Pixabay
Short Summary: Preterm birth (PTB) is the leading cause of neonatal deaths, with approximately 15 million cases each year. Identifying pregnant women at risk and preventing PTB at an early stage is crucial. Physiological parameters can aid in understanding the complex factors behind PTB. Continuous monitoring of these parameters holds significant promise for successful prevention. We propose an IoT-based platform tailored for PTB prevention in everyday settings, with three main contributions: 1) a customized architecture featuring wearable electronic devices suitable for 7-9 months of continuous monitoring, 2) a personalized PTB prevention solution using advanced artificial intelligence methods, and 3) a comprehensive performance assessment through clinical trials to validate the effectiveness of our monitoring approach. Our goal is to address the challenges of PTB and improve the health outcomes for both mothers and newborns.
Principal Investigators: Profs. Amir Rahmani and Anna Axelin
Duration of my involvement in the project: 2017 – 2020
Selected publications:
- Sarhaddi, F., Azimi, I., Axelin, A., Niela-Vilen, H., Liljeberg, P. and Rahmani, A.M., 2022. Trends in heart rate and heart rate variability during pregnancy and the 3-month postpartum period: Continuous monitoring in a free-living context. JMIR mHealth and uHealth, 10(6), p.e33458.
- Sarhaddi, F., Azimi, I., Labbaf, S., Niela-Vilén, H., Dutt, N., Axelin, A., Liljeberg, P. and Rahmani, A.M., 2021. Long-term IoT-based maternal monitoring: system design and evaluation. Sensors, 21(7), p.2281.
- Niela-Vilén, H., Auxier, J., Ekholm, E., Sarhaddi, F., Asgari Mehrabadi, M., Mahmoudzadeh, A., Azimi, I., Liljeberg, P., Rahmani, A.M. and Axelin, A., 2021. Pregnant women’s daily patterns of well-being before and during the COVID-19 pandemic in Finland: Longitudinal monitoring through smartwatch technology. PLOS One, 16(2), p.e0246494.