Healthy Home, Care Research and Technology Centre, UK DRI
Lead Research Organisation:
Imperial College London
Abstract
This programme will produce safe, reliable and usable technologies for the collection of dementia relevant data. User compliance and acceptability is a key challenge and where possible we are developing passive monitoring using environmental sensors, reserving wearable technologies for when close proximity monitoring is necessary. New technologies will link to the Healthy Home. We will develop adaptive AI/machine learning algorithms to predict clinically relevant information that adapt to seasonality, environmental and contextual changes. The Healthy Home will allow integration and fusion of existing and new forms of data and will provide flexible layers for developing and integrating continual and adaptive AI/machine learning algorithms and new observation generation and clinical pathways and response mechanisms.
Technical Summary
The UK Dementia Research Institute (UK DRI) is an initiative funded by the Medical Research Council, Alzheimer's Society and Alzheimer's Research UK. Funding details for UK DRI programmes will be added from 2020.
This programme focuses on developing safe, reliable and usable technologies for the collection of dementia relevant data. User compliance and acceptability is a key challenge and where possible we are developing passive monitoring using environmental sensors, reserving wearable technologies for when close proximity monitoring is necessary. We will develop adaptive and personalised AI/machine learning algorithms to predict clinically relevant information that adapt to seasonality, environmental and contextual changes. The Healthy Home will allow integration and fusion of existing and new forms of data and will provide flexible layers for developing and integrating continual and adaptive AI/machine learning algorithms and new observation generation and clinical pathways and response mechanisms.
This programme focuses on developing safe, reliable and usable technologies for the collection of dementia relevant data. User compliance and acceptability is a key challenge and where possible we are developing passive monitoring using environmental sensors, reserving wearable technologies for when close proximity monitoring is necessary. We will develop adaptive and personalised AI/machine learning algorithms to predict clinically relevant information that adapt to seasonality, environmental and contextual changes. The Healthy Home will allow integration and fusion of existing and new forms of data and will provide flexible layers for developing and integrating continual and adaptive AI/machine learning algorithms and new observation generation and clinical pathways and response mechanisms.
Publications
Amerineni R
(2021)
Fusion Models for Generalized Classification of Multi-Axial Human Movement: Validation in Sport Performance.
in Sensors (Basel, Switzerland)
Enshaeifar S
(2020)
A Digital Platform for Remote Healthcare Monitoring
Fathy Y
(2019)
Quality-Based and Energy-Efficient Data Communication for the Internet of Things Networks
in IEEE Internet of Things Journal
Fletcher-Lloyd N
(2021)
Home monitoring of daily living activities and prediction of agitation risk in a cohort of people living with dementia.
in Alzheimer's & dementia : the journal of the Alzheimer's Association
Hine C
(2021)
SURVEILLANCE FOR INDEPENDENCE: DISCURSIVE FRAMEWORKS IN SMART CARE FOR DEMENTIA
in AoIR Selected Papers of Internet Research
Hine C
(2022)
Ethical considerations in design and implementation of home-based smart care for dementia.
in Nursing ethics
Hotho A.
(2021)
Preface
in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Kalantari E
(2022)
Network analysis to identify symptoms clusters and temporal interconnections in oncology patients.
in Scientific reports
Kolanko M
(2022)
Clinically relevant monitoring of long-term night-time behaviour and physiology from the homes of people living with dementia.
in Alzheimer's & Dementia
Li H
(2021)
Verifying the Causes of Adversarial Examples
| Title | TIHM/UKDRI Dataset |
| Description | The dataset contains in-home sensory observation and measurement data from homes of people with dementia (n=170 homes). This includes high-resolution data from the participants' in-home movement, use of home appliances, daily activities, sleep, vital signs and daily wellbeing questionnaire. The cohort was part of the previous TIHM study. The study has continued as part of the UK DRI CR7T study since September 2019. The TIHM and UK DRI CR&T studies have collected over 40,0000 days of in-home observation and monitoring data and over 20,000 nights of sleep data collected using smart connected sensors. The dataset is currently used in our research to develop analytical models and digital biomarkers to predict adverse health conditions in dementia care. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2020 |
| Provided To Others? | No |
| Impact | We are using this dataset to develop digital biomarkers. The research is currently focused on identifying digital biomarkers that can be used for early detection of behavioural changes, infections, sleep disorders, risk of falls and social isolation. The current dataset and the previously collected dataset from the TIHM study are partially labelled and annotated for events such as infections, hospital admissions, hypertension, agitation and changes in the wellbeing of people living with dementia. The participants have given their consent for their data to be used for research and the study has obtained approval from an NHS ethics review panel. |
| Description | Royal Society Apex Award, Emergent everyday ethics in infrastructures for smart care |
| Organisation | University of Surrey |
| Department | School of Psychology |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | I am the co-Investigator in Prof Christine Hine's Royal Society Apex Award (University of Surrey). In this project, Prof Hine (a social scientist) and I work together to explore how ethical challenges arise and are managed in designing smart environments. New smart technologies offer great promise to improve care for people living with long-term conditions such as dementia and to enable them to live in their own homes for longer. Significant ethical challenges arise, however, as decisions are made about what features the technology should contain, who has access to data collected by monitoring devices and what actions should be taken in response. |
| Collaborator Contribution | The research has started in October 2020 and is on-going. The work will entail observing the work of participants and interviewing engineers, healthcare professionals, carers and patients who are involved in development of smart technologies for care settings. The aim of these interviews is to identify from each participants' perspective when and how they become aware of ethical challenges, how they distinguish the ethical challenges, and how they deal with the various kinds of issue to negotiate acceptable outcomes. As a result, we will learn more about whether ethical issues can be anticipated in advance and develop ways to build ethical decision-making into the lifespan of a project. |
| Impact | The research is at an early stage. The work will investigate how ethical socio-technical infrastructures are built. The outcomes will be aimed at the community developing smart technologies for care settings and focus on highlighting the recommendations for governance mechanisms that emerge from participant perspectives. |
| Start Year | 2020 |
| Title | IoT-basiertes intelligentes Bettsensorsystem für kontaktlose Atemfrequenzüberwachung |
| Description | Die Erfindung gibt ein Verfahren zum Bestimmen einer Atemfrequenz (RR) eines Patienten an, wobei das Verfahren umfasst: Messen eines Gewichts eines Patienten durch Abrufen gewichtsbezogener Informationen von einer Mehrzahl bettmontierter Lastsensoren (2), Übertragen der gewichtsbezogenen Informationen an einen Server. |
| IP Reference | DE102020117244 |
| Protection | Patent application published |
| Year Protection Granted | 2021 |
| Licensed | Yes |
| Impact | This patent has been licensed to Minebea-Mitsumi-Intec Ltd. Minebea-Mitsumi is developing a non-contact sleep and vital sign monitoring device and this work is the result of a Minebea funded project to develop new machine learning models to analyse the raw sensory data and improve the accuracy of vital sign monitoring using their con-contact sensors. |
| Title | Continual learning by using task conditional neural networks |
| Description | Task Conditional Neural Networks (TCNN) leverage the probabilistic neural networks to estimate the probability density of the training samples. Then produce the task likelihood during the test state to fire the task-specific neurons corresponding to the test samples. TCNN can detect and learn the new tasks fully-automatically without informing the changes to the model |
| Type Of Technology | Software |
| Year Produced | 2019 |
| Open Source License? | Yes |
| Impact | Working with nonstationary data, which is subject to changes is a key challenge for conventional machine learning models. The paper provides a new model for deep representation and continual learning in which the model adapts to changing data or multiple tasks over time without forgetting what is learned in the past. |
| URL | https://codeocean.com/capsule/6003668/tree/v1 |
| Title | Minder Digital Platform |
| Description | We have developed a new digital platform called Minder. Minder is designed for collecting, integrating and processing in-home observation and measurement data using smart technologies. The platform stores and manages the in-home sensor data according to the NHS privacy and security requirements. It is designed to support longitudinal studies that involve in-home activity, wearable and vital sign data collection and processing and online interaction with patients, carers and health and care practitioners. Minder includes an FHIR/HL7 compliant electronic health record (EHR) storage, clinical dashboard and reporting user interface. The first version of the platform will be in operation for data collection and supporting the UK DRI CR&T study. The platform includes a hardware abstraction layer to allow new technologies to be integrated into the platform. |
| Type Of Technology | Software |
| Year Produced | 2020 |
| Impact | This will substantially increase the dataset available for longitudinal studies in dementia care and other long-term conditions. The system is designed by involving clinicians, patients, carers and provides a solution to integrate in-home observation and measurement data and machine learning and analytical algorithms for continuous analysis of the data. This helps to provide time interventions, proactively plan for the care needs of people affected by dementia. The system also allows clinicians to access high granular and continuous data on the progression of dementia. This can help analyse the symptoms and trajectory of cognitive decline and will can contribute to accelerating discovery science in dementia. Through 2021 we will be completing the CE marking process and aim to have the platform accredited as a Class IIa medical. |
| URL | https://dashboard.dev.minder.care/ |
| Title | nargesiPSH/Folded-Hamiltonian-Monte-Carlo: First release of project |
| Description | Missing values exist in nearly all clinical studies because data for a variable or question are not collected or not available. Inadequate handling of missing values can lead to biased results and loss of statistical power in analysis. Existing models usually do not consider privacy concerns or do not utilise the inherent correlations across multiple features to impute the missing values. In healthcare applications, we are usually confronted with high dimensional and sometimes small sample size datasets that need more effective augmentation or imputation techniques. Besides, imputation and augmentation processes are traditionally conducted individually. However, imputing missing values and augmenting data can significantly improve generalisation and avoid bias in machine learning models. A Bayesian approach to impute missing values and creating augmented samples in high dimensional healthcare data is proposed in this work. We propose folded Hamiltonian Monte Carlo (F-HMC) with Bayesian inference as a more practical approach to process the cross-dimensional relations by applying a random walk and Hamiltonian dynamics to adapt posterior distribution and generate large-scale samples. The proposed method is applied to a cancer symptom assessment dataset and confirmed to enrich the quality of data in precision, accuracy, recall, F1 score, and propensity metric. |
| Type Of Technology | Software |
| Year Produced | 2021 |
| Open Source License? | Yes |
| Impact | Lack of samples and having missing values are common issues in healthcare and in remote monitoring research that we have been conducted at the UK DRI. The are various methods to impute for missing values and/or generate augmented values. Especially deep generative models have gained significant attention. However, for high dimensional data with a small number of training samples, deep generative models do not perform well. This work has provided a significantly efficient method using Bayesian machine learning and concepts from high energy physics to create an efficient sampling and data imputation method that can work with a smaller number of samples. |
| URL | https://zenodo.org/record/4580970 |
| Description | Keynote at the International Symposium on Test Automation & Instrumentation, China |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Keynote talk at the International Symposium on Test Automation & Instrumentation; Title of the talk: Using Network-enabled Sensing Technologies in Dementia Care |
| Year(s) Of Engagement Activity | 2020 |
| URL | http://www.istai.org.cn/hy/en/ch/index.aspx?meeting_id=ISTAI2020 |
| Description | Keynote talk at International Workshop on Sensors and Actuators on the Web |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Keynote talk entitled "Search, Discovery and Analysis of Sensory Data Streams" at SAW2019: 1st International Workshop on Sensors and Actuators on the Web, at the International Semantic Web Conference in Auckland, New Zealand. |
| Year(s) Of Engagement Activity | 2019 |
| URL | http://saw.gitlab.emse.fr/2019/#keynotes |
| Description | Plenary talk at the UK Dementia Congress, November 2019. |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | Plenary talk, "How Technology and AI Can Help People Affected by Dementia", the UK Dementia Congress, Yorkshire, November 2019. |
| Year(s) Of Engagement Activity | 2019 |
| URL | https://careinfo.org/wp-content/uploads/2019/10/UKDC-2019-12pp-Brochure-v5-online.pdf |
| Description | TEDx Talk on How technology can help people affected by Dementia |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | TEDx talk organised by the University of Piraeus in Athens attended by over 150 people. The event attracted a wide range of audience and this talk was one of the series on the theme changing perspective. The talk has also been selected and has made available on the TED website. |
| Year(s) Of Engagement Activity | 2019 |
| URL | https://www.ted.com/talks/payam_barnaghi_how_technology_can_help_people_affected_by_dementia |
| Description | invited talk, AI and in-Home Monitoring Technologies to Improve Dementia Care, Neurotechnology for Dementia workshop |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Around 50 academics, scientists and industry experts attended this workshop on Neurotechnology for Dementia. The discussions focused on the new developments in applying technology and AI to improve the outcomes in neurodegenerative conditions and in particular dementia. |
| Year(s) Of Engagement Activity | 2019 |
| URL | https://www.imperial.ac.uk/dementia-research-institute/seminars--events/upcoming-events/neurotechnol... |