Neurological patients due to stroke may suffer from disrupted action due to cognitive deficits which prevent them from maintaining independent lives. CogWatch will focus on neurological patients with symptoms of Apraxia and Action Disorganisation Syndrome (AADS) who, while maintaining their motor capabilities, commit cognitive errors during every-day goal-oriented tasks which premorbidly they used to perform automatically.
Most common rehabilitation ICT systems are focused on treating physiological symptoms of stroke, such as hemiparesis and are not appropriate for rehabilitation of cognitive impairments. Moreover, they are based on robot and/or virtual environment platforms which are expensive and impractical for home installations. In addition, they are designed as rehabilitation stations which patients have to access and adapt to the way the systems operate. As a consequence, this affects the continuity of the therapy and weakens its impact.
A new Personal Healthcare System (PHS) for cognitive rehabitation of action after stroke is proposed which will be affordable, customisable and capable of delivering continuous cognitive rehabilitation at home, when it is needed. The proposed CogWatch project plans to exploit intelligent tools and objects, portable and wearable devices as well as ambient systems to provide personalised cognitive rehabilitation at home for stroke patients with AADS symptoms.
The COGWATCH Project is partially funded under the 7th Framework Programme by the European Commission. It is developed by 8 organisations of 3 European Countries (Germany, Spain and United Kingdom).
Description of target users and groups
Stroke patients that suffer from Apraxia or Action Disorganisation Syndrome (AADS) and the healthcare professionals that take care of them.
Description of the way to implement the initiative
Many stroke patients can suffer from Apraxia or Action Disorganisation Syndrome (AADS) which is shown by impairments of cognitive abilities when carrying out daily living activities, such as washing, dressing and food or drink preparation.
CogWatch will develop intelligent objects and tools which will help to retrain patients on how to carry out activities of daily living, by providing persistent multimodal feedback to them whilst attempting to complete the tasks.
More specifically it will:
- Guide patients’ actions and make them aware of cognitive errors when they occur;
- Inform patients what actions they need to take in order to correct their errors;
- Alert patients if their safety is at risk when handling tools and objects inappropriately;
- Conduct a literature review to gather up to date evidence on the effectiveness of the latest assessment and rehabilitation practices;
- Improve the assessment and classification of AADS patients by using established methods including cognitive screening and Functional Magnetic Resonance Imaging (FMRI).
The project will increase the scientific knowledge of AADS. Through carefully designed studies, it will contribute to gaining knowledge about:
- the severity and diversity of symptoms;
- multimodal cues needed to facilitate action guidance and make patients aware of errors committed and imminent risks; and
- the intensity and frequency of rehabilitation needed to result in sustainable improvements.
More importantly, CogWatch will gather the requirements of healthcare professionals and caregivers.
It will use surveys, questionnaires and interviews to identify system features that maximise usability and minimise the effort of intervention. The project researchers are currently carrying out action analysis based on video, markers and object data that were collected during the patient studies and using current psychological models of action to devise labels for the hierarchical labelling of these actions.
The project will develop a recognition system that can identify when an action is performed based on heterogeneous data from multiple sensors. This will be used to recognise an intention to carry out a specific task, when a task is performed and when it is completed.
Advanced statistical techniques will support an action prediction system that will be able to predict when specific actions and tasks will happen. This will allow researchers to develop outcome management that will predict the consequence of patients’ behaviour and provide action guidance cues, error recovery and risk avoidance.
Data from sensorised objects and tools and video motion capture are synchronised and fed through action recognition software into a Markov Decision Process-based Task Model. This determines the next expected action to provide multimodal feedback to the user while he or she attempts to complete activities of daily living such as making a cup of tea.Technology choice: Proprietary technology
Main results, benefits and impacts
Achieving cognitive rehabilitation at home for patients with AADS will have a significant impact on their personal life and on their families. Physical independence improves the emotional life of the patient by improving self image and confidence that in turn will boost a patient’s motivation to continue rehabilitation.
Additionally, independence increases patient inclusivity leading to greater socialising with family members and friends rather than being served by them. CogWatch could also be used in the future to monitor and assist other neurological patients, such as dementia or closed head brain injury patients, showing the same action disruption symptoms as AADS patients.
Personal independence also has great implication for the healthcare system that provides care for AADS patients.
By developing a customised telesupervisory rehabilitation system, CogWatch will reduce the hospitalisation rate and number of home visits by healthcare professionals. This will have significant economic benefits for national healthcare systems. Stored data about the progress of rehabilitation at the central repository of the CogWatch manager will allow healthcare professionals to design a more effective model of disease management.
The data will also be accessible by scientists and engineers who will use it to increase knowledge of the disease and improve rehabilitation modules and hardware.
Return on investmentReturn on investment: Not applicable / Not available
Track record of sharing
The project is being actively disseminated by the consortium through:
- Academic Journals (e.g. Journal of Cognitive Neuroscience, IEEE: Speech and Language, Sensors, Neuropsychologia);
- Oral Presentations (e.g. Symposium of Behavioural Neurology, RoboCITY);
- Posters (e.g. International Conference on Human-Computer Interface, Congress of the German Society of Neurorehabilitation, International Neuropsychological Society, British Neuropsychological Society, International Conference on Health Informatics, Recent Advances in Assistive Technology & Engineering);
- Book Chapters (e.g. Understanding Stroke Ed R.Sassoon, in press);
- Wider Community (e.g. College of Occupational Therapists SSNP, UK Stroke Forum, UK Stroke Assembly, MEDICA Tradefair, NIHR Stroke Research Network, European Knowledge Tree, UK Annual Parliamentary report on R&D in Assistive Technology);
- Media (The Guardian, 2020health Think Tank Blog).
The first lesson learned is that it is possible to adapt speech recognition techniques to the recognition of actions in the performance of activities of daily living (ADL).
Second there has been recognition of the need for individualisation of feedback cues given to AADS patients in guiding and supporting their ADL task performance.
Third, we have been pleasantly surprised how stakeholders (Helathcare professionals, carers and stroke survivors) are very interested and supportive of the efforts of the researchers to provide ICT based approaches to rehabilitation.Scope: International