Internet Explorer is no longer supported by Microsoft. To browse the NIHR site please use a modern, secure browser like Google Chrome, Mozilla Firefox, or Microsoft Edge.

Research specification - Digital approaches for dementia early detection, diagnosis and stratification


Published: 14 October 2021

Version: 1 - October 2021

Print this document


The NIHR Invention for Innovation (i4i) Programme is inviting applications for a dementia-specific funding call aimed at the research and development of digital approaches for the early detection, diagnosis and stratification of individuals with dementia. The i4i Programme supports the research and development of innovative medical technologies that have demonstrated proof-of-concept and have a clear developmental pathway towards new products for ultimate NHS use.

Dementia research is an ongoing priority for the Department of Health and Social Care (DHSC) and for the NIHR. There is an identified need for robust assessments to validate digital biomarkers as accurate and precise diagnostic or prognostic methods that correlate with established assessments, and to evaluate their clinical and patient use through appropriate studies.

This call encourages multi-disciplinary teams across different organisations and healthcare settings to address these priorities, harness innovation and support a step forward in the management of dementia.


The World Health Organisation (WHO) defines Dementia to be a syndrome in which there is deterioration in cognitive function beyond what might be expected from normal ageing. Dementia is an umbrella term which includes multiple subtypes of the condition, affecting daily functioning through a loss of ability to think, remember, comprehend, calculate, learn or judge. There is also an associated impairment in emotional control, social behaviour and motivation.

Dementia is an increasing burden to healthcare globally due to a growing, ageing population. In 2019, the London School of Economics' Care Policy and Evaluation Centre estimated that 850,000 people are living in the UK with dementia, with these figures estimated to rise to 1.6 million by 2040. Motor and sensory changes may precede the first clinical manifestation of cognitive symptoms by several years. Early and accurate detection of cognitive decline will enable optimal management of the condition, and allow patients to access support and treatment when they are likely to be most effective.

In the UK, the government is working closely with charities, industry, the NHS and universities to encourage further investments in dementia research. UK bodies have risen to the ‘UK Industrial Strategy’s Ageing Society Grand Challenge’, to support its mission towards an ageing population that can enjoy at least five extra years of healthy and independent living, by 2035. In 2015, the Prime Minister’s ‘Challenge on Dementia 2020’ highlighted the ‘diagnostic gap’ (ensuing from a worldwide undetected dementia prevalence rate of 61.7% [1]), the need for timely and equal access to diagnosis for everyone, and for a more joined-up delivery of care.  

At present, the UK dementia patient journey is usually triggered by a GP referral to memory assessment services where patients suspected to be in the early stages of cognitive decline are seen by a specialist, such as a psychologist, psychiatrist or neurologist. Suspected dementia patients may then be subjected to a combination of cognitive testing, MRI/CT imaging, blood tests, and in some cases a lumbar puncture for cerebrospinal fluid (CSF) sampling. These diagnostic tests can be invasive, endure high costs, need to be performed in a clinical environment and cannot be carried out on a regular basis, often presenting an incomplete view of the patient's condition [2]. Hence, there are significant opportunities for innovations to identify people in the early stages of the condition, through consistent and non-invasive monitoring of prodromal signs. 

However, an early but non-specific dementia diagnosis may not be sufficient to prescribe an effective early intervention. Dementia subtypes are biologically diverse and have different therapeutic targets, which influence prognosis and choice of treatment. Therefore, distinguishing patients according to the dementia subtype will improve the management of the condition, and facilitate the stratification of patients for inclusion in suitable research studies.   

Research priorities

Digital biomarkers are objective physiological and behavioural measurements of key functions affected by memory disorders and dementia, that are collected by means of digital approaches. Examples include speech and language, sensory acuity, eye movement, motor function, keystroke dynamics, gait, sleep patterns, and cognitive processing. Digital biomarkers offer a unique opportunity to collect longitudinal, consistent measurements over long periods of time and monitor progression of the condition, through non-invasive assessments that do not rely solely on clinical facilities for follow-up and timely detection. 

The priorities below present identified unmet needs and form the scope of the i4i Dementia funding call. 

Prediction, early detection and diagnosis of dementia 

Digital biomarkers collected through technologies (software and hardware), electronic health records or other clinical datasets that can be assessed online or in primary care, may improve the capability to accurately predict, detect and diagnose dementia, to enable the earliest possible coordination of care. Predictions of future dementia risk and onset of dementia are expected to be most powerful when a digital biomarker is combined with clinical or genetic testing, or by combining multiple digital biomarkers.  

However, there are still significant gaps in the evidence that validates the quality and utility of digital biomarkers as reliable indicators of early stage dementia or to rule out the condition. The lack of evidence is especially observed in large-scale studies of digital biomarkers in real world settings (for example, homes, primary care, and in the community), resulting in suboptimal evidence to meet the validity, reliability, sensitivity and robustness standards required by regulators. Additionally, further considerations are needed for which combinations of digital biomarkers are most informative of an individual’s health status, and those most amenable to long term patient use and engagement. 

Stratification of people at risk of developing dementia or the subtyping and prognosis of dementia patients

As previously defined, dementia is an umbrella term which includes multiple subtypes of the condition affecting memory, cognitive abilities, and behaviour. It has been widely recognised that a solution to effectively stratify individuals is needed, since tailoring interventions for specific subtypes is an important component of appropriate clinical management and care. Technologies that can determine the subtype and stage of dementia will be of significant prognostic value and help to inform the choice of intervention. In light of the recent FDA approval of Aducanumab in the United States, additional treatment options for Alzheimer's disease are now available, further highlighting the need for accurate subtyping to ensure correct drug prescribing.

At present, the monitoring and follow-up of individuals that have had early cognitive symptoms assessed in clinical practice, is not typically commissioned in England. Patients are often referred back to primary care without further support or intervention and are reliant on symptomatic disease or declining cognition to be referred back to the clinic. There is a need at this stage to stratify those at higher risk of developing dementia, so that they can be monitored more closely and prioritised for further investigation without the need for newly manifesting symptoms. Stratification of the condition or of the prodromal phase should therefore classify individuals with respect to the probable rate of progression of the condition.

Call details 

Applicants are invited to submit proposals addressing one or both of the above-mentioned research priorities. 

The competition will support proposals covering any of the following aims:

  • the large-scale pragmatic validation of digital biomarkers for dementia (DBDs);
  • the development of DBDs into advanced diagnostic or prognostic prototypes or products;
  • clinical evaluations toward near market-ready solutions

Entry point requirements 

  • At least 1 DBD with proof-of-concept and a technical approach (for example, using statistical, machine learning, deep learning or other data engineering methods) that can be used to interrogate the DBD’s clinical relevance. 

Please note that the proof-of-concept (PoC) for the selected digital biomarker/s should demonstrate a potential for use as an early warning sign or risk/susceptibility indicator in the appropriate patient population. The PoC data should ideally have been generated from a minimum of 50 subjects and must be adequately described within the funding application form. However, where less PoC data has been obtained, applications should fully justify the readiness of the proposed innovation for this funding.

  • The selected digital biomarker/s should leverage technologies for data collection such as clinical databases (for example, electronic health records), smartphone, tablet or desktop applications (for example, eye-tracking software), wearables (for example, smart watches), sensory (for example, smart home systems, sleep sensors), or VR/AR platforms (for example, for interactive predefined tasks). 

Project requirements

  • For pragmatic validation studies and clinical evaluation studies, standard-of-care comparator/s should be considered (for example, clinical cognitive function assessments, clinical biomarkers, genetic or polygenic risk scores, and so on), alongside a clear description of the study design including the endpoints, target population, data collection and statistical analysis. 
  • Recruitment of study participants should consider the diversity of patient populations, including those less likely to present memory concerns to a GP, or where there is prevalence of risk factors such as an ageing population, or a lack of local access to dementia services.
  • End user (patients, service users, familial carers or other carers) involvement to qualitatively assess product accessibility, acceptability, usability and long-term engagement, and consideration of how the outcomes (early warning alerts, diagnosis, prognostic strata, predictions, and so on) will be read out and integrated into follow-up clinical assessments and/or health records. 
  • Technical activities may include (but are not limited to): verification of the technical approach (for example, a classifier), algorithm development and refinement, software and hardware development and integration, user interface and/or healthcare professional dashboard design, information governance and interoperability work.
  • Clinical activities may include (but are not limited to): comparative trials to validate the clinical relevance of a DBD (or DBDs) or to test the efficacy of diagnostic/prognostic prototypes; and qualitative, real-world evaluations to determine whether the innovation is fit for purpose.
  • Where applicable, further eligible project activities may include health economic analyses; CE/UKCA marking and preparations for other regulatory submissions; in-depth market analysis; intellectual property (IP) protection and other relevant commercial activities. 

Applicant eligibility 

  • Project teams must be multidisciplinary, involving relevant collaborations between technology developers, data scientists and clinical staff. We encourage clinical applicants to not only comprise clinicians involved in the diagnosis and treatment of dementia, but also those involved in a target setting (for example, GPs) or in the care of at-risk patient populations (for example, renal, stroke physicians and healthcare professionals).
  • The lead organisation must be based in England, and should be best placed among the project parties, to lead the research, and either own or have full access to the background IP.
  • The lead organisation and collaborating organisations may be any of the types of eligible organisations listed below. Collaborations may be sought across the UK, as deemed appropriate. The following types of organisations are eligible for funding:
    • SMEs (have a staff headcount no greater than 250 and annual turnover no greater than €50 million, including start-ups and newly spun out companies);
    • NHS service providers (including Trusts, community care providers and tertiary care centres);
    • Primary care providers (including GP practices, primary care networks (PCNs) and GP Federations). Please note that primary care providers should obtain formal sponsorship in order to apply for NIHR funding as the lead organisation, as per the Research Governance Framework;
    • Higher education institutions (including universities and research institutes);
    • Not-for-profit organisations (including charities and Community Interest Companies). 
  • Specialist services or expertise may be brought into the team through consultancy or sub-contract arrangements; however, appropriate justification must be provided. Sub-contractors may be based outside of the UK if the required expertise or service cannot be reasonably contracted from within the UK. 
  • Applicants must review the NIHR standard research contract before application submission, and agree in principle with its core terms and conditions as they are non-negotiable.

Areas out of scope

  • In vitro diagnostics (for example, for blood biomarkers, cerebrospinal fluid, p-tau, neurofilament biomarker levels and amyloid-beta levels, and so on);
  • Improving the diagnostic accuracy of clinically available imaging modalities (i.e. MRI, CT, PET, and OCT);
  • Proof-of-concept studies;
  • Discovery of new digital biomarkers not previously demonstrated to be linked to dementia pathology;
  • The digitisation or computerisation of existing neuropsychological tests; 
  • Assistive and care management technologies (for example, to prevent falls and fractures).

Budget and duration 

  • All funding proposals are expected to lead to projects of up to 60 months in duration, and may cost up to £2 million.  
  • The release of a project’s budget will be subject to stage gates based on project milestones, go/no-go decision points and periodic progress monitoring. 
  • We cover 100% costs for SMEs, not-for-profit organisations and primary care providers; 100% of the direct research costs for NHS service providers, and 80% FEC for higher education institutions. There are no set rules on the split between collaborators. However, as the contracted organisation, the lead applicant will receive funding payments and would be required to distribute predefined payment shares to the co-applicants (that is, the collaborators) and subcontractors where applicable. 

Support for Awardees 

Dementias Platform UK (DPUK)

The NIHR Invention for Innovation programme is partnering with DPUK to support this call’s awardees. DPUK Data Portal is a world-class repository and analysis environment for dementia-optimised cohort data. Researchers will be able to access high-quality, multi-modal preclinical and clinical data comprising records for more than 3.5 million people.

Awarded projects will be supported as follows: 

  • Upon execution of the NIHR i4i contract and study ethics approval, a dedicated DPUK Lead will make first contact with awardees to discuss the needs of the project, with a particular focus on patient recruitment, data management and analysis.
  • DPUK will then evaluate each project individually, to determine relevance to DPUK’s cohorts, and to establish the best approach of support which may cover:
    • Facilitating additional recruitments depending on the study designs, by utilising a research ready cohort of about 57,000 participants with genetic, biomarker and cognitive data to validate against;
    • Data storage and data analysis thorough DPUK infrastructure and analytical software packages; 
    • Ongoing consultation with DPUK and signposting to other resources where applicable. 

Join Dementia Research (JDR)

Join Dementia Research is a UK-wide service that allows people to register their interest in participating in dementia research. Once registered, they are matched to suitable studies and given the option to take part.  The service is run by the NIHR in partnership with Alzheimer Scotland, Alzheimer's Research UK and the Alzheimer's Society.  More than 50,000 volunteers with or without a dementia diagnosis, are currently registered on Join Dementia Research.

Awardees will benefit from contact with a dedicated Join Dementia Research officer to discuss recruitment mitigation plans. JDR will help to identify and contact patients that may be eligible for your study, and will conduct a feasibility test to ensure that your recruitment strategy is deliverable. To find out more about the JDR service, details are available on the Researchers page


  1. Lang L, Clifford A, Wei L, Zhang D, Leung D, Augustine G, Danat IM, Zhou W, Copeland JR, Anstey KJ and Chen R. Prevalence and determinants of undetected dementia in the community: a systematic literature review and a meta-analysis BMJ open 2017, 7(2): e011146.
  2. Babrak LM, Menetski J, Rebhan M, Nisato G, Zinggeler M, Brasier N, Baerenfaller K, Brenzikofer T, Baltzer L, Vogler C, Gschwind L, Schneider C, Streiff F, Groenen PMA and Miho E. Traditional and Digital Biomarkers: Two Worlds Apart? Digital Biomarkers 2019, 3(2):92-102.