Blockchain-enabled application for cognitive assessment in dementia
Frankl’s application to the Wellcome Trust Open Research Fund
Update: We’re delighted to announce that Frankl’s application to the Wellcome Trust was successful. In awarding the team £50,000, the Open Research Fund committee described the project as having “the potential to impact health research, through openness, across multiple fields”.
Summary
The proposal will develop an app for screening memory function in elderly people concerned about declining cognition. It uses a simple word-pairing test that has demonstrated sensitivity and specificity in detecting prodromal Alzheimer’s disease.
The app will be built following a model developed by Frankl Open Science, a Sydney-based blockchain start-up. It will incorporate three key features that facilitate and incentivise open science and FAIR data sharing:
- Data collected via the application is archived “on the fly” in secure repositories.This facilitates data sharing within collaborations; between relevant clinicians, patients, and caregivers; and allows broader data sharing of deidentified data as and when appropriate.
- Access to the application is via a cryptocurrency micropayment. If the data are later shared openly then part of this payment is returned to the researcher. This incentivises researchers to share their data even if the study is not published.
- Writing these transactions to the Ethereum blockchain provides a public, immutable, timestamped record of the data’s existence. Adding the data’s archived location to this metadata makes the data findable. Adding a hash (digital fingerprint) of the data allows the data’s provenance to be established.
In addition to its immediate clinical and scientific benefits, the application will serve as a working example for a new way of conducting open research. The strategy of Frankl Open Science is to share code and other resources so that scientists can integrate the above features into their own data collection protocols.
Vision
At Frankl our mission is to make open science easier and more rewarding for scientists. We recognise that there are significant practical barriers and disincentives to the adoption of open science practices. Our approach is to develop tools that have “open” as the default setting but — crucially — are desirable for researchers to use even without these open features.
Our initial focus is cognitive assessment — an area of expertise within the Frankl team. There is scope to substantially improve the user experience for researchers, clinicians, and educators, as well as the individuals completing the assessments. There are also clear market opportunities. This is important for the long term sustainability of the Frankl project.
Aims
The aim of this proposal is to develop an application for assessing memory abilities in patients with diagnosed Alzheimer’s disease dementia or other forms of dementia, people with Mild Cognitive Impairment (who have higher risk of developing dementia), and cognitively normal elderly people with subjective memory concerns.
The assessment will take the form of a simple paired associate learning task that measures the ability to form new memories. Participants are presented with pairs of words to memorise. After a short delay, they are given one word from each pair and required to select its partner.
Variants of this test have proven sensitive to memory impairments in patients with early stage dementia. The test is simple and quick, meaning that it can be conducted at multiple intervals in longitudinal research to reliably track changes in memory over time and ultimately find utility as a convenient and accurate screening tool in primary care settings.
The application will act as a vehicle for the features described above, including: integrated data management that pushes data to a secure repository; cryptocurrency micropayment that can be refunded when data are shared; blockchain record to establish the existence and provenance of data and increase findability.
Target audience
The application will be of immediate benefit to researchers and clinicians working with individuals with actual or suspected memory impairments. The data management will allow secure sharing of data within collaborations or between relevant clinicians, patients, and caregivers. The application will also be of interest to researchers across all scientific domains, serving as a working demonstration of blockchain-enabled research.
Activities
The project will incorporate four major activities:
- Development of a minimal viable product (MVP)
- Pilot phase in which the MVP is provided to clinicians and researchers for feedback
- Refinement of the application
- Wider release of the application
Influence on open research practices
The application developed within this proposal will facilitate and incentivise open research practices in the following ways:
- Making data sharing easier: The application builds data management into data collection by pushing the data to a secure repository as it is collected.
- Ensuring data are FAIR: Data sent from the app will be fully annotated. Consistent data protocols across apps will allow machine readability. Blockchain record of metadata will ensure findability.
- Incentivising data sharing: The Frankl cryptocurrency token provides a mechanism by which researchers and clinicians can be rewarded for sharing their data.
- Addressing the file drawer problem: Writing metadata to blockchain provides a permanent public record of the existence of all the data collected using this test.
- Creating a scientific supply chain: Hashed data on the blockchain provides a means of establishing the provenance of scientific data.
- Enhancing reproducibility: In the longer term, Frankl will provide a platform for other researchers to add their own tests to the platform and be rewarded with tokens when other researchers re-use their tests. This will enhance reproducibility of data collection methods.
Monitoring and evaluation
Software development aspects of this project will be delivered in accordance with an Agile methodology. Progress will therefore be monitored in terms of the number of tasks delivered, the number of commits to a production branch of code, and conversely the number of blocked or backlogged issues. These will be measured closely throughout the build, and especially at the commencement and finalisation of each 2 week ‘sprint’.
Success indicators will include the existence of the MVP and final prototype, positive feedback from users including researchers, clinicians, and patients (or individuals completing the test).
Indirect success indicators include engagement with the cognitive and clinical neuropsychology fields regarding open science practices, FAIR principles, and the development of standards for data and metadata. In the medium term (i.e., beyond the timeframe of this project), success will be demonstrated by the adoption of the application in research and clinical practice, and the re-use of the code we have developed.
Our longer term vision is for Frankl to become a marketplace for open science software applications in which researchers and clinicians can share applications that they have developed and receive Frankl cryptocurrency tokens when they are re-used. The application proposed here represents the first step towards this goal and the development of such an ecosystem is the ultimate indicator of its success.
