Intelligent digital tools for screening of brain connectivity and dementia risk estimation in people affected by mild cognitive impairment

Project description

AI-Mind, through an intelligent diagnostic tool for early, accurate screening and risk assessment to predict the development of dementia, offers doctors the opportunity to apply preventive interventions for modifiable risk factors (e.g. comorbidities such as cardiovascular disease, diabetes, tobacco, alcohol, depression, dyslipidaemia, diet and sedentary lifestyle) and initiate therapeutic drugs and rehabilitative measures early in the course of the disease in correctly identified cases. In AI-Mind, we aim to develop two new artificial intelligence (AI) based digital tools, the AI-Mind Connector and the AI-Mind Predictor, to analyse existing and routinely collected data in an innovative manner. By extracting salient features from EEG data and more modern MEG, we will convert EEG from an easily accessible with rather restricted analytical power to an easily accessible low-cost tool with a much higher potential and predictive power. The AI-Mind Connector will fully automate the identification of early brain network disturbances. After enriching data from AI-Mind Connector with genetic and cognitive information, AI-Mind Predictor will provide an early marker of risk for dementia in people with MCI with high sensitivity and specificity (>95%). We will integrate these two digital clinical decision-making tools into a cloud-based diagnostic support platform and validate the solution in five European clinical centres. Through the tools described above, AI-Mind will allow doctors to significantly delay patients’ loss of autonomy in daily activities and improve the lives of patients and caregivers. The ongoing coronavirus pandemic underlines the significant vulnerability of people affected by dementia (due to, e.g., the need for isolation of demented individuals, infection prophylaxis and reduced emotional contact with caregivers) and the urgent need for early detection and intervention. Currently, there is an ongoing artificial intelligence revolution that is re-defining the health sector. Personalised and precise healthcare is on the verge of advancing from concept to reality, and the European Commission emphasises ‘how artificial intelligence and supercomputing offer new opportunities to transform healthcare’ in the EU White Paper on Artificial Intelligence 1 published in February 2020. However, the everyday situation of performing dementia risk evaluations in clinics around Europe deviates significantly from the European ambition to use advanced predictive and preventive diagnosis and intervention methods. The world of ML is still too disconnected from computational biology, in part due to the limited sizes of biological datasets and lack of clinical validation and in part because clinical data generation remains largely artisanal. There is misalignment between AI technology providers and clinical practitioners, with a lack of focus/context among the former and a resistance to change among the latter. This problem can only be solved by synergistic actions between these sectors, which AI-Mind seeks to achieve. AI-Mind will be the first diagnostic AIsupported decision-making tool in clinical neurological screening worldwide and to be introduced first in our five European clinical centres.

The global societal and economic cost of dementia was estimated at 1.1% of the global gross domestic product (GDP) in 2015. In Europe, the rapid ageing of the population is predicted to increase the cost of burden to between €725 and €828 billion in 2050. Fifty percent of dementia cases globally are attributable to only seven risk factors: diabetes, hypertension, obesity, depression, physical inactivity, smoking and low education. All of these are largely modifiable2 . This underlines the urgent need for sufficient, early preventive screening of dementia risk and intervention evaluation methods for our ageing population. Proactive adaptation of traditional clinical MCI diagnostic procedures to the digital AI age is essential for the sustainability of European healthcare systems in the face of increasingly ageing societies and limited health budgets. The AI-Mind Connector and Predictor will identify MCI patients that are at risk of dementia and facilitate early preventative strategies. This will reduce the expense of unnecessary investigations of low-risk people and, potentially, prolong the functional independent of patients at risk of dementia. Thus, the tools will have significant positive social and economic implications for the healthcare system and those personally affected by dementia. Therefore, now is the right time to introduce AI-Mind. AI-MIND envisages to deliver over 4000 EEG data points taken from 1000 European subjects and their link to genetic (APOE and P-TAU181) and cognitive function information. In addition, we will be able to deliver EEG data from the British Dementia platform DPUK. Partnering with VirtualBrain/VirutalBrainCloud data will accomplish our goal of developing an automated AI based machine and deep learning algorithms for analysing EEG information. Besides data, AI-MIND will profit immeasurably from the VirtualBrainCloud infrastructure, legal aspects, and innovative research collaborations.

Partnering Organisations

Partnering Project Coordinator

Dr. Ira Ronit Hebold Haraldsen

Name: Dr Ira Haraldsen, MD, PhD
Institute: Oslo University Hospital

Biography: Dr. Ira Ronit Hebold Haraldsen, MD PhD, Neurology and Psychiatry. Adm. expertise of clinical department- and research group leadership. Research background in neuroendocrinology, neurobiology of ageing, and translational innovation project management. She is the PI of AI-Mind (project No. 964220), a 14 million Euro Research and Innovation Action (RIA) H2020-SC1-BHC-06-2020 project. Dr. Haraldsen has a proven track record in heading, participating and stimulating several national and EU projects (ENIGI 2007-2017,) Glasgow-Oslo Sheep Study, Cost actions BM1105. and BM1303. She is leading the Cognitive Health Research group (CoHR) at OUS which has published only during the last 3 years more than 60 papers.



The organization of functional neurocognitive networks in focal epilepsy correlates with domain-specific cognitive performance, Christoffer Hatlestad-Hall, Ricardo Bruña, Aksel Erichsen, Vebjørn Andersson, Marte Roa Syvertsen, Annette Holth Skogan, Hanna Renvall, Camillo Marra, Fernando Maestú, Kjell Heuser, Erik Taubøll, Anne-Kristin Solbakk, Ira H. Haraldsen, 25 June 2021, Journal of Neuroscience Research,

Source-level EEG and graph theory reveal widespread functional network alterations in focal epilepsy, Christoffer Hatlestad-Hall, Ricardo Bruña, Marte  Roa Syvertsen, Aksel Erichsen, Vebjørn Andersson, Fabrizio Vecchio, Francesca Miraglia, Paolo M.Rossini,Hanna Renvall, Erik Taubøll, Fernando Maestú, Ira H.Haraldsen, July 2021, Clinical Neurophysiology,

Human brain networks: a graph theoretical analysis of cortical connectivity normative database from EEG data in healthy elderly subjects, Fabrizio Vecchio, Francesca Miraglia, Elda Judica, Maria Cotelli, Francesca Alù & Paolo Maria Rossini, 13 March 2020 GeroScience,

Classification of Alzheimer's Disease with Respect to Physiological Aging with Innovative EEG Biomarkers in a Machine Learning Implementation, Fabrizio Vecchio, Francesca Miraglia, Francesca Alù, Matteo Menna, Elda Judica, Maria Cotelli, Paolo Maria Rossini, 15 June 2020, Journal of Alzheimer’s Disease, 10.3233/JAD-200171


Key facts

Origin: EBRAINS Research Infrastructure Voucher Programme 2020

Funding: HBP SGA3


News: Collaboration between key strategic initiatives is a step towards an artificial intelligence revolution in brain healthcare; CORDIS, 2021.06.28

The Human Brain Project welcomes 5 new partnering projects, HBP website; 2021.09.06