Abbreviations
Medical expertise
Below, you find resources on MCI, dementia, risk factors, diagnosis, the use of AI in clinical practice and innovative tools like digital cognitive testing and genetic biomarkers. Designed primarily for clinicians, healthcare professionals, and patient organisations, this toolbox supports both traditional and AI-driven decision-making in MCI.
Clinical insights
Resources focusing on the definition and investigation of MCI and dementia. Information on known risk factors and details on risk assessment methods. Existing schemes for dementia prediction, including timely intervention strategies for dementia.
Early dementia diagnosis, MCI-to-dementia risk prediction, and the role of machine learning methods for feature extraction from integrated biomarkers, in particular for EEG signal analysis. Read here: DOI
The organization of functional neurocognitive networks in focal epilepsy correlates with domain-specific cognitive performance. Read here: DOI
Intelligent digital tools for screening of brain connectivity and dementia risk estimation in people affected by mild cognitive impairment: The AI-Mind clinical study protocol. Read here: DOI
AI-Mind deliverables
D1.1, August 2021: Report map of MCI diagnostics and management: This deliverable reports on the current practices for diagnosing and managing MCI in the European countries participating in the AI-Mind project: Finland, Italy, Norway, and Spain. It identifies the tools and methods in use and examines clinicians’ acceptance of AI-based diagnostic tools. The report explores MCI diagnostics, treatment and management, and assesses healthcare professionals’ expectations and concerns about using AI for dementia risk estimation.
D5.1, November 2021: Clinical study dossier (Clinical study protocol, Informed Consent Form): This deliverable outlines the clinical study dossier, focusing on the background and standardized procedures that guide the AI-Mind study’s implementation. It details the methodologies used to ensure that enrolment, procedures, and test administration are consistent across all participating clinical sites.
Presentation: A new pathway for identifying MCI and opening a window of therapeutic possibilities. Watch recording.
Presentation: AI-Mind: Artificial intelligence tool for Alzheimer’s Disease diagnosis. Watch recording.
Poster: A study protocol for AI-Mind: Artificial intelligence combined with neurophysiological and neuropsychological measures for dementia prediction. Access the poster here.
Info cards: dementia, MCI, risk factors. Available here.
Digital cognitive testing
Collected resources provide an overview and share information about the available tools for digitalised cognitive testing and their effectiveness as well as acceptance in the clinical settings.
Stay up to date – visit soon to access any new publications on this topic.
AI-Mind deliverables
D5.4, August 2023: Midterm report on the AI-Mind Predictor versus SOA including a report on the use of digitalised cognitive tests: This deliverable outlines the conceptual development framework of the AI-Mind Predictor and its expected accessibility in clinical practice compared to state-of-the-art diagnostic procedures for evaluating the clinical trajectory of individuals with MCI. It also examines the scientific basis for using digitalized cognitive tests, specifically the Cambridge Neuropsychological Test Automated Battery (CANTAB). Additionally, the report explores the level of trust in CANTAB across different countries, based on insights gathered through a survey conducted among clinical sites.
Poster: Exploring Novel RT-Based Measures Calculated from Detailed Paired Associate Learning Task Data: Preliminary Analyses. (CTAD 2024) Access the poster here.
Info cards: dementia, MCI, risk factors. Available here.
Brain screening methods (EEG & MEG)
Resources provide an introduction to high-density EEG, covering the basics of how brain signals are captured, including MEG, noise is reduced, and features are extracted.
Introducing Region-Based Pooling for handling a varied number of EEG channels for deep learning models. Read here: DOI
Bayesian reduced rank regression models generalizable neural fingerprints that differentiate between individuals in magnetoencephalography data. Read here: DOI
Advancing EEG prediction with deep learning and uncertainty estimation. Read here: DOI
EEG electrodes and where to find them: automated localization from 3D scans. Read here: DOI
AI-Mind deliverables
D2.1, October 2021: Standardisation of available and prospective data collection: This deliverable outlines the shared requirements, technical solutions, and methods for collecting clinical data at AI-Mind partner sites. It defines protocols for using research devices, materials, and data-handling practices during participant visits. Clinical data have been gathered from five sites in four countries (Norway, Finland, Italy, and Spain), covering EEG/MEG data, cognitive testing, blood samples, and questionnaires. The focus is on handling prospective data but also sets minimum standards for retrospective EEG data handling.
D2.2, February 2022: Implementation of the central databank: The technology of the AI-Mind tools uses a mix of retrospective and prospective data, which is processed and stored in a central data repository (CDR). The CDR ensures data persistence, security, and integrity for all AI-Mind applications. This deliverable explains the CDR’s implementation on the TSD platform, highlighting key components like the data lake (DL) and data warehouse (DW), details the design for managing raw data and outlines the structure for future standardised and curated data.
Webinar: Fundamentals of source reconstruction on EEG and MEG. In this webinar, you will learn how the measured signals arise, how we can detect them noninvasively outside the head, and how EEG and MEG relate to each other. In particular, the sensitivity of EEG and MEG to different types of neural signals (synaptic currents, action potentials) is discussed. Watch recording.
Webinar: The neurophysiological basis of EEG and MEG signals. This webinar aims to instruct in the mathematical basis of source reconstruction, the different methods and their assumptions, and the suitability of each method for EEG or MEG. Watch recording.
Presentation: Stability of magnetoencephalography (MEG) and EEG spectral features in Mild Cognitive Impairment. Watch recording.
Interview: How does a brain work and what insights can we gain through the use of magnetoencephalography (MEG)? Watch recording.
Biomarkers
The resources in this section introduce traditional and new biomarkers in the field of MCI, with a focus on the genetic and biomarker analysis used in the AI-Mind project.
Intelligent digital tools for screening of brain connectivity and dementia risk estimation in people affected by mild cognitive impairment: The AI-Mind clinical study protocol. Read here: DOI
Recommended reads from AI-Mind partners (not AI-Mind results):
- Roche recommends: GREEN: A lightweight architecture using learnable wavelets and Riemannian geometry for biomarker exploration with EEG signals. Read here: DOI Open source data and code available on GitHub.
Stay up to date – visit soon to access any new reports on this topic.
Info cards: Blood sample analysis. Available here.