Premium
Automatic early detection of cognitive decline
Author(s) -
Maimon Neta,
Molcho Lior,
Loterrman Tomer,
Pressburger Narkiss,
Sasson Ady,
Intrator Nathan
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.043529
Subject(s) - cognition , cognitive decline , electroencephalography , wearable computer , computer science , population , elementary cognitive task , cognitive impairment , montreal cognitive assessment , cognitive psychology , audiology , psychology , artificial intelligence , medicine , dementia , neuroscience , disease , environmental health , pathology , embedded system
Abstract Background While it is agreed that cognitive decline is underdiagnosed, there is still no single universally accepted and affordable screening method that satisfies all needs in the detection of cognitive impairment. The goal of this presentation is to demonstrate an automatic cognitive decline assessment with Neurosteer Aurora, utilizing a novel EEG analysis as a method for characterization of cognitive activity and detection of early cognitive impairment. Method Neurosteer has developed a miniature wearable neurological sensor combined with an automatic assessment tool for a wide range of brain monitoring and treatment needs. It is based on a single EEG channel and relies on advanced mathematical analysis for a decomposition of the EEG signal into multiple components. These components are then processed with machine learning algorithms to create high level biomarkers for real‐time brain activity interpretation. Additionally, Neurosteer has developed a fully automatic assessment and biomarkers for cognitive decline. It is based on a verbal and musical cognitive assessment during brain activity recording. Machine learning algorithms analyze the combined data to produce a simple‐to‐read report indicating the level of different aspects of cognitive decline. We applied this technology on three different healthy population (N=14, 40, 215) and cognitively impaired (N=24, 10, 9, 8) undergoing several automatically administered cognitive tasks. Result A few biomarkers were extracted using machine learning tools on the collected data. The biomarkers distinguished between different levels of cognitive load as well as resting state patterns. The new groups of cognitively impaired populations were examined with Neurosteer Aurora, completing short and automatic auditory assessments. The patterns of the predefined biomarkers were significantly different for each population, differentiating between clinical populations and negatively correlated with the severity of cognitive decline. Conclusion Using the novel brain activity interpretation and biomarkers enables an automatic and easy‐to‐administer cognitive decline assessment in the clinical population. Such assessment is a step forward towards a wide‐spread screening, allowing early and personalized intervention to slow down cognitive decline, improving health and quality of life outcomes.