Identity Assurance through EEG Recordings

Identity Assurance through EEG Recordings

Massimiliano Zanin (Innaxis Foundation & Research Institute, Spain) and David Papo (Universidad Politecnica de Madrid, Spain)
DOI: 10.4018/978-1-5225-0435-1.ch022


The problem of identity assurance, i.e. determining if a claimed identity can be trusted, has been gaining relevance in the last decade, due to the increasing use of on-line services. While this trend can be seen for many biometric sensors, very few studies have considered the use of brain electric signals. This contribution proposes a first solution, based on the reconstruction of motifs (patterns of connectivity between three electroencephalographic sensors) and on the assessment of their stability across different trials for a single subject. Results indicate that, although computationally costly, this approach is promising in terms of the classification scores obtained.
Chapter Preview


The idea that the brain activity can be described through the electric (and magnetic) field it generates during a task is not new, and was proposed back in 1875 (Swartz, 1988). Electrophysiological techniques such as the ElectroEncephaloGraph (EEG) measure the voltage fluctuations generated by the ionic current within the neurons. More recently, it was recognized that coordination, e.g. synchronization, between electrical activity at different brain regions represent a basic modus operandi of brain information transfer and processing. However, only over the last fifteen years have such structures of interactions been described as networks, thanks to the raising field of complex networks analysis (Albert & Barabasi, 2002; Boccaletti, Latora, Moreno, Chavez & Hwang, 2006).

Network theory is a statistical mechanics understanding of an old branch of pure mathematics: graph theory. In order to represent a system by means of a network, all unnecessary details are deleted, to extract only its constituent parts and their interactions; these are then respectively represented by nodes and links. The structure created by such interactions is then called the network topology. Most social, biological, and technological networks (including, of course, the brain) display substantial non-trivial topological properties, i.e. patterns of connection between their elements are neither purely regular nor purely random (Costa et al., 2011). These properties can be thought of as features describing the network’s structure. The topological properties of a network can directly or indirectly be retrieved from the so-called adjacency matrix, which represents which nodes are connected to which other nodes in a network (Costa, Rodrigues, Travieso & Villas Boas, 2007).

Complete Chapter List

Search this Book: