Experimental results on multi-modal fusion of EEG-based personal verification algorithms

Abstract

Recently, the use of brain activity as biometric trait for automatic users recognition has been investigated. EEG (Electroencephalography) signal is more often used in the medical field for diagnostic purposes. However, early EEG studies adopted similar signal properties and processing tools to study individual distinctive characteristics. As a matter of fact, features related mostly to a single region of the scalp were used, thus losing information on possible links among brain areas. In this work we approached the investigation of the EEG signal as possible biometric by focusing on two recent methods based on functional connectivity, which, in contrast with previous approaches, tend to estimate the complex interactions between EEG signals by measuring the time-series statistical interdependence. Thanks to their potential complementary, we explored their fusion by feature-level and match score-level approaches. Experimental results have shown a performance improvement with respect to that of the individual systems

EEG-based personal identification: comparison of different functional connectivity metrics

“…These findings highlight a tendency to obtain higher performance (lower EER) for the higher frequency bands. This result confirms previous findings [16], [18] which suggested a potential role of myogenic activity, which is known to affect high frequencies [30], on the definition of individual EEG characteristics. Moreover, these reported results do not show any potential effect induced by the use of metrics which correct for signal spread.…”

“…In contrast, the second category is mainly oriented to detect characteristic patterns of induced brain activity at rest (both during eyes-closed and eyes-open. In line with the extensive use of tools from modern network science to understand brain complex organization (Stam, 2014), measure of functional connectivity (La DelPozo-Banos et al, 2015;Han et al, 2015;Garau et al, 2016) and network metrics have been recently proposed Fraschini et al, 2015) as EEG-based biometric traits. However, it seems still evident that there exists a gap between current investigations of EEG signal as neurophysiological marker and its application in personal verification systems.…”

“…Nevertheless, the reason why these metrics convey different subject specific information has not been investigated yet. Following what reported in (Garau et al, 2016;Fraschini et al, 2015), the present paper aims to investigate and compare the impact of some of the most commonly used techniques to estimate functional connectivity on the ability to detect personal unique discriminative features based on inter-regional interaction profiles. In order to answer this question, we focused our attention on measures based on different properties of the original signal.…”

Variability and stability of large-scale cortical oscillation patterns

“…Our long-term classification results suggest interesting biometric applications for EEG network analyses. Although many EEG features have been examined for their biometric potential (Del Pozo-Banos et al, 2014 ), large-scale oscillatory patterns, including functional connectivity profiles, have only recently received attention for subject identification (Garau, Fraschini, Didaci, & Marcialis, 2016 ; Maiorana et al, 2016 ; Rocca et al, 2014 ). Compared with these reports, our study examined a substantially longer interval between recordings, demonstrating the long-term permanence of these oscillatory patterns.…”

Combining Cryptography with EEG Biometrics

“…Several studies presented the fusion of EEG with other modalities to get a multimodal biometric system such as in [ 45 , 46 ]. Also see a survey of security and privacy challenges in BCI applications in [ 47 ].…”