Subject, session and task effects on power, connectivity and network centrality: A source-based EEG study

Abstract

Inter-subjects' variability in functional brain networks has been extensively investigated in the last few years. In this context, unveiling subject-specific characteristics of EEG features may play an important role for both clinical (e.g., biomarkers) and bio-engineering purposes (e.g., biometric systems and brain computer interfaces). Nevertheless, the effects induced by multi-sessions and task-switching are not completely understood and considered. In this work, we aimed to investigate how the variability due to subject, session and task affects EEG power, connectivity and network features estimated using sourcereconstructed EEG time-series. Our results point out a remarkable ability to identify subject-specific EEG traits within a given task together with striking independence from the session. The results also show a relevant effect of task-switching, which is comparable to individual variability. This study suggests that power and connectivity EEG features may be adequate to detect stable (over-time) individual properties within predefined and controlled tasks.

Clinical applications of EEG power spectra aperiodic component analysis: A mini-review

“…Furthermore, already in 2001 Robinson et al (Robinson et al, 2001) provided a theoretical basis for the conventional division of EEG spectrum into frequency bands, highlighting, at the same time, how the exact bounds are not universal but dependent on the individual. The common thread of individuality has been explored in the following years with studies that showed the important role that the subject-specific characteristics of EEG features may play for clinical purposes (Arns, 2012;Demuru et al, 2017;Fraschini et al, 2015;Pani et al, 2020;Rocca et al, 2014) In this context, it has been recently demonstrated by Demuru and Fraschini (2020) that the aperiodic component also is characterized by strong subject-specific properties and that its features may help to characterize and make inferences at the single subject level, with a better performance than that of the classical frequency bands. Not surprisingly, the traditional approach derives from partially arbitrary choices of the principles to be considered fundamental for the study of EEG signals and it would be definitely interesting to know what the differences in the EEG analysis routine would be if in the past the 1 / f exponential distribution had been defined as a core characteristic of EEG signals (Donoghue, 2020).…”

Robustness of individualized inferences from longitudinal resting state EEG dynamics

“…Our proposed strategy is based on the observation that RS change classification is qualitatively isomorphic to the well-studied problem of RS-based person identification. Numerous studies demonstrate that individual RS activity is highly distinctive to the extent that a person can be identified relative to others solely from their RS activity (Campisi & Rocca, 2014;Finn et al, 2015; see, e.g., Huang et al, 2012;Pani et al, 2020;Valizadeh et al, 2019). In that framework, identification is a form of population inference with a focus on multivariate relationships in a person's RS activity that generalize to samples of the person's own activity but not to the activity of others.…”

“…Numerous prior studies demonstrate that RS activity can serve as a 'fingerprint' for person identification (Campisi & Rocca, 2014;Finn et al, 2015;Huang et al, 2012;Pani et al, 2020;Valizadeh et al, 2019). Although our focus was not on the neural basis of individual differences and trait identification (Demuru et al, 2017;Finn et al, 2017;Gratton et al, 2018;Smit et al, 2005Smit et al, , 2006, a person identification approach, using multi-class classifiers, provided a convenient technical platform for our test of individual-specific change classification.…”

“…Systematic inter-day differences were evident from the dissimilarity between samples from all participants and all days (90 samples per participant per day) (Figure 4b). The dissimilarity between any two samples was described by their correlation distance (= 1 À r, where r is the Pearson's correlation coefficient) (Diedrichsen & Kriegeskorte, 2017;Dimsdale-Zucker & Ranganath, 2019;Pani et al, 2020). For all 24 participants, the mean dissimilarity between samples from the same day was lower than between samples from different days (cross-day) (t 23 = À6.74, p < .0001).…”

On the characterization of cognitive tasks using activity-specific short-lived synchronization between electroencephalography channels

“…First, there may be an excessive information flow from parietal to frontal regions to compensate for the imagination inability during kinesthetic motor imagery task which may cause unexpected short-lived synchronization (Bauer et al, 2015;Gu et al, 2020;Menicucci et al, 2020). Another reason may be the insufficient number of training task periods available for statistical analysis: in the PhysioNet dataset, there are only 45 motor imagery activity task periods in total for each subject which may cause subject-specific interregional synchronization Pani et al, 2020;Xie et al, 2018). By filtering the subject-specific synchronization modulations or using a greater number of motor imagery task periods, more reliable and biophysically relevant channel pairs may be expected to emerge (Allen et al, 2014).…”