TY - JOUR
T1 - Moment to moment variability in functional brain networks during cognitive activity in EEG data
AU - Dasari, Naga M.
AU - Nandagopal, Nanda D.
AU - Ramasamy, Vijayalaxmi
AU - Cocks, Bernadine
AU - Thomas, Bruce H.
AU - Dahal, Nabaraj
AU - Gaertner, Paul
N1 - Publisher Copyright:
© 2015 Imperial College Press.
PY - 2015/9/1
Y1 - 2015/9/1
N2 - Functional brain networks (FBNs) are gaining increasing attention in computational neuroscience due to their ability to reveal dynamic interdependencies between brain regions. The dynamics of such networks during cognitive activity between stimulus and response using multi-channel electroencephalogram (EEG), recorded from 16 healthy human participants are explored in this research. Successive EEG segments of 500ms duration starting from the onset of cognitive stimulation have been used to analyze and understand the cognitive dynamics. The approach employs a combination of signal processing techniques, nonlinear statistical measures and graph-theoretical analysis. The efficacy of this approach in detecting and tracking cognitive load induced changes in EEG data is clearly demonstrated using graph metrics. It is revealed that most cognitive activity occurs within approximately 500ms of the stimulus presentation in addition to temporal variability in the FBNs. It is shown that mutual information (MI), a nonlinear measure, produces good correlations between the EEG channels thus enabling the construction of FBNs which are sensitive to cognitive load induced changes in EEG. Analyses of the dynamics of FBNs and the visualization approach reveal hard to detect subtle changes in cognitive function and hence may lead to a better understanding of cognitive processing in the brain. The techniques exploited have the potential to detect human cognitive dysfunction (impairments).
AB - Functional brain networks (FBNs) are gaining increasing attention in computational neuroscience due to their ability to reveal dynamic interdependencies between brain regions. The dynamics of such networks during cognitive activity between stimulus and response using multi-channel electroencephalogram (EEG), recorded from 16 healthy human participants are explored in this research. Successive EEG segments of 500ms duration starting from the onset of cognitive stimulation have been used to analyze and understand the cognitive dynamics. The approach employs a combination of signal processing techniques, nonlinear statistical measures and graph-theoretical analysis. The efficacy of this approach in detecting and tracking cognitive load induced changes in EEG data is clearly demonstrated using graph metrics. It is revealed that most cognitive activity occurs within approximately 500ms of the stimulus presentation in addition to temporal variability in the FBNs. It is shown that mutual information (MI), a nonlinear measure, produces good correlations between the EEG channels thus enabling the construction of FBNs which are sensitive to cognitive load induced changes in EEG. Analyses of the dynamics of FBNs and the visualization approach reveal hard to detect subtle changes in cognitive function and hence may lead to a better understanding of cognitive processing in the brain. The techniques exploited have the potential to detect human cognitive dysfunction (impairments).
KW - brain network variability
KW - cognition
KW - Functional brain network
KW - graph metrics
KW - moment-to-moment
KW - mutual information
UR - http://www.scopus.com/inward/record.url?scp=84947028811&partnerID=8YFLogxK
U2 - 10.1142/S0219635215500211
DO - 10.1142/S0219635215500211
M3 - Article
C2 - 26365114
AN - SCOPUS:84947028811
SN - 0219-6352
VL - 14
SP - 383
EP - 402
JO - Journal of Integrative Neuroscience
JF - Journal of Integrative Neuroscience
IS - 3
ER -