Biology subjects

Kozhemiako, N.

Publications and source records attributed to Kozhemiako, N..

2 recordsLinked to original sources

bioRxiv2019

Atypical developmental features of cortical thickness trajectories in Autism Spectrum Disorder

Neuroimaging studies have reported numerous region-specific atypicalities in the brains of individuals with Autism Spectrum Disorder (ASD), including alterations in cortical thickness (CT). However, there are many inconsistent findings, and this is probably due to atypical CT developmental trajectories in ASD. To this end, we investigated group differences in terms of shapes of developmental trajectories of CT between ASD and typically developing (TD) populations.\n\nUsing the Autism Brain Imaging Data Exchange (ABIDE) repository (releases I and II combined), we investigated atypical shapes of developmental trajectories in ASD using a linear, quadratic and cubic models at various scales of spatial coarseness, and their association with symptomatology using the Autism Diagnostic Observation Schedule (ADOS) scores. These parameters were also used to predict ASD and TD CT development.\n\nWhile no overall group differences in CT was observed across the entire age range, ASD and TD populations were different in terms of age-related changes. Developmental trajectories of CT in ASD were mostly characterized by decreased cortical thinning during early adolescence and increased thinning at later stages, involving mostly frontal and parietal areas. Such changes were associated with ADOS scores. The curvature of the trajectories estimated from the quadratic model was the most accurate and sensitive measure for detecting ASD. Our findings suggest that under the context of longitudinal changes in brain morphology, robust detection of ASD would require three time points to estimate the curvature of age-related changes.

neuroscience↗

bioRxiv2019

Multiple constrained minimum variance beamformer (MCMV) performance in connectivity analyses

Functional brain connectivity is increasingly being seen as critical for cognition, perception and motor control.Magnetoencephalography and electroencephalography are modalities that offer noninvasive mapping of electrophysiological interactions among brain regions, yet suffer from signal leakage and signal cancellation when estimating brain activity. This leads to biased connectivity values which complicate interpretation. In this study, we test the hypothesis that a Multiple Constrained Minimum Variance beamformer (MCMV) outperforms the more traditional Linearly Constrained Minimum Variance beamformer (LCMV) for estimation of electrophysiological connectivity. To this end, MCMV and LCMV performance is compared in task related analyses with both simulated data and human MEG recordings of visual steady state signals, and in resting state analyses with simulated data and human MEG data of 89 subjects. In task related scenarios connectivity was estimated using coherence and phase locking values, whereas envelope correlations were used for the resting state data. We also introduce a novel Augmented Pairwise MCMV (APW-MCMV) approach for signal leakage suppression in resting state analyses and assess its performance against LCMV and more conventional MCMV approaches. We demonstrate that with MCMV effects of signal mixing and coherent source cancellation are greatly reduced in both task related and resting state conditions, while in contrast to other approaches 0-and short time lag interactions are preserved. In addition, we demonstrate that in resting state analyses, APW-MCMV strongly reduces spurious connections while better controlling for false negatives compared to more conservative measures such as symmetrical orthogonalization.

neuroscience↗