Spontaneous network activity in the form of Up and Down states plays an important role in the function of neural circuits and reflects intrinsic connectivity. Such cortical dynamics are shaped by genes, experience and intrinsic cellular properties and form the substrate upon which external stimuli and neuromodulation will impinge to determine cortical responses.

The scope of this work was to build a methodological framework for assessing developmental changes in intrinsic cortical activity patterns obtained from LFP recordings of spontaneously active mouse brain slices.

To overcome the complexity seen in the recorded signals, which are characterized by a high variability in Up-state waveforms, we resorted to a representation that preserves the dynamic invariants of the underlying network and built a pattern-analytic scheme that includes three pipelined stages.

In the first stage, upstate waveforms are represented as dynamical trajectories and similarities are quantified by means of a non-parametric multivariate statistical test. Subsequently, for each LFP recording a prototype is extracted among all the available up-state traces based on an algorithm that selects the most typical event in the space of trajectories. In the second stage, all the available prototypical trajectories are brought to a common space and compared against each other. Using the associated labels reflecting the age of the animal, we mine the representative trajectories with morphological characteristi cs specific for each age group. In the third stage, the up-state waveforms corresponding to the representative trajectories are presented in an orderly fashion that reflects the spectrum of variations related with ageing.

The adopted scheme underlined the utility of in-vitro Up states as an index of normal cortical development and maturation and potentially as a neurophysiological biomarker (endophenotype) of neurodevelopmental disorders, paving the path for a better understanding of their underlying cellular mechanisms.

Dimitrios A. Adamos[1]*, Nikolaos A. Laskaris[1], Pavlos Rigas[2], Charalambos Sigalas[2] and Irini Skaliora[2]

[1] Neuroinformatics Group, Aristotle University of Thessaloniki
[2] Neurophysiology Laboratory, Biomedical Research Foundation of the Academy of Athens

Presented in AREADNE 2014

Poster will be shortly available here

Post to Twitter Post to Yahoo Buzz Post to Delicious Post to Digg Post to Facebook Post to Google Buzz Post to LinkedIn Post to Slashdot Post to StumbleUpon Post to Technorati