hide
Free keywords:
-
Abstract:
Microelectrode arrays are a privileged recording modality to study neural processes with a very fine spatial and
temporal resolution. They capture the activity of small populations and permit assessment of synergistic interactions
between cells. Patterns of rhythmic ongoing activity are of particular interest because they reflect the
intrinsic dynamics of neural populations and the way such dynamics may optimize the processing of incoming
information. In this study, we identify the various coherent spatio-temporal patterns of rhythmic activity occurring
across time using a two steps approach. First, signals were bandpass filtered in a relevant frequency band and
subsequently Hilbert-transformed. Second, the complex patterns of activity occurring across time were clustered
using a graph cut algorithm based on a phase shift invariant similarity measure. This invariance is a key-property
of our approach to isolate wave propagation phenomena. We apply our method to Local Field Potentials recorded
in the inferior convexity of the Prefrontal Cortex (icPFC) in two anesthetized macaques using a multi electrode array.
We found a dominant travelling wave pattern in the beta band (15-25Hz), propagating along the ventral-dorsal
plane, emerging and vanishing across time both in the absence of visual stimulation (spontaneous activity) and
during binocular stimulation with movie clips. By computing mutual information, we showed that the amplitude of
this wave actually carries sensory information during the presentation of several movies. Altogether, our analysis
provides evidence for travelling wave phenomena reflecting the distributed computation in icPFC, which is known
to be involved in higher order sensory processing. More generally, our approach enables the unsupervised analysis
of the complex spatio-temporal neural dynamics in ongoing signals, providing key information to understand
cooperative mechanisms in spatially distributed neural populations.
