Datensatz

Zusammenfassung

Sharp-wave ripples (SPW-Rs), transient episodes of neural activity combining a sharp wave of dendritic depolarization and a high-frequency oscillation, are a major feature of the cortico-hippocampal communication. Experimental evidence relates these episodes to offline consolidation of memory traces. However, the circuitry dynamics that governs these episodes remains poorly understood. Using multi-site extracellular recordings of the hippocampal CA1, we have previously provided evidence for the existence of differentiated SPW-Rs, whose LFP signatures come in four types that suggest a dynamical coupling between SPWs and ripples[1] (Figure 1a). Here we develop a methodology to extract dynamical components from peri-event activity of the CA1 stratum radiatum (SR) and stratum pyramidale (PL). We hypothesize that SPW-Rs can be approximated by an instantaneous linear superposition of independent sources with different spectral signatures that account for their dynamics. Inspired by the second order blind identification (SOBI) algorithm[2], our approach estimates the spatial spread and spectral signature of these dynamical sources by implementing a Joint Approximate Diagonalization of the cross-spectral density matrix between recording sites. Our analysis reveals that dynamical components with distinct spatial signature can be isolated: a broad-band-spectrum component spanning the gamma and ripple bands, originates in SR; and a second component, with clear spectral peak in the ripple band (80-200Hz), originates in PL (Figure 1b c). Differences between SPW-R subtypes are reflected in the spectra of the second component (ripple). We suggest that differences between SPW-Rs are related to local recurrent dynamics within PL, endowing both pyramidal cells and interneurons. Our preliminary findings illustrate the convenience of our approach for extracting meaningful dynamical components from multi-site LFPs during neuronal events and for unraveling their underlying mechanisms.