Decoding the Hippocampal (and cortical) oscillatory Complexity to predict behavior and its pathological alterations – HippoComp

Coherent oscillations of neuronal activity, ubiquitous across brain scales, would play core roles in information transfer and processing. Oscillatory coordination is implicated in the formation of sensory or behavioral representations, or in the flexible routing of information between neuronal populations as suggested by both experiments and theory. In the hippocampal formation, theta (?) oscillations during exploration behavior provide a reference for phase-coding of place sequences while coherence in different gamma (?) frequency sub-bands temporally segregates information originating from different sources. Furthermore, behavior also correlates with dynamic coherence between hippocampus and other cortical regions. Altogether, these and other findings establish ?-? oscillatory dynamics in the hippocampus and its broader networks as a signature for computations underlying spatial navigation and, more generally, episodic and spatial memory.
Beyond the healthy brain, these key cognitive functions are particularly disrupted in pathologies like neurodegenerative diseases and dementias, with disruptions of hippocampal oscillatory dynamics correlating with behavioral and cognitive degradation before any detectable classical histopathological alterations in animal models of AD. In humans as well, hippocampal functional connectivity is disrupted at early preclinical stages in relation with cognitive decline, revealing alterations of long-range oscillatory coordination linked to pathology. It remains an open question whether these changes in coordinated neural dynamics are a consequence of neurodegeneration or one of its causes. Beyond the advancement of neurodegeneration, it is likely that the degradation of structured neural dynamics has a direct mechanistic impact on the emergence of functional impairments. This hypothesis has major implications. First, designing neuromarkers which faithfully track deviations from “healthy” neural dynamics may lead to earlier predictions about the evolution of deficits. Second, being able to intervene on dynamics could eventually slow or even invert the evolution of cognitive and behavioral impairments, by acting on the “software” –local neural codes and global routing/interfacing processes (mediated by multi-scale oscillatory coordination, refitted to be closer to non-pathological trajectories)– rather than on the “hardware”, i.e. the integrity of structural circuits.
HippoComp will rely on a synergy between electrophysiological recordings (single units and LFPs at the meso-scale of hippocampus; EEG at the macro-scale of whole cortex), machine-learning approaches and non-linear time-series and temporal multiplex network analyses to fulfill the following general aims: i) deepen our general understanding of how complex oscillatory dynamics at multiple scales mediate coding and computations relevant to behavior and memory function; ii) to identify alterations of oscillatory-mediated information processing that translate since early AD pathology stages into functional impairments; and, iii) to show that manipulations (visual gamma stimulation, vGENUS) that preserve/restore function in AD do so by acting to restore a “healthy” oscillatory dynamics (and the associated organization of neural computations).
Explicitly pursuing these aims requires a radical redefinition of analyses approaches to the study of the relations between oscillations, oscillatory coherence and behavior and cognition. Notably, we will shift emphasis from well-behaved average oscillatory properties to the haphazard and apparently random-like spectrum of transient oscillatory events, ongoing “in real time” in parallel with actual behavior, to show that such complexity, usually averaged out as noise, may be a needed resource for efficient neural information processing.