Coordinated emergence of hippocampal replay and theta sequences during post-natal development

Abstract

Hippocampal place cells encode an animal's current position in space during exploration[1]. During subsequent sleep, hippocampal network activity recapitulates patterns observed during recent experience: place cells with overlapping spatial firing fields during locomotion show a greater tendency to co-fire ('reactivation') [2] and temporally ordered and compressed sequences of place cell firing observed during wakefulness are reinstated ('replay') [3-5]. Reactivation and replay are thought to be network mechanisms underlying memory consolidation [6-10].Compressed sequences of place cell firing also occur during exploration: during each cycle of the theta oscillation, the set of active place cells shifts from those signalling positions behind to those signalling positions ahead of an animal's current location [11,12]. These 'theta sequences' have been linked to spatial planning [13].Here we demonstrate that, before weaning (post-natal day 21, P21), offline place cell activity reflects predominantly stationary locations in recently visited environments. By contrast, sequential place cell firing, describing extended trajectories through space during exploration ('theta sequences') and subsequent sleep ('replay'), emerge gradually after weaning in a coordinated fashion, possibly due to a protracted decrease in the threshold for experiencedriven plasticity.

Hippocampal Sequencing Mechanisms Are Disrupted in a Maternal Immune Activation Model of Schizophrenia Risk

“…In humans the hippocampus is involved in both episodic memory processes and thinking about the future (Schacter et al, 2007), and theta sequences may therefore have a fundamental role in these processes Jaramillo & Kempter, 2017;Terada et al, 2017;Wang et al, 2015). In support of this hypothesis, the developmental emergence of theta sequences coincides with the maturation of hippocampal memory (Muessig et al, 2019), and increases in hippocampal theta power have been observed in humans undertaking sequential planning tasks (Kaplan et al, 2020).…”

Ripple Band Phase Precession of Place Cell Firing during Replay

“…It has yet to be established whether large populations of grid cells in the deeper layers of MEC exhibit theta phase precession on linear tracks (Hafting et al, 2008), although previous studies have described this phenomenon in open field recordings (Climer et al, 2013;Jeewajee et al, 2013), in contrast to our results on the Z-shaped track. Elsewhere, recent studies have demonstrated a link between theta sweeps during active exploration and enhanced replay of trajectories through that environment (Feng et al, 2015, Drieu et al 2018Farooq et al, 2019;Muessig et al, 2019). Hence, future work should examine the development with experience of coordinated theta phase precession across place cells during active exploration and ripple band phase precession during subsequent replay trajectories.…”

Hippocampal theta sequences in REM sleep during spatial learning

“…Fitted trajectory slopes (Figure S1D) were significantly higher in REM decodes than in wake (wake and REM sleep distributions are different, p < 3.41e-59, rank-sum test), indicating higher virtual decode velocities, which were also mirrored in the absolute decode distances elapsed in these decodes (Figure S1E) and their differences (wake and REM sleep distributions are different, p < 1.46e-47, rank-sum test). Though recent work has provided a fascinating possible role for stationary representations of space in theta decodes (Muessig et al, 2019), we imposed virtual decode velocity cutoffs of greater than 1m/s and less than 10m/s to only analyze non-stationary decodes, or decodes with velocities approaching those reported of sharp-wave ripple replay to avoid potential, though unlikely false negatives.…”

“…This cutoff was used to enforce contiguity of decoding, rather than contiguity of decoded space. In addition, the virtual speed of the decode was restricted to the range of 1 m/s to 10 m/s, to avoid any possibility of stationary decodes, which have been described previously (Farooq and Dragoi, 2019;Muessig et al, 2019) but were beyond the scope of this analysis, as well as decodes that traversed the track faster than a single theta cycle, which were deemed to be artifactual or ripple-like given previous literature. In addition, only the tail ends of all sequence scores were considered, discarding any sequences whose true sequence scores were below the median of the population.…”