What actually happens during a ripple
A sharp-wave ripple is a short, synchronized event in the hippocampus: a large slow electrical wave with a fast ripple oscillation riding on top of it. During these events, small groups of neurons that were active together earlier fire again in a compressed sequence. The idea is that the brain is rehearsing a recent pattern while it is offline, so the memory can be strengthened without the original experience being present.
The replay is not random. In mice learning a new track, place cells (neurons that fire at particular locations) that took part more in awake synchronized events were reactivated more strongly during the post-experience offline events [1]. When the awake ripples were disrupted, the experience-driven increase in how much individual neurons contributed to later offline events was blocked [1]. That suggests the awake events help tag which neurons get folded into the later replay.
The ripple's content is partly set from outside the hippocampus
Ripples are not generated in isolation. Silencing the medial entorhinal cortex, a major input route from the neocortex into the hippocampus, reduced how often ripples occurred, while barely changing their size or duration [2]. The same manipulation changed which CA1 neurons fired during ripples, and that change reversed as soon as silencing stopped [2]. The authors interpret this as the entorhinal cortex supplying both a steady enabling signal that helps ripples happen and a transient signal that shapes what is replayed [2].
This matters for the consolidation question because it means the hippocampus is not replaying a fixed tape. The content of the rehearsal can be influenced by cortical input, which is one route by which a memory could be shaped toward what the wider brain currently needs.
Does the same process work for non-spatial memory?
Most of the direct evidence concerns spatial memory. The mouse study above used place cells and a novel run, so its findings are about spatial experience [1]. The entorhinal-cortex work also does not test non-spatial memory [2]. So the claim that ripples consolidate recently formed cell assemblies is well supported for spatial material and only weakly extended beyond it.
There is one suggestive human result. In people, hippocampal ripple rate was higher during an offline rest period, and ripples were linked to how non-spatial conceptual tasks drew on a grid-like schema, a structured way of organizing information [3]. That is consistent with ripples helping to integrate non-spatial experience offline, but it is about aligning new information to an existing framework, not about consolidating a freshly formed assembly of cells. The abstract does not establish that cell assemblies are replayed and stabilized for non-spatial memories.
What the picture adds up to
Put together, the evidence supports a two-stage view for spatial memory: awake synchronized events mark which neurons belong to a new ensemble, and offline ripples then replay and strengthen that ensemble [1], with cortical input helping to time ripples and shape their neuronal content [2]. For non-spatial memory, the human ripple data point toward offline integration of concepts [3], but the specific step of consolidating a recently formed cell assembly has not been demonstrated there.
A further open point is how ripples get started in the first place. Work on hippocampal slices indicates that input from the perforant path, a fiber bundle carrying cortical signals into the hippocampus, can influence the initiation of ripple activity, and that cell-assembly reactivation is discussed as a route to cortical memory consolidation [4]. That is consistent with the cortical-control theme, but it comes from an in vitro preparation and does not settle the non-spatial question.
Sources used in this answer
Awake hippocampal synchronous events are incorporated into offline neuronal reactivation
In mice, place cells that participate more in awake synchronous events are more strongly reactivated during post-experience offline events, and disrupting awake sharp-wave ripples blocks the experience-induced increase in neuronal contributions to those offline events.
Hippocampal sharp-wave ripples and their spike assembly content are regulated by the medial entorhinal cortex
Silencing the medial entorhinal cortex reduces the incidence of hippocampal sharp-wave ripples and alters the neuronal composition of CA1 firing during ripples, with the content change reversing immediately after silencing.
Human hippocampal ripples predict the alignment of experience to a grid-like schema
In humans, hippocampal ripple rate is higher during an offline rest period and ripples are linked to the alignment of non-spatial conceptual experience to a grid-like schema.
Influence of perforant path synaptic excitation on the initiation of hippocampal sharp-wave ripple activity in vitro
In hippocampal slices, perforant path synaptic excitation influences the initiation of sharp-wave ripple activity, and ripples are discussed as reactivating cell assemblies for memory consolidation to the neocortex.