Abstract
Schlingloff D, Freund TF, Hangya B, Gulyás A. J Neurosci. 2025;12;45:e0500252025 Diverse network oscillations, thought to represent different information processing modes of cortical networks, are accompanied by synchronous neuronal activity at various temporal scales. Sharp wave-associated ripple oscillations, supporting memory consolidation in the hippocampus, are among the fastest physiological oscillations characterized by strong interneuronal synchrony. In contrast, when hippocampal activity turns epileptic, pathological fast-ripple oscillations appear. The distinction of the two oscillations is diagnostically relevant; however, how differential mechanisms of the same network generate the two activities is not well understood. Here, we addressed this question using an in vitro hippocampal model that allowed targeted recording of cell types and local pharmacological manipulations in mice of either sex. We showed that inhibition did not contribute to current and rhythm generation of fast-ripples, unlike physiological ripple oscillations. Instead, pathological fast-ripples emerged when perisomatic inhibition from parvalbumin-expressing basket cells collapsed and depended on the quasi-simultaneous onset of stereotypical pyramidal cell (PC) bursts, leading to pseudosynchrony. This was accompanied by a loss of spatial coherence. In epileptogenic conditions, deep CA3 PCs selectively ramped up their burst activity before fast-ripple onset, while normally nonbursting superficial PCs acquired burst capability. These results point to PC pseudosynchrony as the underlying mechanism of fast-ripples, with differential contribution of known PC types.
Commentary
Fast synchronous activity that comprises network oscillations (ripples) occurs in several brain regions and underlies both normal and abnormal processes. 1 Two important forms of ripples are physiological or sharp wave ripples (SWRs, 150–250 Hz) that coordinate memory consolidation in the hippocampus 2 and pathophysiological “fast” ripples (fRIPs, 250–600 Hz) that correlate with local epileptic activity in the hippocampus. 3 Features of fRIPs can help localize and characterize seizure onset and propensity, endowing them with translatable clinical utility, particularly related to the epileptogenic zone of interest for epilepsy surgery.3,4
Ripples of different frequencies are likely generated by distinct but perhaps overlapping mechanisms. Nonepileptic SWRs emerge from the highly connected excitatory network comprising the hippocampal CA3 region, with a prominent role of synchronous rhythmicity induced by firing of deep parvalbumin-expressing inhibitory basket cells (PVBCs) onto excitatory pyramidal cells. However, the role of inhibition related to PVBC activity in producing epileptic fRIPs is less well known. At first glance, it might be assumed that the interictal-to-ictal transition involves reduction of local inhibition, increased local excitation, or both. The role of local inhibition in the generation of fRIPs is a critical knowledge gap. The present paper uses hippocampal slices to study differences between SWRs and fRIPs. 5
The authors employ several technical modifications to examine their hypothesis that the transition from the interictal to the ictal state involves complex dynamic changes in neural networks. 6 First, they used a dual flow system whereby the flow of artificial cerebrospinal fluid could be separately controlled above and below the slice in their submerged chamber, to allow optimal metabolic supply to both the superficial and deeper cell layers. Second, they induced fRIPs by bathing hippocampal slices in elevated potassium (8.5 mM K) to produce epileptiform bursting. Third, they used “loose” patch clamping to monitor neuronal firing as well as to facilitate biocytin filling. Loose patch recording is basically a cell-attached single-cell extracellular recording (without the need to form a tightly sealed patch as in standard patch-clamp recording) to register neuronal firing without rupturing the neuronal membrane, such as in single-unit recording. A potential limitation is that, unlike regular intracellular (whole-cell) patch clamp, loose patch clamping cannot provide detailed information about membrane properties. Finally, pharmacological blockers of action potential firing or inhibitory synaptic currents were puffed locally over small segments of the hippocampal CA3 region, allowing precise control of firing or synaptic currents across the extent of the neuron. The sodium channel blocker tetrodotoxin (TTX) was puffed onto the slice to suppress action potential firing, while exposure to the GABA-A receptor antagonist, gabazine, diminished inhibitory synaptic currents. This technique of precise focal drug application assumes that the entire neural circuit (the whole slice) is not exposed to the blockers, as they demonstrated using distant control electrodes.
In the baseline, nonepileptic condition, SWRs at ∼200 Hz were recorded. Local TTX applications to small areas of CA3 blocked action potentials, but TTX doses were low enough to avoid blocking all transmitter release, and inhibitory synaptic activity remained intact. In this situation, SWR power was reduced modestly. When gabazine was puffed onto the area to decrease GABA-A receptor-mediated synaptic inhibition, ripple power fell significantly, supporting the hypothesis that inhibitory synaptic currents contribute strongly to SWRs. The main source of inhibitory synchrony in this setting is deep PVBCs that normally produce synchronized inhibition onto pyramidal neurons.7,8
When epileptic burst activity (interictal discharges) was induced by bathing the slices in elevated potassium (8.5 mM K), fRIPs were produced and could be markedly suppressed by local TTX application. Addition of gabazine did not further affect fRIPs power, leading to the conclusion that epileptic fRIPs are generated by action potential firing but are relatively unaffected by inhibitory synaptic currents. In this condition, synchronous firing of PVBCs is diminished, in accordance with the loss of perisomatic inhibition. Pyramidal neurons fire in bursts when inhibition diminishes; when PVBCs stop firing (are desynchronized), fRIPs occur.
Additional experiments provided evidence for the existence of 2 distinct populations of pyramidal cells—deep and superficial—that exhibit somewhat different firing characteristics and properties. Deep CA3 pyramidal neurons increase their bursting before fRIPs, while superficial pyramidal neurons, which do not ordinarily burst, become bursters. The authors denote this situation as “pseudosynchronous” action potential bursting of pyramidal cells, a situation favoring epileptic firing.
The authors concluded that when PVBC stop firing (no more inhibition), fRIPs appear; that is, inhibition does not contribute to pathological fRIPs firing, and when local inhibition is reduced, fRIPs replace physiological ripples. When TTX is used to block action potentials, there is a decrease in ripple power, more so in interictal discharges/fRIPs than in the nonepileptic SWRs. Interestingly, as SWRs transitioned into fRIPs, the fRIPs began at the exact time that PVBCs stopped firing.
In summary, this paper provides strong evidence that inhibition, at least in this acute high K+ model, contributes to physiological SWRs but not to epileptic fRIPs. Instead, fRIPs emerge when inhibition from deep PVBC “collapses” due to depolarization block, leading to “pseudosynchronous” action potential bursting of pyramidal cells. The findings point to distinct physiological mechanisms underlying each type of ripple, adding to the possibility of clinical utility of ripples for the localization of epileptic pathology. That goal—translatable clinical correlation to aid surgical management—has become more feasible as knowledge about ripples has advanced over the past few decades, facilitated by technical innovations to study these phenomena on small scales and over limited time spans. The discovery that ripples exist not only in the hippocampus but also in widespread neuronal regions, including the neocortex, amygdala, piriform cortex, lateral septum, and others 1 speaks to their importance in both cognitive processes and epilepsy. Furthermore, although ripples have been studied most avidly in temporal lobe epilepsy and its animal models, their existence in other epilepsies, such as posttraumatic epilepsy and developmental epileptic encephalopathies, such as infantile spasms and focal cortical dysplasias, broadens their potential importance in diverse forms of neuronal hyperexcitability. Indeed, inhibition may play a more prominent role in regulating the various high-frequency oscillations at younger ages. 9 Remaining challenges include investigations of mechanisms governing neuronal synchrony, ictogenesis, and epileptogenesis in these alternative areas and syndromes. Many of these questions may be approachable by combining in vitro models with ex vivo tissue from surgical resections. 10
