5. DISCUSSION
The importance of shifts in E/I balance across multiple brain regions in the development of ASD pathology has been recognized for many years \cite{RN9,RN22,RN130}. In the hippocampus, accumulating data now links synaptic changes to the abnormal connectivity and E/I imbalances observed in vivo in ASD model animals. Nevertheless, it is important to note that synaptic alterations in ASD can vary across the different models, brain regions, and age groups under study. Regardless, it remains essential to uncover the key mechanisms necessary for initiation and maintenance of E/I shifts throughout the lifespan, as well as comprehend how those shifts impact brain connectivity and function. As the models discussed here illustrate, these E/I changes can relate to the availability of receptors at the synapses \cite{RN131,RN31,RN9}, to changes in inhibitory circuits \cite{RN60,RN59}, or alterations in morphology \cite{RN132}; many routes to achieve a similar functional outcome.
In understanding the development of ASD, it is crucial to consider, not only neurons, both also the role glia play in these disorders. Accumulating evidence of morphological, transcriptomic, and functional changes in hippocampal astrocytes across rodent models of ASD has begun to clarify the role they play in the shaping synaptic and cognitive phenotypes related to the disorder. Astrocytes play an essential role in sculpting neural circuits by coordinating synapse formation and function, promoting neuronal survival, and guiding axonal growth in the developing brain \cite{RN67,RN68,RN69}. Recent studies have also demonstrated that astrocytes actively participate in pruning dysfunctional synapses \cite{RN133,RN73} and regulate synaptic transmission through gliotransmitter release and removal, forming tripartite synapses in the hippocampus \cite{RN134,RN70}, thus abnormal astrocytic function also often leads to a disruption of E/I balance. A recent study reported the critical importance of astrocytes in ASD pathology by transplanting astrocytes derived from ASD individuals into the hippocampus of newborn mouse pups (P1 to P3) \cite{RN135}. The transplanted mice exhibited distinct repetitive behaviors and memory impairment resembling ASD, along with exaggerated astrocytic Ca2+ fluctuations in vivo, as well as reduced long-term potentiation, neuronal network firing, and spine density in vitro. In addition, microglia have been well documented to be primarily responsible for synapse pruning, a crucial process for maintaining E/I balance in the developing brain \cite{RN73,RN74}. However, once again, the nature of these glial alterations varies depending on factors such as the specific model, sex, and age under consideration. Nonetheless, current data make clear that glial cells are not merely passive victims of neurocentric pathology, but rather play an active role in the orchestration of ASD pathology.
While shifts in E/I balance can have a plethora of consequences during development and in the adult brain, as well as on single cell and network activity, several in vivo measures of circuit function, including coupling between slow and fast oscillations, modulation of spiking by the local field potential, and the coordinated activity of ensembles of neurons, both by theta during movement and by ripples during rest, provide temporally sensitive and precise readouts of the consequences of disruptions in the E/I network. While it is well accepted, both experimentally and theoretically, that shifts in E/I balance in either direction- a more inhibited or a more excited network- can have disruptive effects on circuit function and cognition, this is also true on the level of in vivo activity patterns. The most common alterations across the models reviewed were shifts in the fine patterns of temporal coordination of hippocampal activity. This perhaps is unsurprising given the need for a well-tuned E/I balance to achieve precision of spiking at the millisecond, or even tens of milliseconds, timescale. Interestingly, in models in which single cell properties, like place field size and average pyramidal cell firing rate, were assessed, very few changes were noted. However, when population activity was examined, more significant changes were observed, be the ensemble level, with discoordination observed in the Fmr1 mice \cite{RN114}, the truncation of replay sequences in the scn2a-/+ mice \cite{RN129} or the ensemble level hypersynchrony observed in the Mecp2 mice \cite{RN119}.
E/I shifts in network activity may also appear in a state-dependent fashion- both in terms of the animal’s cognitive state, i.e. memory encoding or recall, and in terms of the physiological state of the circuits, such as an active theta-dominated state or the large-irregular activity and SWRs that accompany quiet wakefulness and slow wave sleep. While a well-balanced E/I network is important for all, mechanistic differences in the cell-types involved, the dominant excitatory inputs, and the timescales involved may reveal specific dysfunctions in select models. While there is clearly no singular physiological endophenotype of ASD across the mice assessed, there are recurrent themes in the data. A loss of oscillatory coordination and disorganization and inflexibility of population activity, both during rest and after learning, stand out as hallmarks of hippocampal dysfunction in ASD model mice. Moving forward the field would benefit from the application of more standard protocols to facilitate comparisons between the various models, as well as the application of high-density recording and/or imaging approaches to assess the impact of ASD risk mutations on the levels of coordination across the population.