The claustrum is a synaptic hub in the amniotic brain thanks to its far-projecting excitatory neurons supporting reciprocal claustrocortical connections. Its sheet shape and deep location between the insular cortex and the striatum have made its study complicated and its exact function remains elusive. However, existing hypotheses are numerous and all point to crucial processes such as perceptual binding (Crick and Koch, 2005), salience detection or attention (Atlan et al. 2018), but also cognitive control (Madden et al. 2022) and modulation of cortical networks dynamics like sleep slow-wave synchronization (Narikiyo et al. 2020). When studying mouse brain horizontal slices containing both the claustrum and the insula, we discovered a synchronized 0.3-1Hz oscillation in claustral principal cells during bath-application of an acetylcholine receptor agonist. This rhythmic network activity could be an essential component of the claustrum’s function, especially regarding cortical synchronization hypotheses. We aim at exploring this oscillation’s cellular, synaptic and neuromodulatory mechanisms in vitro and in computo, using electrophysiology, optogenetics, and modeling tools. Ongoing exploration of synaptic activity in each claustral neuron subtype during cholinergic-induced oscillations shows that principal cells receive tonic excitatory inputs and rhythmic chloride inhibitory inputs, suggesting that rhythmic GABA-A inhibition drives the oscillation of principal cells. Recordings in an interneuron subset exhibiting intrinsic rhythmic properties, together with Adaptive Exponential Integrate-and-Fire (AdEx) modeling, suggest that somatostatin (SST)-positive interneurons could be the oscillator driving principal cells’ rhythmic activity. In parallel, we found that parvalbumin (PV)-positive interneurons receive rhythmic excitatory inputs, presumably from principal cells, and oscillate in phase with the latter. Calcium imaging will allow examining the extent of the oscillation synchronization while AdEx modeling will help us complement our experimental work by eventually creating whole claustrocortical network models to test further hypotheses.