Human induced pluripotent stem cell (hiPSC) - derived neurons offer the possibility of studying human-specific neuronal behaviors in physiologic and pathologic states in vitro . However, it is unclear whether these cultured neurons can achieve the fundamental network behaviors that are required to process information in the human brain. Investigating neuronal oscillations and their interactions, as occurs in cross-frequency coupling (CFC), is potentially a relevant approach. Microelectrode array culture plates provide a controlled framework to study populations of hiPSC-derived cortical neurons (hiPSC-CNs) and their electrical activity. Here, we examined whether networks of two-dimensional cultured hiPSC-CNs recapitulate the CFC that is present in networks in vivo . We analyzed the electrical activity recorded from hiPSC-CNs grown in culture with hiPSC-derived astrocytes. We employed the modulation index method for detecting phase-amplitude coupling (PAC) and used an offline spike sorting method to analyze the contribution of a single neuron's spiking activities to network behavior. Our analysis demonstrates that the degree of PAC is specific to network structure and is modulated by external stimulation, such as bicuculine administration. Additionally, the shift in PAC is not driven by a single neuron's properties but by network-level interactions. CFC analysis in the form of PAC explores communication and integration between groups of nearby neurons and dynamical changes across the entire network. In vitro , it has the potential to capture the effects of chemical agents and electrical or ultrasound stimulation on these interactions and may provide valuable information for the modulation of neural networks to treat nervous system disorders in vivo .
Significance: Phase amplitude coupling (PAC) analysis demonstrates that the complex interactions that occur between neurons and network oscillations in the human brain, in vivo , are present in 2-dimensional human cultures. This coupling is implicated in normal cognitive function as well as disease states. Its presence in vitro suggests that PAC is a fundamental property of neural networks. These findings offer the possibility of a model to understand the mechanisms and of PAC more completely and ultimately allow us to understand how it can be modulated in vivo to treat neurologic disease.