Background

To date, 3D biological models with varying complexity are used to recapitulate the physiological features of animal and human nervous system circuitry, allowing for the study of physio-pathological brain activity [1–3]. Recent advances in stem cell technology have introduced the possibility of producing region-specific neural organoids from human-induced pluripotent stem cells (hiPSCs). Animal models offer fully mature, structurally accurate brain architectures, which are crucial for examining circuit functions and understanding how diseases affect the neuronal electrical activity [4,5]. In contrast, human-derived neural organoids and the simpler brain spheroids, which can be derived from human neural progenitors but lack the structural and regional organization typical of organoids, can replicate unique human features, thereby reducing challenges in translating research findings to humans [6–8]. Furthermore, as these models can be maintained in culture for extended periods (up to years), they provide opportunities for detailed neurodevelopmental studies. Additionally, neural organoids can generate diverse patterned cell types, such as cortical, thalamic, and spinal cord cells, and achieve greater functional maturation than 2D cultures [9,10], though not equivalent to adult human brains. This enables the study of complex, in vivo–like brain network formation [11–15].

Several techniques exist to study neuronal activity in 3D models, including patch-clamp recordings, calcium imaging, and multielectrode arrays (MEA). In particular, MEAs represent the ideal technique for characterizing the electrophysiological activity of neuronal networks in these models, enabling simultaneous non-invasive recordings from multiple locations within the network, maintaining single-cell resolution, and providing excellent spatial resolution with the recent introduction of microchip-based high-density MEAs (HD-MEAs) [16–18]. For these reasons, HD-MEA devices are already in use with several 3D models, such as acute slices [18–21] and organoids [2,22–24]. However, planar probes commonly used for electrophysiological recordings are insufficient to exploit the complex network architecture that provides the main advantage of using these models. For instance, the chip is in contact only with the surface of the sample, interfacing the electrodes solely with the outermost cell layers. As a result, it fails to capture the activity occurring within the deeper, denser, and more interconnected layers. Even more critical, signals may be dampened by the presence of a dead cell layer in acute slices or by scaffolding materials, such as Matrigel, that are required to embed organoids in some protocols [25–27]. Another important aspect concerns vitality, which is compromised when the tissue is pushed against the chip. This dramatically reduces the recorded surface supply of oxygen and nutrients required to sustain the sample's viability. To fully harness the potential of HD-MEA in combination with 3D tissues, a recent technological breakthrough is represented by the integration of 3D electrodes that can penetrate the tissue and record neuronal activity from inner layers. In addition, these new devices have been equipped with a microstructure at the base of the 3D electrodes that forms microchannels below the tissue, thereby enhancing viability and accelerating pharmacological testing [28].

This methodology is well-suited for detailed investigations of neural network dynamics within 3D biological models, opening new opportunities to understand the relationship between structure and function in preparations, such as brain slices and organoids, that display a significant circuit architecture. This makes the technique ideal for a range of applications across multiple areas of neuroscience, including neurodevelopment, neuropharmacology, and neuropathology. For example, the technique's sensitivity would allow the uncovering of subtle modifications in neuronal firing and network dynamics in disease models with altered neuronal excitability and anatomical connectivity. This aspect is particularly relevant for characterizing organoids' functional network connectivity and its modification during development and in pathological models. Moreover, another promising field of application is certainly the testing of neuroprotective or neurotoxic compounds, which will benefit from the high efficiency of activity recordings from 3D models that more closely recapitulate in vivo conditions and can be maintained in more physiological environments than 2D models. Overall, this methodology provides a versatile platform for researchers working with diverse 3D neural models to investigate physiological mechanisms and disease-related dysfunctions with high reproducibility.

Indeed, the introduction of this new generation of 3D HD-MEA devices addresses the limitations that previously hindered the effectiveness of recordings by using planar devices. However, their unique surface, made up of micropillars approximately 90 μm tall, arranged in a 64 × 64 matrix with an electrode pitch of 60 μm, combined with the complexity of 3D biological models, requires specific adjustments to the experimental preparation and the application of this technology, which are crucial to ensure high-quality recordings. This protocol offers a comprehensive overview of preparing, mounting, recording, and analyzing acute 3D HD-MEA experiments from various 3D models: cerebellar, prefrontal cortex, and cortico-hippocampal acute slices, brain spheroids, and cortical and spinal cord organoids. For clarity, the paper is organized into two independent sections, one for slices and the other for spheroids and organoids. The protocols reported here are intended to enhance researchers’ ability to record functional activity from different types of 3D biological samples.