The AP is the fundamental electrophysiological unit that underlies neuronal excitability and intercellular communication (49). Generated within a single neuron, it represents a transient, all-or-none electrical event triggered by the orchestrated opening and closing of voltage-gated ion channels. This rapid depolarization and subsequent repolarization of the cell membrane enables neurons to encode and transmit information both within local circuits and across long-range projections. As the most basic measurable signal in electrophysiology, APs are essential for monitoring neural activity, assessing functional connectivity, and interpreting the dynamic interactions between cells in both healthy and disease states. In the context of brain organoids, detecting and analyzing APs is critical for validating neuronal maturity and evaluating the emergence of functional network properties (49-53).

In organoid electrophysiology, APs are typically detected as sharp, transient deflections in extracellular voltage recordings using MEAs or as direct intracellular signals captured via patch-clamp techniques. Raw voltage traces are band-pass filtered (e.g., 300∼3,500 Hz) to isolate high-frequency spike events (41, 54). Then, Individual spikes are identified by comparing their amplitude against a predetermined threshold, which distinguishes neural signals from background noise. This threshold is commonly set at a multiple of the baseline signal’s standard deviation, with values such as 5σ frequently used in brain organoid studies (55, 56).

Quantitative analysis of detected spikes provides key insights into neuronal function and development. Essential parameters include the firing rate (spikes per unit time), the inter-spike interval (ISI), and physical characteristics of spike waveforms, such as amplitude, duration, and full-width at half-maximum (FWHM) (15). Tracking changes in these properties over time serves as a critical indicator of neuronal maturation. For instance, a developmental trajectory study of cerebral organoids (COs) demonstrated that as cultures matured (between days 30 and 64), spike amplitude increased while low-amplitude random signals evolved into higher-fidelity signals—indicative of network maturation (57). Furthermore, assessing how these parameters respond to pharmacological compounds or optogenetic stimulation enables detailed evaluations of neuronal function and network activity, establishing organoids as powerful tools for drug screening and neurophysiological research (41, 54).

As neuronal networks mature, firing patterns transit from sporadic, isolated spikes to organized, high-frequency clusters known as bursts. These bursts, separated by periods of relative quiescence, indicate functional development and emerging synaptic communication (57). Analysis of spike trains to identify and characterize these bursts provides key metrics for evaluating network health and activity.

The mean firing rate (MFR), representing the average number of spikes per second, provides a broad measure of network excitability. However, a deeper understanding comes from analyzing burst dynamics. Neurons in developing networks often exhibit high-frequency burst patterns that reflect the progressive organization of activity. Burst detection is commonly performed using algorithms that define bursts based on spike density and the ISI. For example, bursts may be classified as sequences of at least five spikes with an ISI of less than 100 ms (58). Key parameters derived from burst analysis include the burst rate (bursts per minute), mean burst duration, the number of spikes per burst, and the percentage of spikes occurring within bursts (59).

Tracking changes in these properties over time provides insight into the maturation of brain organoid. The progression from sporadic, random spiking in early-stage organoids to robust, organized bursting in mature organoids reflects the establishment of functional synaptic networks (57). Furthermore, studies integrating optogenetics with high-density MEAs have demonstrated that specific “leader” or “hub” neurons play a pivotal role in network-wide bursts. Their activity can initiate synchronized firing across neuronal populations, underscoring their importance in orchestrating collective activity (54). Observing how burst properties respond to pharmacological or photostimulation allows for detailed assessments of network dynamics, making these analyses valuable in studying neuronal maturation and functional connectivity.

Beyond the spiking activity of individual neurons, the synchronized electrical behavior of a local neuronal population gives rise to LFPs. These composite signals primarily reflect the collective effect of synaptic currents, afterpotentials, and other sub-threshold membrane potential fluctuations within a defined tissue volume (60, 61). LFPs offer a macroscopic perspective on network dynamics and play a crucial role in identifying emergent, rhythmic patterns of activity known as neural oscillations.

The analysis of LFPs typically involves low-pass filtering raw extracellular recordings (e.g., cut-off frequency of 300 Hz). Fourier transforms or wavelet analysis are then used to decompose the signal into its constituent frequency bands, including delta (1∼4 Hz), theta (4∼8 Hz), alpha (8∼12 Hz), beta (13∼30 Hz), and gamma (30∼100 Hz) (41). These frequency bands represent distinct oscillatory regimes critical for various aspects of neuronal computation and network coordination. A landmark discovery in brain organoid research is the emergence of complex oscillatory waves that recapitulate features of early human brain development. Studies by Trujillo et al. 2019 (15) demonstrated that cortical organoids, after several months in culture, generate oscillatory activity resembling the electroencephalogram (EEG) patterns observed in preterm human infants. Initially, these networks display infrequent, simple electrical events, but they progressively develop robust, rhythmic oscillations indicative of functional maturation

One key hallmark of this maturation is phase-amplitude coupling (PAC), in which the phase of a low-frequency oscillation (e.g., theta) modulates the amplitude of a high-frequency oscillation (e.g., gamma). The emergence and evolution of such complex oscillatory dynamics serve as a critical benchmark for functional network development, demonstrating that the organoid has achieved a level of network-level information processing (41, 62, 63). Notably, LFP analysis has provided valuable insights into disease modeling. For instance, in Rett syndrome organoid models, researchers observed a loss of low-frequency and gamma oscillations, replaced by recurring epileptiform-like spikes and high-frequency oscillations (HFOs). These alterations closely mirror electrographic abnormalities seen in patients, reinforcing the translational potential of LFP recordings in neurological disease research (63). Furthermore, LFPs recorded from transplanted organoids demonstrate functional integration, as they respond to sensory stimuli and exhibit phase-locking with host brain oscillations (64). This finding underscores the ability of organoid-derived neuronal circuits to synchronize with biological brain networks, offering promising avenues for regenerative neuroscience and neuroengineering applications.

Functional connectivity describes the temporal relationship between spatially remote neurophysiological events and is a powerful measure for assessing how information is integrated and communicated across a neural network (65). In brain organoids, analyzing functional connectivity is essential for understanding how individual neurons and neuronal populations self-organize into coherent, functional neural circuits (66).

Several analytical methods are used to quantify functional connectivity from MEA recordings. The most straightforward is cross-correlation, which measures the similarity between two spike trains as a function of the time lag between them (41). More sophisticated methods have been developed to account for variations in firing rates. The spike time tiling coefficient (STTC), for instance, is a rate-independent measure that calculates the proportion of spikes in one train that lie within a defined temporal window (e.g., ±50 ms) of spikes in another train (58, 67). Another approach is Correlated Spectral Entropy (CorSE), which assesses synchronization based on the correlation of temporal changes in the spectral content of signals from different electrodes (58). These analyses are used to construct functional connectivity maps, where neurons or electrodes are represented as nodes and the statistical relationships between their activities are represented as edges. Such maps can reveal the underlying topology of the network, including the identification of highly connected “hub” neurons that may act as leaders in initiating network-wide activity (54, 55).

In studies of transplanted organoids, functional connectivity analysis has been crucial for demonstrating integration with the host brain. For example, Wilson et al. 2022 (64) showed that spikes in transplanted organoids become phase-locked to slow oscillations in the host visual cortex, indicating the formation of functional synaptic connections. Giandomenico et al. 2019 (68) utilized a MEA system (MEA1600, Multi Channel Systems) with a 3-D grid MEA (60 electrodes) for extracellular recordings in air-liquid interface cerebral organoids (ALI-COs), inferring functional connectivity from correlated spontaneous activity using the STTC. Their analysis revealed densely connected local networks with highly correlated neuronal activity across both short and long-range connections, suggesting specific spatial patterns of connectivity. Mansour et al. 2018 (69) employed electrode arrays from Innovative Neurophysiology Inc. (16 tungsten contacts) for in-vivo extracellular recordings in transplanted human brain organoids. They assessed neuronal interactions by calculating cross-correlations for neuron pairs, revealing that late-stage grafts (115 dpi) exhibited strong correlated activities and a more active and coordinated neuronal network compared to early-stage grafts. Crucially, their optogenetic experiments further demonstrated functional connectivity from the grafted organoid to the host brain, showing laser-evoked LFP changes in host regions upon stimulating the graft. These quantitative approaches, often complemented by anatomical methods like rabies virus-based synaptic tracing, are essential for characterizing the development and function of neural circuits in organoid models (55).

The whole-cell patch-clamp technique stands as the definitive gold standard for the measurement of the electrophysiological properties of individual neurons within brain organoids. Its outstanding precision allows researchers to measure intrinsic membrane properties, ion channel currents, and synaptic activity of a single cell with sub-millisecond temporal resolution. The core principle involves forming a high-resistance (giga-ohm) electrical seal between a glass micropipette electrode and the neuronal membrane. This configuration provides low-resistance access to the entire cell interior, enabling direct control and measurement of the cell’s membrane potential and ionic currents. This high-fidelity recording is fundamental for validating the functional maturation of neurons derived from human pluripotent stem cells and for dissecting the cellular basis of network activity.

A primary application of patch-clamp in organoid research is to confirm the functional identity and maturity of individual neurons. By injecting current and measuring the voltage response, researchers can determine a neuron’s intrinsic properties, such as resting membrane potential, input resistance, and its capacity to fire APs (70, 71). Numerous studies have successfully employed this method to demonstrate that neurons within human brain organoids progress to a functionally mature state, capable of generating robust, repetitive AP trains upon depolarization (3, 72, 73). For example, Renner et al. 2020 (74) confirmed that cells in their automated midbrain organoids (AMOs) displayed typical neuron-like sodium and potassium currents and could fire APs, validating their differentiation protocol. Similarly, Sakaguchi et al. 2015 (75) used patch-clamp on dissociated cultures from their hippocampal organoids to record voltage-dependent Na＋/K＋ currents and spontaneous excitatory postsynaptic currents (sEPSCs), confirming the generation of functional hippocampal-like neurons.

Furthermore, patch-clamp is indispensable for dissecting synaptic transmission and network integration, particularly in complex assembloid models that fuse different brain regions. Birey et al. 2017 (76) demonstrated that interneurons that migrated from subpallial to cortical spheroids received a threefold increase in synaptic input, mostly excitatory, confirming their functional integration into the cortical network. In a more complex cortico-motor assembloid, Andersen et al. 2020 (77) elegantly used this technique to show that optogenetic stimulation of cortical neurons evoked monosynaptic postsynaptic currents in connected spinal motor neurons, thereby validating the entire circuit’s integrity. The technique’s precision is also critical for disease modeling. Ghatak et al. 2019 (78) leveraged patch-clamp recordings to reveal a phenotype of neuronal hyperexcitability in Alzheimer’s disease organoids, characterized by an increased frequency of sEPSCs. In a model of cyclin-dependent kinase-like 5 (CDKL5) deficiency disorder, Negraes et al. 2021 (79) identified a lower rheobase current and increased Na＋ current density, providing a cellular mechanism for the associated epilepsy. This level of detail is also crucial for assessing the maturation of human neurons in an in-vivo environment. Revah et al. 2022 (80) performed patch-clamp recordings on human organoids transplanted into the rat cortex, revealing that transplanted neurons exhibited more mature electrophysiological properties—including a more hyperpolarized resting membrane potential and higher maximal firing rates—compared to their in-vitro counterparts, demonstrating the powerful influence of the in-vivo environment on human neuronal maturation.

Despite its high resolution, the patch-clamp technique has significant limitations. Its primary drawback is its low throughput, as it permits the recording of only one cell at a time, making it unsuitable for capturing emergent, large-scale network phenomena. Moreover, the technique is highly invasive for 3-D organoid structures. Accessing individual neurons has historically required acutely slicing the organoid, a process which inherently disrupts the native cellular architecture and severs countless network connections (76, 77). To address this, methods for patching in intact, unsliced organoids are being developed, though they remain technically challenging due to the tissue’s opacity, softness, and the small size of the neurons (81).

In conclusion, while the patch-clamp technique is limited by its low throughput and invasive nature, it remains an indispensable tool in brain organoid research. It is not a large-scale screening tool, but rather a precise instrument for detailed mechanistic investigation. It provides the ultimate resolution for assessing single-neuron function, validating cellular maturity, uncovering subtle pathophysiological changes in disease models, and confirming the successful assembly of functional, multi-regional circuits. Its strength lies in providing the foundational “ground truth” data that is essential for interpreting and validating findings from higher-throughput methods such as calcium imaging and MEAs.

While patch-clamp electrophysiology provides unparalleled detail at the single-cell level, understanding the emergent properties of neural circuits requires methods that can simultaneously monitor the activity of large cell populations. Calcium imaging has become a cornerstone technique for this purpose, enabling researchers to visualize the dynamic activity of hundreds to thousands of neurons across a developing network. This optical method functions on the principle that neuronal APs trigger a rapid, transient influx of calcium ions (Ca^2＋) into the cytosol. By using fluorescent indicators that bind to Ca^2＋, these electrical events can be indirectly observed as flashes of light, providing a powerful proxy for neural spiking (82), and spatially, temporally and non-synchronous activity (83). Indicators range from synthetic chemical dyes, such as fluo-4 acetoxymethyl ester (Fluo-4 AM), to genetically encoded calcium indicators (GECIs) like GCaMP, which can be expressed in a cell-type-specific manner for long-term studies (3, 73).

The key advantage of calcium imaging is its ability to provide a panoramic view of network dynamics. This has been essential for characterizing one of the most critical hallmarks of functional maturation in organoids: the transition from sporadic, individual neuronal firing to coordinated, synchronous network-wide events. Early studies by Lancaster et al. 2013 (1) demonstrated spontaneous calcium transients in COs, confirming the presence of active neurons. Building on this, Sakaguchi et al. 2019 (84) elegantly captured the self-organization of synchronized calcium transients in dissociated COs, demonstrating their modulation by neurotransmitters and highlighting the method’s utility for pharmacological screening. This followed earlier work where spontaneous calcium transients in dissociated hippocampal organoid cultures confirmed the generation of a functional neuronal network (75). Similarly, Zafeiriou et al. 2020 (85) identified spontaneous, giant depolarizing potential (GDP)–like synchronized calcium bursts in their bioengineered neuronal organoids (BENOs), a key feature of the developing fetal brain that indicates early network formation.

Calcium imaging has also been instrumental in validating the functional integration of complex, multi-part assembloid systems and transplanted grafts. For instance, Birey et al. 2017 (76) utilized calcium imaging to show that neurosteroids could increase the frequency of spontaneous calcium spikes in subpallial spheroids, demonstrating a functional response to external factors. In more complex cortico-motor assembloids, imaging of GCaMP signals was used to confirm that optogenetic stimulation of cortical neurons successfully triggered downstream calcium transients in connected spinal neurons, thereby validating the entire circuit from cortex to spinal cord (77). Taking this a step further, Revah et al. 2022 (80) performed two-photon calcium imaging on human organoids transplanted into the rat brain, demonstrating that the human graft could be functionally activated by sensory stimuli delivered to the host animal, providing powerful evidence of true in-vivo circuit integration.

The technique is also a powerful tool for disease modeling. For example, Khan et al. 2020 (86) identified a significant decrease in the amplitude of depolarization-induced calcium influx in cortical neurons derived from individuals with 22q11.2 deletion syndrome, linking a highly penetrant genetic risk factor for psychiatric disease to a specific cellular excitability phenotype. Similarly, Negraes et al. 2021 (79) utilized calcium imaging to demonstrate neuronal hyperexcitability in organoids from patients with CDKL5 deficiency disorder (CDD). In models of Rett syndrome, Samarasinghe et al. 2021 (63) showed that organoids with methyl-CpG binding protein 2 (MECP2) mutations displayed epochs of spontaneously synchronized calcium transients, mirroring the epileptiform activity seen in patients. Further demonstrating its utility in high-throughput applications, Renner et al. 2020 (74) showed that aggregate-wide synchronous calcium spikes in their midbrain organoids could be measured on a standard plate reader, establishing calcium imaging as a viable, scalable functional readout for drug screening. LaMontagne et al. 2025 (87) have further expanded on this by showing that reduced graphene oxide (rGO)–poly(viny) alcohol (PVA) nanofibers can mediate light-induced, synchronous calcium handling in cardiomyocytes and neurons, suggesting novel ways to mature and probe organoid networks.

However, the technique is not without significant limitations. The primary trade-off is temporal resolution; the kinetics of calcium transients are inherently slower than the underlying electrical APs, which limits the ability to resolve the precise timing of individual spikes within rapid bursts. The most substantial challenge, especially for intact organoids, is optical access. The dense, 3-D nature of the tissue causes significant light scattering, which typically limits high-resolution imaging to the superficial cell layers, often just 50∼100 μm from the surface (73, 82). This means that the vast majority of neurons deep within the organoid remain unobserved.

In conclusion, calcium imaging serves as an indispensable bridge between single-cell electrophysiology and network-level function. It is the premier method for visualizing the spatial and temporal patterns of activity across large neuronal populations, revealing how synchronous activity emerges and how functional circuits are established and modulated. While challenged by issues of temporal resolution and optical penetration in dense 3-D tissues, its synergy with optogenetics and other recording modalities makes it a powerful and widely adopted tool for interrogating the functional development of human brain organoids in both health and disease.

Human brain organoids have emerged as invaluable 3-D in-vitro models for investigating human brain development, neurological disorders, and drug discovery. While traditional patch-clamp techniques offer single-cell resolution, their limited throughput and invasive nature pose challenges for large-scale network analysis. 2-D MEAs circumvent these limitations by providing non-invasive, high-resolution, and large-scale recording capabilities, making them an indispensable tool for deciphering the complex electrophysiological dynamics of brain organoids. The evolution of MEA technology, particularly high-density complementary metal-oxide-semiconductor (CMOS)–based systems, has significantly advanced our ability to capture intricate neuronal activity, from individual APs to synchronized network oscillations across extended periods. This section reviews recent studies that have leveraged 2-D MEAs to explore the functional properties of human brain organoids, detailing the specific MEA platforms employed and the profound insights gained into organoid neurophysiology.

Several research groups have successfully employed 2-D MEAs to characterize the electrical activity of brain organoids. Sharf et al. 2022 (41) utilized MaxOne (Maxwell Biosystems) to map spontaneous extracellular APs in organoid slices. They found distinct firing patterns, detected theta frequency oscillations, and observed phase-locking between spikes and LFPs, demonstrating how drug treatment altered network connectivity. Similarly, Trujillo et al. 2019 (15) also employed a commercial MEA system to monitor human cortical organoids, noting a consistent increase in spontaneous electrical activity and the emergence of periodic, regular oscillatory network events that transitioned to more complex, synchronous patterns resembling preterm human EEG. Fair et al. 2020 (57) used a 64-channel MEA platform over five months to profile COs. Their longitudinal analysis revealed the maturation of rapid firing rates and network bursting events, correlating with cellular development and activation of the neurotrophin (NTR)/TRK receptor signaling pathway. Yokoi et al. 2021 (88) utilized a planar MEA to analyze frequency components below 500 Hz from cortical organoids, establishing it as a reliable platform for pre-clinical seizure liability assessment. They observed concentration-dependent changes in periodic activity and frequency components in response to convulsants and antiepileptic drugs (AEDs). Renner et al. 2020 (74) developed a high-throughput workflow for human midbrain organoids, incorporating MEA measurements. Their AMOs exhibited spontaneous, highly synchronized neural activity, indicating their potential as a high-throughput screening platform for neurotoxicity and drug effects.

More recently, LaMontagne et al. 2025 (87) measured electrical activity of brain organoids using the Maestro Pro MEA system from Axion Biosystems (CytoView MEA 48 plates). They demonstrated that integrating brain organoids with rGO-polymer nanofibers and light stimulation improved neural function over time, leading to more spikes, activated electrodes per burst, and increased sensitivity to light, along with retinal cell differentiation. Jin et al. 2025 (89) developed a custom 32-channel planar MEA with Pt black electroplating. This MEA, with significantly reduced impedance, enabled robust real-time detection of spontaneous neural activities and pharmacological responses in human induced pluripotent stem cell (iPSC)–derived COs, showing functional network maturation by day 180 and a clear increase in spike rate with potassium chloride (KCl). Müller et al. 2024 (90) investigated epileptiform activity in glucose transporter 1-deficiency syndrome (GLUT1-DS) patient-derived brain organoids using a custom-made MEA biochip and the SpikeOnChip (SPOC) acquisition system (GliaPharm SA, described in (91). They found that GLUT1-DS organoids exhibited distinct epileptiform activity, heightened sensitivity to glucose deprivation (increased spike and burst frequencies), and a higher burst power spectrum density (PSD) in the 4∼40 Hz range compared to healthy controls. Adding to this body of research, Negraes et al. 2021 (79) employed a Maestro MEA system (Axion Biosystems) with 12-well plates to study CDD cortical organoids. Their findings revealed early hyperexcitation and an overly synchronized network in CDD organoids, characterized by changes in MFR and synchrony index over time. Chen et al. 2021 (92) also utilized MEAs to assess neural network activity in human brain organoids exposed to serum, modeling sporadic Alzheimer’s disease. They observed a substantial reduction in neural network activity, including decreased number of spikes, bursts, MFR, and synchrony index, in serum-treated organoids compared to controls. Finally, Popova et al. 2021 (93) conducted MEA recordings using Axion Biosystems’ 64-electrode plates on neuroimmune organoids. Their work showed that microglia transplantation accelerated neural network synchronization, resulting in increased synchronization and frequency of oscillatory bursts, alongside a reduction in synaptic density, suggesting a role for microglia in synaptic remodeling.

In summary, the application of 2-D MEA technology, ranging from commercial platforms like Axion Biosystems’ Maestro Pro and MaxOne from Maxwell Biosystems to high-density research-developed arrays and custom-built systems, has been instrumental in advancing our understanding of human brain organoid function. These studies demonstrate the ability of MEAs to capture complex electrophysiological signals, including APs, LFPs, oscillatory dynamics, and network bursting events. The insights gained from these measurements have elucidated aspects of neuronal maturation, network connectivity, the effects of pharmacological agents, and even disease-specific electrophysiological phenotypes, collectively reinforcing the role of 2-D MEAs as a powerful and versatile tool for modeling human brain physiology and pathology.

Recent advancements in brain organoid research have highlighted the need for high-resolution electrophysiological monitoring systems tailored to the unique properties of these 3-D neural structures. Conventional commercial MEAs, while effective for 2-D neuronal cultures, often lack the spatial resolution and adaptability required for organoid studies. To address these limitations, researchers have developed custom-engineered MEA systems optimized for brain organoid maturation, functional analysis, and transplantation (Fig. 3). These novel platforms integrate high-density electrode arrays, transparent materials for multimodal imaging, and flexible designs to enable long-term recordings and precise stimulation. By refining electrical input parameters, such systems facilitate enhanced neuronal differentiation, synaptic connectivity, and functional integration with host brain circuits post-transplantation. Studies utilizing these custom MEAs have demonstrated superior organoid viability and improved electrophysiological activity. This chapter explores the development and application of these advanced MEA technologies, emphasizing their role in overcoming traditional limitations and advancing brain organoid research toward clinical translation.

Wilson et al. 2022 (64) investigated the functional integration of human cortical organoids transplanted into the mouse brain using a multimodal monitoring approach combining transparent MEAs and two-photon imaging. The study demonstrated vascularization of the transplanted organoid and its ability to generate electrophysiological responses to visual stimuli, matching the activity of the surrounding cortex. Increased multi-unit activity (MUA), gamma power, and phase locking of stimulus-evoked MUA with slow oscillations indicated functional connectivity between the organoid and host brain. Immunostaining confirmed the presence of human-mouse synapses, supporting the organoid’s integration into the sensory network. The study’s strengths lie in its innovative use of transparent electrodes for chronic in-vivo monitoring, providing insights into organoid maturation and connectivity. However, limitations include variability in organoid survival and integration, as well as challenges in distinguishing endogenous versus transplanted neural activity. This work advances the potential of organoid transplantation for studying brain disorders and neural repair strategies.

Suzuki et al. 2023 (94) developed a high-density CMOS-MEA system optimized for brain organoid research, offering significantly improved spatial and temporal resolution compared to conventional MEAs. Traditional MEAs, limited by a small number of electrodes, struggle to record precise single-neuron activity, whereas this system utilizes 236,880 electrodes to enable high-resolution neural recordings at the single-cell level. This advancement allows for detailed analysis of neural network dynamics within brain organoids, facilitating drug response studies and neurological disease modeling. The study demonstrated enhanced detection of network bursts and propagation patterns, providing insights into synaptic connectivity and functional maturation. Strengths of this system include its ability to capture large-scale neural activity while maintaining cellular resolution, making it a powerful tool for in-vitro-to-in-vivo extrapolation. However, challenges remain in data complexity and long-term stability of measurements. Despite these limitations, this CMOS-MEA represents a significant step forward in brain organoid electrophysiology.

Schröter et al. 2022 (95) utilized high-density CMOS-based microelectrode arrays (HD-MEAs) developed at ETH Zurich, overcoming the limitations of traditional passive MEAs, which offered low electrode density (<100 electrodes/mm²) and broad inter-electrode spacing, primarily restricting analysis to population activity. HD-MEAs, with their superior spatiotemporal resolution, enabled efficient de-mixing of neuronal activity and robust single-unit analysis from human cerebral organoid (hCO) slices—a feat challenging with techniques like patch-clamping or calcium imaging. This advanced capability allowed for comprehensive neurophysiological insights by analyzing extracellular AP waveform features, differentiating signals from axons and somatic compartments, and quantitatively inferring AP propagation velocities (averaging 0.41±0.15 m/s) in hCOs for the first time. Furthermore, the study successfully elucidated functional connectivity at the network level, providing a deeper understanding of developing neuronal networks in brain organoids and supporting longitudinal tracking of single neurons. This research establishes HD-MEAs as a versatile platform for the functional characterization of hCOs, significantly advancing our comprehension of brain organoid models.

Osaki et al. 2024 (96) introduced an innovative electrophysiological measurement platform tailored to capture inter-regional network dynamics between connected brain organoids. By engineering a custom polydimethylsiloxane (PDMS)-based MEA device, the researchers successfully linked two COs via reciprocally projecting axonal bundles and recorded synchronized activity from both regions simultaneously. This novel platform enabled the detection of significantly enhanced and spatially complex oscillatory patterns compared to conventional or fused organoids, highlighting the functional significance of long-range axonal connectivity. Importantly, optogenetic stimulation of the axonal tract reliably entrained activity in both organoids and induced robust short-term plasticity, suggesting that these bidirectional axon bundles act as structural and functional conduits for network integration. The combination of high-resolution electrical recording with microfluidic support and optical transparency established a versatile, multimodal system capable of mimicking in-vivo-like inter-regional interactions. While limitations such as variability in organoid maturation and the lack of long-term plasticity remain, this study demonstrates a new approach for modeling large-scale brain circuits and investigating the electrophysiological basis of neurodevelopmental and psychiatric disorders.

Li et al. 2025 (56) investigated the effects of ES on cortical organoid maturation and its impact on transplantation outcomes. Using a MEA, they applied ES to organoids, finding that it enhanced differentiation and maturation via the calcium–calmodulin (CaM) dependent protein kinase II (CAMKII)–protein kinase A (PKA)–phosphorylation of cyclic-AMP response binding protein (pCREB) pathway. ES-pretreated organoids exhibited improved viability, increased synaptic density, and more complex functional activity. When transplanted into the damaged sensory cortex of mice, these organoids demonstrated superior integration with host neural circuits, forming extensive axonal projections and synaptic connections. Electrophysiological recordings confirmed functional connectivity between grafts and the host brain, with ES-treated organoids showing enhanced neural activity and responsiveness to sensory stimuli. Strengths of this study include its novel approach to optimizing organoid transplantation through bioelectric modulation and its comprehensive analysis of structural and functional integration. However, limitations include variability in organoid responses to ES and the need for improved electrode designs for long-term stimulation. This work provides valuable insights into enhancing brain organoid transplantation for neural repair.

Traditional planar MEAs have been instrumental in tracking the maturation of neuronal networks, but their 2-D nature fundamentally limits their application to the complex, 3-D architecture of brain organoids. Planar MEAs can only record from the outermost cell layers that are in direct contact with the substrate, leaving the vast internal volume of the organoid unprobed. This limitation has spurred the development of a new generation of 3-D MEAs designed to non-invasively or minimally invasively interface with the entire organoid, enabling volumetric recording of neural activity and preserving the tissue’s structural integrity for long-term studies (97). These advanced platforms represent a significant technological leap, moving from simple surface recordings to comprehensive, tissue-wide functional analysis by creating flexible, self-assembling, or soft electronic interfaces that conform to, integrate with, or penetrate the organoid structure (Fig. 4).

One prominent strategy involves implementing electrodes that physically wrap around or are enveloped by the growing organoid. This approach is exemplified by the work of Huang et al. 2022 (98), who developed self-folding “shell MEAs” inspired by EEG caps. Their device consists of an optically transparent, self-folding polymer (SU-8) bilayer with integrated gold wiring and conductive polymer (PEDOT:PSS) electrodes. The folding is driven by intrinsic stress generated through differential ultraviolet (UV) cross-linking, allowing the 2-D patterned leaflets to form a 3-D shell. This structure conformally wraps around an organoid (400∼600 μm) to maximize the surface recording area. The key advantage is a scalable fabrication process that yields a device providing more intimate contact than planar MEAs, resulting in a significantly higher signal-to-noise ratio and greater sensitivity in detecting APs, as demonstrated by the increased spike detection upon glutamate stimulation. A limitation is that it still primarily records surface activity. A similar concept was advanced by Park et al. 2021 (42), who engineered 3-D multifunctional mesoscale frameworks (3-D MMFs). These frameworks, implemented by mechanically guided assembly of thin polyimide layers, form a soft, compliant “cage” that gently envelops spheroids, enabling stable, multimodal interrogation via integrated electrical, optical, and chemical sensors. A key feature is that the cage can be reversibly opened and closed by stretching its elastomeric substrate, allowing for gentle insertion and removal of the organoid. This device successfully recorded spontaneous APs and coordinated bursting events propagating across the 3-D surface and was even used to monitor the transection and functional recovery of a neurite bridge between two spheroids, highlighting its unique utility for injury and regeneration models.

To overcome the challenges of recording from freely floating organoids, which is necessary to maintain their native architecture, several suspended mesh platforms have been devised. McDonald et al. 2023 (99) fabricated a hammock-like suspended mesh MEA from polyimide, where the organoid is placed on the mesh and grows to envelop it over time. This design, featuring 61 low-impedance titanium nitride (TiN) microelectrodes, is distinct in its simplicity and stability, enabling non-invasive recording of spontaneous APs from within the tissue bulk for over a year, demonstrating its potential for exceptionally long-term studies without impeding organoid growth. Similarly, Yang et al. 2024 (100) introduced “Kirigami electronics” a highly deformable, basket-like structure that also supports organoids in suspension. Inspired by the art of paper cutting, this SU-8 based structure spontaneously transforms from a 2-D precursor to a 3-D basket upon release, offering a stable yet flexible scaffold. The platform allows for chronic recording for up to 120 days and was used to detect disease-related phenotypes (increased firing rates in a DiGeorge syndrome critical region gene 8 [DGCR8]＋/− model) and probe corticostriatal connectivity in assembloids via optogenetic stimulation, showing its utility for long-term functional studies of complex circuits. Wu et al. 2024 (101) expanded on the concept of soft, integrated interfaces by developing a 128-channel neuro-interface from a stretchable liquid metal-polymer conductor. This highly deformable mesh, made of a gallium-indium alloy within a polyurethane matrix, gently envelops hippocampal organoids. Its primary advantage is the combination of high channel count and extreme flexibility, successfully detecting APs and network activity from up to 85 channels simultaneously. This represents a significant increase in recording density over other flexible devices, which is crucial for detailed circuit mapping, though its application is currently focused on hippocampal models.

A distinct and powerful strategy involves integrating nanoelectronics directly into the tissue during its formation, creating “cyborg organoids.” This was first demonstrated by Li et al. 2019 (102), who co-cultured 2-D stem cell sheets with ultra-flexible, stretchable mesh nanoelectronics. During organogenesis, driven by cell-cell attraction forces, the 2-D cell sheet spontaneously folds into a 3-D structure, seamlessly embedding the nanoelectronics throughout the entire tissue. This approach was further refined by Le Floch et al. 2022 (103), who used a similar method with improved, softer mesh nanoelectronics to achieve chronically stable, tissue-wide electrophysiological mapping at single-cell resolution in developing brain organoids for up to 6 months. This technique is exceptionally powerful as it accommodates the organoid’s growth and morphological changes with minimal interruption, enabling the capture of emerging single-cell APs and network activity from the earliest developmental stages. Its main limitation is the complexity of the initial integration step. Lin et al. 2025 (104) advanced “cyborg organoid” technology by integrating stretchable mesh nanoelectronics into 3D organoids through a natural 2D-to-3D tissue reconfiguration during organogenesis. Their improved biomimetic mesh features high tissue flexibility (∼10% filling ratio, submicron thickness), subcellular electrode size (∼10 to 20 μm) with low impedance, and scalability via lithographic fabrication. This approach enables long-term, stable, single-cell electrophysiological recordings across the entire organoid with minimal disruption to development. It supports comprehensive functional mapping of organoid growth—including cardiac, brain, and pancreatic types—and opens possibilities for multimodal data integration and artificial intelligence (AI)–driven closed-loop control.

Finally, some approaches combine non-invasive support with enhanced functionality. Giandomenico et al. 2019 (68) leveraged an air-liquid interface culture method, which greatly improves nutrient and oxygen availability to sliced organoids. This allowed for improved neuronal survival and the growth of thick, long-range axon tracts that could functionally innervate external mouse spinal cord-muscle explants. To record directly from the organoid body, they used a commercial 3-D MEA, demonstrating how culture method innovation can be paired with existing hardware to probe functional output. A mechanically compliant approach was introduced by Ryu et al. 2021 (105) who fabricated transparent, 3-D mesostructures of parylene-C that reversibly open and close to gently envelop and mechanically restrain organoids. The primary innovation here is not electrophysiology, but enabling precise nanoindentation measurements of their viscoelastic properties, a key biophysical parameter. The device’s transparency and gentle handling highlight its potential for integration with other sensing modalities, though it is not primarily an electrical interface.

Collectively, these 3-D MEA technologies represent a paradigm shift in organoid electrophysiology. By moving beyond the planar surface, they enable the long-term, minimally invasive, and volumetric analysis of neural network development and function. While challenges related to scalability, data analysis complexity, and standardization remain, these innovative platforms are paving the way for a deeper understanding of the human brain in health and disease.

While many 3-D recording strategies focus on interfacing with the surface of brain organoids or embedding electronics during early developmental stages, an equally critical approach involves the use of implantable or penetrating MEAs (Fig. 5). These devices are specifically designed to access the deep interior of mature organoids, enabling direct monitoring of neural circuits that remain inaccessible to surface-based techniques. Unlike in-vivo brain tissue, where immune responses such as glial scarring can degrade electrode performance over time, brain organoids lack a fully developed immune system and exhibit minimal glial reactivity. As a result, implantable electrodes can achieve long-term, stable recordings within the organoid core without the chronic signal degradation typically observed in live animal models. The key technological challenge lies in balancing mechanical penetration capability with structural compliance—developing probes that are sharp enough to enter dense tissue while remaining soft or biocompatible enough to preserve the organoid’s integrity and functionality over extended periods. This makes implantable MEAs a particularly promising platform for high-resolution, volumetric electrophysiological mapping in organoid-based studies.

One major strategy involves fabricating arrays of sharp, vertically-oriented microelectrodes that can be inserted into the organoid. For instance, Scholvin et al. 2016 (106) and Quadrato et al. 2017 (16) utilized a high-density silicon MEA, originally designed for in-vivo rodent recordings, to probe deep within intact, mature human brain organoids. The rigid silicon shank, featuring 64 recording sites, was inserted into the organoid, successfully capturing spontaneous neuronal activity, including AP spikes and network bursts from internal cell populations. While this demonstrated the feasibility of using existing in-vivo tools, the rigidity of silicon probes poses a significant risk of tissue damage and is not ideal for chronic recording in soft, developing organoid tissue. To address the mechanical mismatch, Phouphetlinthong et al. 2023 (43) developed a novel “curvy and spiky” MEA (csMEA) with self-bending cantilevers. Fabricated from a bilayer of silicon nitride and silicon dioxide, internal stresses cause the microbeams to spontaneously bend upwards over 200 microns. This creates an array of sharp, protruding electrodes that can penetrate organoids, allowing for the stable recording of APs and LFPs from deep within the tissue for over two weeks with less damage than a rigid probe.

Similarly, a study by Kim et al. 2025 (107) employed another approach to assess the functional characteristics of 3D organoid arrays. While neural organoids hold promise for bio-inspired computing, conventional culture scalability limits their functional capacity. To address this, a novel 3-D stacking strategy, inspired by semiconductor technology, was developed. This approach vertically assembles Matrigel-embedded organoids in a PDMS chamber, creating a stable, multi-layer structure that preserves oxygen and nutrient diffusion. To assess these arrays, Kim et al. 2025 (107) utilized a MEMS-based neural probe with 14 black-platinum microelectrodes for enhanced recording and stimulation. A microdrive precisely lowered the probe to minimize damage. This allowed successful measurement and comparative analysis of neural firing rates and synchronization from single organoids, 2-D arrays, and the new 3-D arrays. Structural analysis confirmed robust inter-organoid connectivity, and electrophysiological recordings revealed significantly enhanced neural dynamics in 3-D arrays, with prolonged culture promoting network maturation and increased functional complexity.

Other approaches focus on creating flexible probes that are more mechanically compliant with the soft organoid tissue. Soscia et al. 2020 (108) developed a 3-D MEA from flexible polyimide probes. A key innovation is that the probes are mechanically actuated into a vertical orientation before the neurons are encapsulated in a surrounding hydrogel. This “pre-insertion” method avoids the damage associated with penetrating a mature, dense organoid and allowed for stable recording of spikes and bursts for over 45 days. Liu et al. 2021 (109) also fabricated a flexible probe using Parylene-C, focusing on a multilayer approach to increase channel count without increasing the probe’s physical footprint. This 16-channel probe was successfully used for acute single-unit recordings both in-vitro in an organoid and in-vivo in a mouse brain, demonstrating its versatility. Shin et al. 2021 (44) took this further by developing a multifunctional, high-density 3-D MEA by stacking and bonding multiple 2-D MEAs. This created a probe with 63 electrodes, integrated microfluidic channels for drug delivery, and an optical fiber for optogenetic stimulation. This device enabled sophisticated experiments, including the measurement of synaptic latency across compartmentalized 3-D neural networks, as well as the recording of neural spikes in human spinal cord organoids. Zips et al. 2023 (110) introduced a method for fabricating 3-D needle-type MEAs via transfer printing of conductive polymer inks, a process that is scalable and cost-effective. These probes successfully recorded spontaneous AP spikes and LFPs from organoids, demonstrating their utility for chronic functional monitoring.

Finally, some of the most innovative strategies combine novel materials and dynamic actuation to create highly advanced implantable probes. Kim et al. 2025 (48) developed a magnetically reshapable 3-D MEA using directly printed pillars of liquid metal (EGaIn). The extreme softness of the liquid metal minimizes tissue damage upon insertion. The groundbreaking feature is that a ferromagnetic coating allows the electrode tips to be tilted and repositioned in situ using an external magnetic field. This allows a single electrode to record signals from multiple distinct locations, dramatically increasing the effective recording density and enabling flexible 3-D mapping of APs and LFPs. In a different approach, Li et al. 2025 (56) used flexible shank probes to show that external ES could enhance the functional maturity of organoids. Following stimulation, these organoids were transplanted into injured mouse brains, where the same type of flexible probes were used to record their activity in-vivo, demonstrating that ES-pretreated grafts showed superior functional integration.

In summary, implantable and penetrating probes provide an essential toolkit for exploring the internal electrical landscape of brain organoids. While they inherently carry a risk of tissue disruption, ongoing innovations in materials science and microfabrication are leading to increasingly sophisticated devices that are softer, more flexible, and even dynamically reconfigurable. These technologies are crucial for bridging the gap between surface recordings and the complex, volumetric network activity that defines organoid function.