Bioelectronic Interfaces and Sensors for Neural Organoids

Abstract

Neural organoids are emerging as promising in vitro models, offering a unique platform to partially recapitulate the structural and functional complexity of the human nervous system. These three-dimensional (3D) constructs, which mimic key aspects of organ architecture, can be reliably derived from pluripotent stem cells (iPSCs) or embryonic stem cells (ESCs). Their ability to faithfully model neural development and disease pathogenesis has positioned them as indispensable tools in neuroscience research. However, to further unleash their potential, there is a pressing need for long-term and stable monitoring of their dynamic functions in a 3D context. This review provides a brief overview on diverse types of neural organoids and their induction protocols. We further highlight recent advancements in bioelectronic interfaces and sensors tailored for 3D culture. Finally, we discuss future directions aimed at advanced methodologies for real-time, multidimensional functional analysis, ultimately paving the way for breakthroughs in understanding neural development and pathology.

Introduction

With the gradual progress of biomedical research and tissue engineering, animal models and 2D-cell models have emerged as powerful tools for exploring the development and pathology of the human nervous system. Nonetheless, the inherent species discrepancies between animal models and human beings, coupled with the limitations of simplicity in 2D human cell cultures, have cast doubt on the validity of these models^1,2. To address these concerns and accurately replicate the intricacies of the human neural system, the advent of organoids, which are 3D self-organized structures derived from iPSCs and ESCs, has brought a new perspective^3,4. The neural organoids by far have been established vary a lot in types, including specific areas of brain^5,6,7,8, spinal cord⁹, retina¹⁰, inner ear¹¹ and so on. Compared to conventional 2D culture models, the 3D neural cultures generated complex hierarchical structures that foster increased cellular communication¹². Moreover, they can mimic part of the key structural and attributes of human organs^13,14, which is crucial for advancing our understanding of disease progression^15,16 and facilitating drug screening^17,18 in the context of human neurological disorders. For example, disease manifestations such as the deposition and misfolding of proteins in Alzheimer's disease (AD), can only be accurately reflected in 3D in vitro models¹⁵. Further, to fully unleash the potential of neural organoids, long-term and stable characterization of their organization structure and physiological functions in a 3D manner is essential.

Establishing the structural and developmental congruence between organoids and human tissues constitutes the most compelling evidence to validate the applicability of organoid models. The image detection is a convenient method to achieve it. For example, fluorescent labeling of developmental stages and cell type-specific proteins, have demonstrated that the developmental trajectory of human brain organoids closely resembles that of the human brain, and under optical microscopy, brain organoids exhibit key structural features similar to the ventricular and subventricular zones of the human brain^5,13. Enhanced clarity and granularity of information can be achieved through the application of advanced instruments, including scanning electron microscopy (SEM), which emphasizes the visualization of surface characteristic texture, and transmission electron microscopy (TEM), which is adept at capturing the internal ultrastructural details¹⁹. To adapt to the intricate complexity of the 3D structure of organoids, noninvasive optical sectioning techniques, such as confocal imaging, multiphoton microscopy and light-sheet technology have enabled the acquisition of more detailed 3D information from intact organoids²⁰. Further, the escalating demand for high-resolution dynamic observation has led to the emergence of 4D live cell imaging using light-sheet microscopy as a gentle and rapid imaging method on organoids^21,22. However, the traditional imaging methodologies are constrained in their capacity to capture a comprehensive spectrum of features, especially in the key features of discharge characteristics and metabolic activities. Additionally, the developmental trajectory of organoids is protracted and intricate, and the traditional terminal detection techniques exhibit inherent limitations when assessing this process.

To address the limitations of traditional methods, there is a pressing need for bioelectronic interfaces and sensors capable of continuous surveillance²³. These tools are essential for monitoring the morphological and functional evolution of organoids during development. Furthermore, due to the substantial potential of organoid models in realms such as regenerative medicine, disease modeling, and personalized therapeutics, there is also a need for bioelectronic interfaces and sensors to dissect the functional alterations of organoids under pathological conditions and accurately simulate the disease progression^24,25. New materials and advanced manufacturing methods are progressively breaking through the constraints of biosensors. High-density multi-electrode arrays can improve the spatiotemporal resolution²⁶ while compact multi-well arrays have made high-throughput drug screening feasible²⁴. Although several commercialized bioelectronic technologies have satisfied partial requirements, there remains some deficiencies (Table 1). The prevalent bioelectronic devices, being planar in configuration, can hardly accommodate the 3D architecture of organoids. Many researchers adapt the organoids to these flat devices either by directly contacting or slicing^26,27, which inevitably results in the loss of a large amount of valuable information. With the development of materials science and manufacturing technology, the neural interfaces are increasingly evolving towards miniaturization, enhanced flexibility, and improved biocompatibility^28,29. In addition, the complexity of organoids requires monitoring from diverse perspectives, including changes in morphology, electrophysiology and molecular levels. The development of multimodal bioelectronic interfaces and sensors, offering a more comprehensive characterization of organoid functions, represents an emerging frontier in the field^30,31,32.

Table 1 Commercial bioelectronic tools for organoids

In this review, we mainly focused on the cutting-edge advancements in neural organoids and evolution of bioelectronic interfaces and sensors for long-time dynamic monitoring. As illustrated in Fig. 1, neural organoids have been successfully developed to recapitulate key physiological characteristics of the nervous system in vivo. We provide a concise overview on the progress achieved in the induction technology of neural organoids over the past decade, which mainly encompasses brain, spinal cord, retina, inner ear and olfactory organoids. Various types of neural interfaces and sensors have been employed to reflect the comprehensive physiological characteristics of these neural organoids. This review critically evaluates the application of advanced neural interfaces for in vitro functional assessment of neural organoids, with particular emphasis on technological innovation of 3D biosensors for long-term dynamic detection in small 3D cultures (such as neural organoids). The enhanced structural conformity of these 3D interfaces with organoids facilitate more durable, real-time, high-throughput and stable monitoring capabilities. These technological breakthroughs hold substantial translational potential, with future applications anticipated in human physiological and pathological modeling, pharmaceutical screening, regenerative medicine and other related fields.

Fig. 1

The modeling, monitoring and application of neural organoids

Different types of neural organoids

Brain organoids

The human brain is such a complex and unique system that its protracted development and genetic characteristics cannot be commendably described by animal models¹. The establishment of brain organoids has cast light on our understanding of our human brain⁵. Different from neuro-spheres, the unique advantages of organoids lie not only in 3D structure but also in the capability to replicate the cellular composition, intercellular tissue structure and developmental trajectory of the human brain.

In general, the generation of brain organoids typically commences with the formation of an embryoid body (EB) or a spheroid, which is subsequently subjected to neural induction media to direct differentiation towards a neuroectodermal lineage^5,33. Two principal methodologies are employed in the generation of brain organoids: unguided and guided approaches (Fig. 2). Unguided methods, represented by the protocol of Lancaster in 2013, rely on the self-assembly and intrinsic differentiation capacities of iPSCs and ESCs⁵. The EBs are embedded in the extracellular matrix (ECM) with few factors and a spinning bioreactor is needed for continuous maturity and long-term cultivation^5,34,35. The products, also called cerebral organoids, which can achieve diameters of up to 4 mm, encompass a diverse array of cellular constituents, including neural progenitors, mature neurons, astrocytes and so on^5,13,36. Encouragingly, the lumen can be observed in the cerebral organoids, and cells are organized into polarized radial structures, forming a ventricular zone, subventricular zone, and cortical plate orderly, which is reminiscent of the early developmental process of newborns^34,37. However, unrestricted induction brings about unpredictable product sizes and heterogeneous internal arrangements, posing challenges for reproducibility in experimental settings. In contrast, guided methods involve the addition of specific exogenous components during the early stages of differentiation, which can constrain the self-patterning of organoids and lead to specific brain regions, such as the forebrain^38,39, midbrain^40,41, cerebellum^42,43, hippocampus^44,45, thalamus^46,47, hypothalamus^8,48, striatum^49,50 and choroid plexus^51,52 (Table 2). Thanks to the establishment of region-specific organoids, customized organoids for different regions can be formed separately before being assembled, forming fused organoids. The method embraces various cell lineages from different brain regions, offering a controlled and reproducible framework for investigating the intricate intercellular interactions^53,54. So far, researches on fused brain organoids have primarily focused on cellular migration^55,56, nerve distal projection^46,49, regulation of downstream tissues^57,58 and so on.

Fig. 2

The production of different types of neural organoids through patterned differentiation using an appropriate combination of cytokines derived from human iPSCs/ ESCs or primary mouse cells^{27,233,234,235,236,237}

Table 2 The induction methods for reginal specific brain organoids

Overall, the brain organoids with hierarchical structures and organizational microenvironments have emerged as invaluable tools for modeling human neurological disorders, including autism⁵⁹, AD¹⁶, Parkinson's disease (PD)⁶⁰ and so on. However, due to the maturity limitation, the current models excel in recapitulating the early embryonic neurodevelopment as well as in the screening of safe drugs during pregnancy^61,62. The establishment of complex neuronal circuitry and description of system aging remain the weaknesses. To compensate these shortcomings, researchers have incorporated elderly primary monocytes into the model, which enhanced the appearance of aging phenotypes in brain organoids successfully⁶³. With the development of technology and the continuous optimization of modeling schemes, brain organoids are poised to complement other models, offering a more comprehensive and realistic reflection of human brain.

Spinal cord organoids

The spinal cord, another pivotal component of the central nervous system, serves for information exchange between the brain and the periphery, and it is also responsible for the generation of motor activity⁶⁴. The spinal motor neurons and sensory neurons within are of vital importance for the maintenance of physiological functions and damage to these neurons can lead to motor neuron diseases, including amyotrophic lateral sclerosis (ALS) and spinal muscular atrophy (SMA)⁶⁵.

The first cerebral organoids established in 2013 have opened a new chapter for in vitro 3D modeling of the central nervous system, subsequently igniting interest in the development of hindbrain and spinal cord organoids with refine protocols⁶⁶. Similarly, embryoid bodies are formed before the neural induction and maturation process. The incorporation of specific factors during these stages guides towards a more precise cellular fate, mitigating the cellular heterogeneity derived from unrestricted self-differentiation. Specifically, the inhibition of the SMAD signaling pathway, in conjunction with the coordinated roles of retinoic acid (RA), WNT and fibroblast growth factor (FGF) signaling pathways, facilitates the differentiation and axial extension of neuro-mesodermal progenitors (NMPs) (Fig. 2)^9,67. Recent studies have demonstrated that human iPSC-derived gastrula can elongate into distinctive tubular structures, with axial symmetry disruption stimulated by WNT signaling and the formation of NMPs close to a signaling center similar to a tail bud, following the initial spherical configuration^68,69. The mature spinal cord organoids exhibit a range of cell types and structures along the dorsal-ventral and rostral-caudal axes within a single entity, overcoming the limitation of traditional 2D cultures and addressing the significant challenge of simulating the intricate tubular architecture of the spinal cord in vitro⁷⁰.

Motor neuron diseases pose significant health risks and organoid models have proven instrumental in mitigating these risks⁷¹. The organoids derived from patient-specific iPSCs are instrumental in elucidating the underlying molecular mechanisms of neurodegeneration and further serve as a drug screening platform⁷², facilitating personalized medical interventions. Numerous studies have focused on ALS⁷³ and SMA⁷⁴, investigating the pathogenesis and potential therapeutics. In addition, stem cell-based therapies have shown potential in replacing lost spinal cord neurons, with some animal and clinical trials reporting that the transplantation of spinal cord-derived neural stem cells can ameliorate disease progression⁷⁵.

Retinal organoids

The retina, an integral component of the peripheral nervous system, originates from the diencephalon during brain development, where a region transforms into a pair of optic vesicles that ultimately differentiate into the retina, the neural tissue directly interfacing with the external environment^76,77. Relying on the substantial correlation between the retina and brain development process, Elke Gabriel et al. reconstructed the induction process of brain organoids to incorporate optic vesicle-containing brain organoids, thereby achieving the functional integration of retinal structure within brain organoids⁷⁸. While more studies induced retinal organoids individually by using iPSCs, retinal progenitor cells (RPCs), and other types of stem cells^79,80. The induction protocols for retinal organoids exhibit variability, yet they generally emulate the developmental trajectory of the human retina, with critical guidance on the optic vesicle's fate^80,81,82. Initially, stem cells are aggregated or induced to form EBs at the beginning, which can be continuously sustained in suspension or added into ECM for further induction and cultivation. Some key signaling pathways, such as WNT and Notch, are of vital importance in cell fate determination^83,84, necessitating the inclusion of exogenous factors such as BMP4, DKK1 and RA in the induction protocols (Fig. 2). With the progress of development, neural retina progenitors differentiate into retinal ganglion cells, amacrine cells, and horizontal cells, followed by photoreceptor cells⁸⁵. After the neural induction and retinal differentiation procedure, the organoids are continuously induced for over 100 days. Ultimately, the mature retinal organoids encompass a spectrum of retinal neuronal cell types arranged in a laminated structure, with photoreceptor cells (PRCs) located in the outer layer and retinal ganglion cells (RGCs) in the inner layer of organoids^86,87,88.

Beyond their significance in developmental theory and therapeutic strategies, the retinal organoids have accelerated advancements in regenerative medicine. Patients with advanced retinal degeneration, often hindered by a scarcity of donor sources, may benefit from retinal organoids, which could potentially alleviate the transplantation gap. The high degree of cellular and structural similarity between organoids and the human retina allows for the extraction and subsequent transplantation of corresponding segments into diseased ocular regions. This surgical approach has demonstrated promising results in animal models, offering a significant breakthrough for patients with visual impairments^89,90,91.

Inner ear organoids

The vertebrate inner ear, which originates from the otic placode, is responsible for balance maintenance and undertakes the function of the auditory system⁹². The sensory cells in the inner ear, also known as hair cells, work as transducers to convert mechanical stimuli into electrical signals. However, these cells are incapable of self-duplication and their absence can result in irreversible hearing impairment and balance disorders^93,94. With the deepening of stem cell biology, the inner ear organoids are promising to become new donors for transplantation, offering a potential therapeutic avenue for individuals with hearing impairments⁹⁵. The generation of inner ear organoids commences with the establishment of EBs, which serve as the foundation for a 3D structural framework. The key branch point of inner ear fate depends on the different orientations of otic placodes, with the Sonic Hedgehog and WNT signaling pathways being pivotal in fate selection (Fig. 2). A suite of key growth factors, including TGF, FGF, and BMP are implicated in the regulation of the pathways mentioned above^96,97. Upon completion of the differentiation process, the inner ear organoids successfully generate the sensory epithelia, consisting of functional hair cells, support cells and spiral ganglion neurons^97,98.

The inner ear organoids have broad prospects in the exploration of developmental and pathological mechanisms, as well as in drug development. Stem cell therapy has been increasingly mentioned in the treatment of hearing loss while the transplanted stem cells often encounter impediments in normal differentiation and migration⁹⁶. To address this challenge, organoids can provide a more differentiated population of neural progenitor cells for delivery into the body, presenting a promising alternative approach⁹⁹.

Olfactory organoids

The initiation of olfactory perception transpires within the olfactory epithelium, where odorant molecules interact with corresponding receptors on olfactory sensory neurons (OSNs). The odor information can be subsequently transmitted to the olfactory bulb and olfactory cortex by OSNs¹⁰⁰. Our team has devoted a lot to establishing the olfactory organoids model based on the regenerative ability of basal cells in the olfactory epithelium^101,102. To cultivate olfactory organoids, mouse olfactory epithelial tissue is meticulously segregated and digested to obtain single basal cells. Within a Matrigel matrix, the basal cells capable of self-differentiating into olfactory organoids, facilitated by factors that modulate the WNT and Notch signaling pathways (Fig. 2)¹⁰³.For the regeneration of olfactory basal cells, the acquisition of olfactory organoids is more convenient and time-saving. Investigating olfactory disorders holds the potential to aid in the early detection of certain neurodegenerative conditions, such as AD¹⁰³. Furthermore, the acute olfactory acuity of biological systems has enabled the development of biomimetic sensors that incorporate biomaterials, such as olfactory organoids, as sensing components. These sensors are capable of closely replicating the biological sensory system in vitro with a high degree of stability and reproducibility¹⁰⁴. Nonetheless, interspecies disparities still present a challenge in the accurate reconstruction of human sensory perception in vitro.

In general, with further exploration in developmental biology, researchers have successfully constructed neural organoids of the human species, including the brain, spinal cord, retina and inner ear. The models emulate the early developmental stages of human embryos, thereby providing a novel avenue for probing the intricacies of early neurodevelopment and associated pathologies. The emergence of organoid models has significantly broadened our investigative repertoire, offering a complementary approach to the limitations inherent in animal and cell models. The induction process of neural organoids is conceptually similar to that of 2D neural cultures, yet it requires the preservation of a 3D architecture through supportive systems such as gel culture or other scaffold frameworks. For instance, biomaterials such as Matrigel, natural hydrogels¹⁰⁵ and hyaluronic acid hydrogels^106,107 have played an important role in the development of neural organoids due to their good biocompatibility and capacity to incorporate neurotrophic factors. Despite these advancements, the current modeling methods for neural organoids are still in their early stages. There are still substantial disparities between these models and living organisms in terms of morphological characteristics, architectural configurations and cellular composition.

The lack of vascularization represents a common problem in the modeling of various neural organoids. Some large neural organoids, such as brain and retina organoids, can reach a diameter of over 1 mm. In the absence of a vascular network, the normal transportation of nutrients and the elimination of metabolic waste are impeded within such a large volume, ultimately leading to the death of organoids in the core¹⁰⁸. Common strategies to address this issue include transplantation into a host organism^109,110 or co-culture with vascular tissues^111,112. On this basis, engineering solutions, such as organ-on-chip technology, with microfluidic devices, have been employed to further refine the available methodologies. These devices are capable of simulating the process of vascular invasion into neural tissue during embryonic angiogenesis and can facilitate the targeted transport of molecules to specific locations via microfluidic channels¹¹³. Numerous studies have reported the advantages of vascularized neural organoids, including enhanced maturation and improved functional outcomes^114,115.

In addition, the formation of organoids relies on the self-assembly of stem cells, which inherently introduces substantial variability in cell type, size, and morphology among individual organoids within the same batch¹¹⁶. This intrinsic heterogeneity makes it difficult to achieve reproducibility and consistency in research. To address these limitations, several strategies have been proposed. Firstly, reducing the operational variability during the formation process and transitioning towards automated cultivation platforms for organoids can enhance reproducibility and standardization^117,118,119. Secondly, controlling the quality of intermediate products during critical induction steps is essential. Specifically, most reported protocols for neural organoid induction initiate with stem cell aggregates known as EBs, and recent studies have made effort in achieving consistent EB states to mitigate variability^120,121,122. Additionally, 3D bioprinting has emerged as a promising technology for the mass production of highly consistent neural organoids. It can precisely print various cell types within defined 3D architectures with bioinks, thereby offering a scalable and uniform approach to neural organoid fabrication^123,124.

Multi-electrode array (MEA) for electrophysiological recording of neural organoids

The traditional biochemical analysis and imaging techniques that primarily delineate the cellular constituents and anatomical characteristics of neural organoids capture only a limited aspect of the multifaceted functions of these models. The core contribution of these neural organoids, however, lies in their capacity to simulate functionally active neurons within the human nervous system and establish interconnected neural networks. Indeed, in some neurological diseases including but not limited to epilepsy¹²⁵, AD¹²⁶, the early stage of retinitis pigmentosa¹²⁷ and the early stage of spinal neurodegenerative diseases¹²⁸, clinical observations often fail to reveal substantial structural alterations. Consequently, it proposes new challenges for researchers to identify the subtle yet critical impairments in the neural circuitry, which underlie these neurological disorders. To ensure the accurate assessment of neuronal firing and the functionality neural networks, bioelectrical interfaces and sensors are needed to acquire electrophysiological data of these neural organoids. The MEA allows for the simultaneous recording of electrical signals across multiple sites, providing insights into the spatiotemporal dynamics of neural network activity and facilitating a more comprehensive evaluation of the functional integrity and maturation of neural organoids¹²⁹. Here, we shall present an overview on various types of MEAs that are frequently employed on neural organoids, along with their structural adaptations designed to address the novel challenges posed by the 3D architecture of organoids.

Planer MEA

The traditional electrode technology, which employs rigid substrate materials, continues to play a fundamental role in the field of neurophysiological research. This technology has been widely applied across various contexts, ranging from in vivo neurons¹³⁰ to stem cell-derived 2D neurons¹³¹, and more recently, in neural organoids³⁷. Specifically, neural organoids are positioned on these electrodes to establish neural interfaces. The encapsulated metal electrodes can then be connected to an amplification system, which enables the continuous and long-term recording of the discharge patterns of these organoids after the preprocessing and analysis of computer. The signals detected by bioelectronic interfaces encompass extracellular action potentials (EAPs), commonly referred to as "spikes," from individual neurons, as well as local field potentials (LFP), which aggregate synaptic currents within a specific region¹³². When the interconnected neural network is formed, synchronous bursts of spike and flourish of LFP can be observed simultaneously, providing compelling evidence for the maturity and credible function of neural organoids^37,133. Take human brain organoids for example, numerous studies have documented the progressive enhancement of electrical activity and characteristic network dynamics across a 10-month period^37,133. These investigations have consistently reported that electrical activity in human brain organoids is initially detected after approximately one month of cultivation, coinciding with the early stages of neurogenesis within brain organoids. Following two months of development, the brain organoids exhibit a gradual increase in the firing rate and amplitude of spike, accompanied by highly synchronized network activity. In the advanced stages of development, the proliferation of inhibitory interneurons within organoids leads to a coordinated balance between inhibitory and excitatory neurons, signifying the maturation of functional neural network communication (Fig. 3a–c). The complex oscillatory waves emerging from brain organoids are similar to that of human infants recorded by EEG devices, affirming the proficiency of brain organoids in mimicking the early developmental stages of the human brain³⁷. The technology has since been extended to explore the electrophysiological characteristics of other neural organoids, such as spinal cord organoids¹³⁴ and olfactory organoids¹⁰⁴.

Fig. 3: Typical neural oscillations in neural organoids detected by planar MEA.

a, b Spontaneous electrical activity can be detected from brain organoids using a MEA system. a A transferred brain organoid on an MEA probe during acquisition of spontaneous electrical activity. b brain organoid recording setup with 64 planar microelectrodes¹³³. c Detection of spontaneous electrical activity in day-34 brain organoids and approximately every month thereafter shows a progressive increase in and complexity of EP properties¹³³. (with permission of Elsevier). d Left, picture of a brain organoid at 2 months, scale bar is 1 mm. Right, organoids (4–6 months old) were embedded in 10% agarose and sectioned into 500 μm thick slices²⁷. e Organoid slices were then placed on high-density CMOS microelectrode arrays²⁷ (with permission of Springer Nature)

The functional neural networks within the human body are highly precise and intricate, yet they remain vulnerable. Upon being compromised by pathological conditions, it may not exhibit overt anatomical alterations, yet they often manifest significant electrophysiological aberrations, which compromise their functional integrity¹³⁵. The human brain, which serves as the central hub for the collection and processing of vast amounts of neuronal information within the central nervous system, has been the focus of numerous studies, which employ neural interfaces to explore the pathogenesis of various neurological disorders. These disorders encompass neurodevelopmental conditions such as Rett syndrome¹³⁶ and autism spectrum disorder¹³⁷, neurodegenerative diseases like AD¹³⁸, Parkinson's disease¹³⁹ and amyotrophic lateral sclerosis¹⁴⁰, as well as epilepsy¹⁴¹ (Table 3). By leveraging key electrophysiological features, such as firing rate and network bust rate, these studies aim to replicate the pathogenesis of human neurological disorders in vitro. The approach also provides a reliable platform for clinically relevant drug screening, for instance, hyaluronan has been shown to inhibit the overexcitability of neural and exert a mitigating effect on autism spectrum disorders in the 3D spheroid model¹⁴².

Table 3 Phenotype of organoids with neurological disorders by neural interfaces

In addition, drugs can also induce neurotoxicity, resulting in damage to the nervous system. Fetal brains, in contrast to adult brains, exhibit heightened susceptibility to such detrimental impacts¹⁴³. Since the brain organoid can exhibit the morphological and functional properties of the early embryonic human brain more accurately⁶¹, it demonstrates superior efficacy in the screening of neurotoxic compounds. For example, the combination of rigid bioelectrode interfaces with brain organoids has been utilized in the evaluation of medication-assisted treatment in opioid use disorder during gestation^62,144. Beyond their well-established applications in modeling neurodevelopmental processes and investigating disease mechanisms, neural organoids have emerged as promising candidates for cell replacement therapy through tissue transplantation. For example, recent reports have demonstrated that rigid bioelectronic interfaces serve as essential tools for evaluating the functional integration of transplanted human retinal organoids in rats, thereby facilitating improvements in visual function^90,145.

Although the current electrode technology can fulfill the fundamental requirements for detecting electrophysiological networks in neural organoids, there remain many aspects requiring improvement. One of the limitations lies in the restricted number of channels, which leads to insufficient spatiotemporal resolution³⁷. The utilization of Complementary Metal Oxide Semiconductor (CMOS) technology has emerged as a promising solution to these limitations. It demonstrates remarkable compatibility with various rigid substrate materials, including silicon and glass¹⁴⁶. The incorporation of CMOS technology has facilitated the development of electrode arrays with enhanced channel density and greater integration, thereby significantly improving detection probability and the richness of acquired information^27,147. When the number of electrodes reaches 26,400, subcellular spatial resolution becomes attainable, which is a representative commonly used CMOS-MEA¹⁴⁸. Further, the CMOS-MEA designed by Ikuro Suzuki et al. has reached an impressive 236880 channels and has been successfully implemented in various neural tissue models, including mouse brain slices, 2D neural cell cultures, peripheral neurons, and human brain organoids²⁶. The acquisition of high temporal resolution is crucial for accurately capturing rapidly fluctuating electrophysiological signals. This capability not only enables expeditious data collection but also facilitates the comprehensive analysis of dynamic properties within neural networks. For instance, it opens a new avenue for investigating the correlation between cell waveform clusters and specific cell types¹⁴⁹. Furthermore, the high-density electrode configuration provides extensive coverage of larger measurement areas, thereby enhancing the capacity to monitor propagation patterns across different regions of large-scale neural organoids, including the modeling of multiple brain region interactions²⁶.

Apart from that, similar to electroencephalography (EEG) or electrocorticography (ECoG), the detection depth of planar MEA is limited to the immediate 100 µm vicinity of the electrodes¹⁵⁰, while the diameter of neural organoids can often reach the millimeter level⁵. Consequently, a large amount of valid information, especially the discharge situation in the core parts of organoids and global electrophysiological network connection, remains beyond effective characterization. To obtain deeper activity records, the slicing of tissues or organoids has been employed to accommodate the planar MEA designed for 2D cultures (Fig. 3d-e)^27,151. Although the slicing of organoids facilitates long-term cultivation and real-time monitoring, the challenges associated with delicate slicing procedures and the inevitable damage incurred still present certain obstacles³⁶. Most critically, regardless of the increase in electrode numbers or the supplement of slicing technology, the data acquired remains confined to a single 2D plane, thus failing to fully exploit the holistic advantages of 3D models¹³².

3D MEA

The organoid model serves as a valuable repository of information, and the application of commercialized planer electrodes has partially harnessed its potential in neural development, disease modeling, drug screening and so on. However, the structural constraints of planar electrodes limit their capacity to access deep tissue regions that generate specific local signals. Biological signal generation within organisms typically exhibits regional specificity, for instance, distinct brain regions possess unique roles and functions, and their communication and interconnectivity collectively underpin the holistic functionality of the brain¹⁵². To fully exploit the potential of this 3D organization, various 3D MEA technologies are emerging.

The traditional rigid probes used to contribute to understanding the electrophysiological activity of the brain in vivo¹⁵³ and some studies have also attempted to apply it to 2D or 3D neural cultures¹⁵⁴. For example, Hyogeun Shin et al. designed a 3D high-density MEA, and their design concept involves sequentially stacking three 2D MEAs of varying dimensions, which were previously employed for in vivo neural recording, to construct a 3D configuration. The 3D MEA, integrated with 63 recording microelectrodes and multifunction of optical stimulation and drug delivery, can record the developing neural network of neuron culture nearby daily. Although the culture analyzed is not as complex as organoids, the spatial mapping of active neurons has been preliminarily established¹⁵⁵. However, this approach has not been widely researched and applied for the reason that these electrodes, originally developed for in vivo transplantation, have poor adaptability to small organoids in terms of size. Also, the invasive implantation process with destructive effects causes damage to the neural network connections of organoids, hindering the long-term stable signal sources. The emergence of increasingly sensitive and flexible detection devices tailored for organoids are becoming viable alternatives.

Microneedle electrode

Compared to the aforementioned invasive approaches of sliced or rigid-implanted electrodes, microneedle electrode represents a comparatively non-invasive form of 3D interfaces, which does not compromise the tissue integrity of organoids. A pivotal advantage of this configuration lies in its enhanced signal transduction efficiency at the tissue-electrode interface¹⁵⁶. The fidelity of electrophysiological recordings from organoids is critically dependent on the quality of this interface. For planar electrodes, the thickness of the interstitial gap between the cells and electrode surface is about 150 nm under optimal contact conditions and the situation is more pessimistic for organoids due to its 3D characteristic¹⁵⁷. In contrast, when organoids are interfaced with microneedle electrodes, the mechanical compliance of the cellular membrane facilitates its deformation and subsequent encapsulation of the electrode surface. This phenomenon significantly improves both the sealing impedance and effective contact area, thereby optimizing interfacial coupling and signal transduction efficiency¹⁵⁸.

A typical microneedle electrode was developed by Oranny Phouhetlintong et al., utilizing the principle of accumulated stress within multi-layer thin films to fabricate a microelectrode array featuring protruding cantilevers (Fig. 4a–c)¹⁵⁹. This configuration comprises 60 vertically aligned electrodes arranged in a uniform square array across four quadrants, with precise spatial distribution. These electrodes, extending several hundred micrometers from the substrate surface, enable the monitoring of intrinsic electrophysiological activity within brain organoids. Notably, the tight interfacial coupling between the organoids and electrode array significantly enhanced signal amplitude, thereby facilitating reliable detection and analysis. Additionally, other microneedle electrode with diverse geometries, such as mushroom-shaped, wire, pillar, tube, volcanic shapes and coronal designs have demonstrated distinct advantages in 2D cellular electrophysiological recordings and hold significant potential for application in 3D organoid studies¹⁶⁰.

Fig. 4: Some typical microneedle electrodes for neural organoids.

a Technological cross section of the MEA showing the layer stack before and after sacrificial layer etching and beam displacement¹⁵⁹. b Schematic of the placement of a cerebral organoid on top of a conventional planar MEA showing the cleft layer between cells and electrodes and a schematic representation of an organoid inserted on a spiky microelectrode¹⁵⁹. c 3D schematic representation of a cerebral organoid impaled on the MEA¹⁵⁹. (with permission of Royal Society of Chemistry). d Schematic illustrations of the pneumatically controlled direct printing stage with high precision¹⁶³. e Schematic illustrations of microelectrode array with 3D liquid metal electrodes for retinal organoids with various heights¹⁶³. (with permission of Wiley). f Configuration of 3D liquid metal MEA¹⁶⁵. g Schematic illustration of magnetically reshapable 3D liquid metal MEA¹⁶⁵ (with permission of Springer Nature). h SEM image of a 4 by 4 grid of μ-needles printed in different heights with print durations ranging from 60 to 300 s onto a gold-covered wafer¹⁶⁶. i Bright-field microscopy image of a μ-needle MEA¹⁶⁶. j Photograph of the experimental setup. Four organoids were immobilized in 4 wells on the μ-needle MEA, which was inserted in the headstage of the amplifier system. After being filled with medium, the glass ring chamber on the MEA was sealed with a PDMS lid equipped with an Ag/AgCl reference electrode¹⁶⁶. k Time traces of signals detected via a μ-needle MEA from a brain organoid¹⁶⁶ (with permission of American Chemical Society)

Traditional fabrication of microneedle electrodes predominantly relies on conventional photolithography techniques, which impose stringent requirements on the manufacturing processes and consequently entail substantial costs^160,161. Also, the traditional method often struggles to accommodate the variable sizes of different neural organoids, thereby limiting their adaptability. In contrast, 3D printing has emerged as a transformative fabrication paradigm, offering enhanced simplicity and exceptional customization capabilities¹⁶². The research team led by Jang-Ung Park has made significant contributions in the development of 3D-printed liquid metal microneedle electrodes. Utilizing high-precision 3D printing technology, the team has developed flexible liquid metal microelectrodes capable of precise targeting to the inner layer of retinal organoids. These electrodes exhibit remarkable structural compatibility with retinal organoids, thereby mitigating mechanical trauma and preserving tissue integrity. The team further demonstrated the utility of this platform in monitoring the development of RGCs within retinal organoids, which exhibit striking morphological and functional similarities to early-stage fetal retina (Fig. 4d–e)¹⁶³. Moreover, the printed electrodes can be integrated into an electrical stimulation system, enabling precise spatial modulation of retinal neural activity and concurrent electrophysiological recording of both pre- and post-stimulation responses¹⁶⁴. Besides, the research team designed a reconfigured 3D MEA using liquid metal direct printing technology for the purpose of recording electrophysiological signals from brain organoids. This innovative approach capitalizes on the magnetic field-assisted reconfiguration capability of the flexible microneedle electrodes, enabling the establishment of multiple recording sites through a single electrode unit. Such spatial reconfiguration significantly enhances recording site density without proportional increases in electrode quantity, thereby optimizing the spatial resolution of neural activity mapping (Fig. 4f–g)¹⁶⁵. Sabine Zips et al. employed another 3D printing paradigm, aerosol jet printing, to readily achieve rapid 3D patterning of functional inks with high aspect ratios and varying heights. The μ-needle electrodes fabricated enable long-term and stable application on neurospheres and brain organoids derived from iPSCs (Fig. 4h–k)¹⁶⁶.

In addition, microneedle electrodes offer a significant advantage in accessing the intracellular membrane potential of individual cells^167,168. While traditional patch-clamp methods can provide high-fidelity electrophysiological recordings at the single-cell level, they are inherently limited by their manual operation and low-throughput characteristics¹⁴. In contrast, microneedle electrodes can be synergistically integrated with auxiliary methodologies, including electroporation¹⁶⁹ and optoporation¹⁷⁰, to enable precise localization of single-cell membrane penetration, thereby facilitating the acquisition of high-throughput and stable intracellular electrophysiological signals. To date, the application has primarily been confined to 2D cell cultures while its potential extension to 3D organoid systems represents a promising direction. Overall, the microneedle electrodes demonstrate the dual capability of precisely monitoring individual cellular activities while simultaneously capturing the collective electrophysiological network dynamics within organoids to a certain depth.

Flexible interface and stretchable mesh

The advent of flexible bioelectronics has ushered in a new era of neural interfaces, wherein flexible interfaces and scaffolds demonstrate superior integration capabilities with biological tissues, potentially mitigating the adverse effects frequently associated with foreign material integration on organoid development and function. These flexible substrates, primarily fabricated from polymers such as SU-8, polyimide and parylene C, are engineered through advanced microfabrication techniques including spin coating, photolithography, and reactive ion etching to create precisely patterned functional layers^28,29. In contrast to microneedle electrode, which primarily target deep electrophysiology signals, flexible MEAs exhibit exceptional compatibility with the 3D architecture of organoid surfaces. This design expands the contact area with the neural organoids, thereby enabling a macroscopic assessment of the overall neural network formation and signal propagation within the organoids. Remarkably, the system also compatible with a commercially 256-channel MEA electrophysiological hardware system, which means the convenience of functional characteristics catching and widespread use (Fig. 5a)^171,172. For the sake of a wider range of applications, the authors further developed improved computational analytical tools to increase the spatial precision, especially in the Z dimension of the 3D tissue, making the engineered 3D neural organoids a visualized and trustworthy model in vitro¹⁷³.

Fig. 5: Some typical flexible interface for neural organoids.

a Light micrograph of a 3D flexible MEA with vertical electrodes¹⁷¹. (with permission of Royal Society of Chemistry). b Schematic of a brain organoid encapsulated in the 3D self-rolled MEA¹⁷⁴. (with permission of IOP Publishing Ltd.). c Schematic of the self-folding 3D shell MEA¹⁷⁵. (with permission of Wiley). d FEM snapshots showing organoids of different sizes (400 to 600 μm) fitting in tailored shell electrodes¹⁷⁶. e Corresponding SEM images of 3D shell electrodes with different levels of folding. Scale bar, 100 μm¹⁷⁶. (with permission of AAAS). f Optical image of a brain organoid enclosed in a 3D mesh MEA¹⁷⁷. g 3D illustration of the positions of the microelectrodes across the surface of the spheroid¹⁷⁷. (with permission of AAAS). h Photos of the whole device and a single well with a suspended mesh MEA¹⁷⁸. i Waveforms overlaid at electrode positions of the mesh MEA¹⁷⁸. (with permission of Wiley)

Nevertheless, the electrodes mentioned above remain constrained to specific regions of organoid surfaces. The limitation has prompted ongoing research efforts to develop optimized flexible interfaces and scaffolds with diverse geometries to maximize tissue-electrode contact and enable comprehensive reflection of the overall network activity. For instance, the 3D self-rolled array developed by Kalmykov et al. is capable of achieving complete spatial envelopment of spherical organoid structures, driven by the residual mismatch stress between the constituent layers (Fig. 5b)¹⁷⁴. Similarly, Cools et al. designed a 3D self-rolled shell MEA, which mimics the anatomical configuration of electroencephalogram caps. The array, with good biocompatibility and a higher signal-to-noise ratio, can perform fluorescence imaging along with signal recording for its optical transparency (Fig. 5c)¹⁷⁵. Addressing the inherent size variability in spontaneously differentiated brain organoids, QI HUANG et al. engineered an adaptable shell MEA with tunable folding characteristics through precise photolithographic parameter modulation, enabling robust electrophysiology recording of organoids ranging from 400 to 600 μm (Fig. 5d–e)¹⁷⁶. Furthermore, the buckled MEA designed by Park et al. utilizes mechanically designed compressive buckling to transform 2D films into 3D structures. This multifunctional neural interface not only facilitates the investigation of synchronous burst events on the surface of individual brain organoid spheres but also enables the study of the cascading processes involved in the formation and regeneration of the overall electrical activity network between assembloids (Fig. 5f–g)¹⁷⁷.

However, the aforementioned neural interfaces are primarily more suitable for acute detection rather than long-term monitoring for that the long-term growth and survival of organoids necessitate adequate oxygenation and nutrient supply. To achieve sustained electrophysiological recordings from neural organoids, researchers have developed innovative strategies for maintaining suspended state of these organoids in culture medium while establishing stable neural interface connections. Mesh scaffolds represent an effective solution, offering several critical advantages for prolonged organoid maintenance. These scaffolds facilitate free floating of organoids, eliminating undesirable mechanical constraints associated with solid substrates. This design not only supports unhindered 3D growth of organoids but also preserves the integrity of their intrinsic electrophysiological networks, thereby enabling reliable long-term signal acquisition and monitoring (Fig. 5h–i)¹⁷⁸.

The growth and development of neural organoids typically involve extended periods, rendering chronic electrophysiological monitoring indispensable for the comprehensive assessment of growth trajectories, functional characterization and pharmacological intervention efficacy in these 3D models. To accommodate the 3D growth and volume expansion inherent in the development of organoids, an ideal long-term sensing device should possess not only sufficient bending stiffness to adapt to the curved boundaries of 3D cultures and form a stable electrical interface, but also a stretchable design that enables the device to extend in response to forces generated by the organoids themselves.

The stretchable mesh MEA reported by Thomas L. Li et al. incorporates an elastic polymer, styrene-ethylene-butylene-styrene (SEBS), in the selection of substrate and encapsulation materials. The electrodes facilitate the establishment of enduring neural interfaces with freely suspended organoid cultures, demonstrating exceptional mechanical resilience with tolerance to 50% compressive and tensile strain (Fig. 6a–b)¹⁷⁹. Yan Wu et al. have also developed a highly stretchable MEA employing a gallium indium (GaIn) alloy integrated with elastic polymers (thermoplastic polyurethane TPU, and polyurethane, PU). The electrodes can freely deform and fold at any angle when the organoids float and grow within the culture medium, exhibiting stretchability of up to 500% (Fig. 6c–d)⁴⁵. Furthermore, the Kirigami Electronics reported by Xiao Yang et al. feature a highly innovative structural design. Drawing inspiration from the principles of Kirigami art, researchers have devised an ultrathin 2D spiral pattern that spontaneously transforms into a 3D basket-like structure upon suspension. The neural organoid is subsequently embedded within the concentrated area of the bottom electrodes. Owing to the unique Kirigami pattern design, the electronics exhibit remarkable deformability, enabling long-term recording on freely floating organoids for at least 120 days without compromising their spherical morphology and differentiation processes (Fig. 6e–f)¹⁸⁰.

Fig. 6: Some typical stretchable mesh for neural organoids.

a An illustration of the stretchable mesh MEA and its fabrication process¹⁷⁹. b Photographs of a mesh electrode with organoids being manipulated by metal forceps. The left image is of an undisturbed mesh, and the right image is the mesh being deformed by the forceps¹⁷⁹. (with permission of Elsevier). c Schematic diagram of coupling the highly stretchable MEA with the brain organoid⁴⁵. d The bottom view captured by a microscope of the stretchable MEA with the brain organoid⁴⁵. (with permission of Springer Nature). e Schematics of the generation of neural organoids from hiPS cells, the concept of vertically deformable Kirigami Electronics and the integration of neural organoids with electronics¹⁸⁰. f The spiral Kirigami Electronics design consists of concentric rings connected by spiral latches with 32 electrodes (gold circles) in the central area¹⁸⁰. (with permission of Springer Nature). g Schematics illustrating the stepwise integration of stretchable mesh nanoelectronics into 3D human iPSC-derived neural tissues through cell self-organization and brain organoids through organogenesis¹⁸². h Optical photograph of stretchable mesh nanoelectronics¹⁸². i Optical photograph of a 2 × 2 devices well, with a single culture chamber for four cyborg brain organoids cultured simultaneously¹⁸². (with permission of Wiley)

The stretchable mesh MEAs discussed previously primarily focus on monitoring neural activity at the surface of organoids, leaving a gap in the detection of deep internal signals within 3D structures. This limitation has been partly addressed by the cyborg system developed by Jia Li’s research team. The soft and stretchable mesh nanoelectronics employed in this system exhibit physicochemical properties that closely resemble those of tissue scaffolds, thereby achieving excellent biocompatibility and tissue-like bending stiffness. The implementation of serpentine mesh architecture further enhances its stretchable performance, enabling seamless integration with developing organoids. Through precise engineering modifications, the researchers successfully incorporated this system into both brain and cardiac organoid models. Specifically, the mesh-like nanoelectronics were transferred into 2D stem cells in advance and stretched continuously along with the 2D-to-3D reconfiguration of organogenesis. The engineered organoids ultimately form a 3D spherical morphology, with a uniform spatial distribution of the embedded sensors throughout the entire sphere. The tissue-like stretchable mesh nanoelectronics do not disturb the typical development of organoids and provide long-term stable electrophysiological recording of single cells with millisecond spatiotemporal resolution (Fig. 6g–i)^181,182.

In conclusion, the organoid model has emerged as a highly promising tool for exploring the intricacies of the human nervous system. To fully decode the intricate biological phenomena within these 3D structures, it is imperative to establish comprehensive electrophysiological monitoring capabilities that span from single-cell resolution activity to functional network dynamics. While planar electrodes, which have been extensively studied and applied, continue to play a significant role in organoid research, there remains an unmet need for MEAs specifically optimized for 3D architectures. The practice of organoid slicing and implantation of rigid electrodes, though effective in providing access to internal neural activity, is constrained by the introduction of inevitable mechanical damage. As an alternative, microneedle electrodes mitigate mechanical disruption and enable precise detection of electrical signals at specific depths within organoids, yet they sacrifice global coverage of the organoid. The advent of flexible and stretchable electronics, coupled with innovative engineering concepts, has redefined the trajectory of 3D MEA development. The emerging 3D MEAs is expected to focus on minimizing interference with the physiological activities of organoids while enabling long-term and stable data acquisition and current 3D MEAs have already demonstrated promising capabilities in spatial mapping of organoid activity. Moreover, given the long history and in-depth research on brain organoids, most optimized neural interfaces have thus far been tailored primarily for brain organoids. As investigations expand to diverse neural organoids, these advanced 3D MEAs will undoubtedly be adapted and refined to address the specific requirements of various neural models, thereby broadening their application scope in neuroscience research.

Impedance biosensors for the physiological recording of neural organoids

Electric cell-substrate impedance sensing (ECIS) is an assay to measure the dielectric properties of cells based on the principles of electrochemical impedance spectroscopy (EIS)¹⁸³. The dielectric properties can demonstrate cell behaviors like proliferation, metabolism, and viability. The impedance measurements have the advances of being non-invasive and label-free and can detect and record data sensitively and quantitatively. Nowadays, impedance measurement devices have been applied for cell research¹⁸⁴, cancer research¹⁸⁵, drug screening¹⁸⁶, environmental pollution¹⁸⁷, food safety¹⁸⁸, and so on.

In terms of neuroscience research, the ECIS system works multiple uses. For example, neurodegenerative diseases are common nowadays, causing a major threat to human health. Neurodegenerative disease is affecting all ages, ranging from congenital leukodystrophies in childhood to AD and Parkinson’s disease (PD) during aging. The diseases have genetic, environmental, or complex etiologies, resulting in obstacles in clinical treatment¹⁸⁹. As a result, it is critical to develop new tools for facilitating disease diagnosis and treatment. Tau hyperphosphorylation was reported implicated in a group of neurodegenerative diseases¹⁹⁰. The ECIS system is expected to monitor tau hyperphosphorylation and retrograde neurite degeneration. An early study conducted by Robitzki et al. applied impedance spectroscopy to measure the hyperphosphorylated tau protein in human neuroblastoma cell line SH-SY5Y for evaluating the pathological risk potential of AD¹⁹¹. To verify the prospect in vivo, Robitzki et al. induced tau hyperphosphorylation in excised organotypic hippocampal slice and monitored the real-time change with electrical impedance spectroscopy technique¹⁹². Researchers successfully detected two distinct phases of neurite degeneration that were reflected by different impedance data.

The distinguishing feature of neural organoid is the 3D structure and complex composition. The ECIS system has been applied to organoids for drug screening and the results showed higher drug resistance in 3D models^193,194. Due to poor adhesion between 3D organoids and planar electrodes, the expandable application of impedance sensors to 3D cultures still has many limitations. Our team has made a pioneering work to explore the ECIS system to 3D application. We developed a biosensing platform that involves a vertical impedance electrode array and a multi-channel detection system where 3D cultured cells were planted between a pair of gold electrodes for anti-cancer drugs testing^195,196 (Fig. 7a). The platform can record 3D cell impedance in the vertical direction and with the penetration of drugs, the survival rate of cells will exhibit temporal and positional dependencies. In addition, we’ve miniaturized the biosensing chip to a silicon chip, resulting in a multi-dimensional microgroove impedance sensor (MGIS)¹⁹⁷ (Fig. 7b). Similar to our work, Lei et al. reported a simple paper-based ECIS system in a 3D in vitro system to monitor cell viability by measuring impedance between 0.1 to 100 kHz¹⁹⁸. Although most available articles focus on cancer organoids, the ECIS system has broad application prospects on neural organoids due to the similar features.

Fig. 7: Schematics of some typical impedance electrode array for neural organoids.

a Schematics of the vertical impedance electrode array¹⁹⁶. (with permission of Elsevier). b Schematics of the MGIS chip¹⁹⁷. (with permission of Springer Nature). c Schematic illustration of ECIS system based real-time imaging and synchronous electrophysiological monitoring for OE organoids and the principle of impedance detection¹⁰³. (with permission of Elsevier)

Recently, we implemented an innovative strategy to investigate olfactory dysfunction in the early stages of neurodegenerative diseases using the ECIS system¹⁰³. The ECIS system serves as a non-invasive tool for monitoring olfactory epithelial organoids derived from basal stem cells. By integrating a live imaging system, we performed a real-time spatiotemporal analysis of the morphological and physiological features of olfactory epithelium organoids. To examine the early pathological changes associated with AD, we established olfactory epithelial organoids from APP/PS1 transgenic mice and observed their aberrant growth in both spatial and temporal dynamics. Abnormal pathogenic protein aggregation and accumulation was found in AD organoids and increased as disease progressed. This multimodal monitor strategy could help investigate the onset mechanisms of olfactory dysfunction, which would shed new light on early diagnosis and treatment of neurodegenerative disease (Fig. 7c).

The impedance sensing strategy is valuable in neural organoid research, although there is a limited number of published articles on the subject. A distinctive advantage of the ECIS system is its ability to enable researchers to monitor organoids in a real-time and non-invasive manner. This innovative monitoring tool is crucial for advancing neuroscience studies, particularly in the area of drug development. In the foreseeable future, the importance of the impedance biosensing strategy is expected to become increasingly evident.

Electrochemical biosensors for biomarker detection of neural organoids

In drug discovery studies involving organoid systems, tracking the metabolic activities of tissue-engineered organoids is crucial. Soluble biomolecules secreted by cells serve as indicators for assessing the functionality of ex vivo organoids and evaluating their response to drug compounds or toxic chemicals¹⁹⁹. Quantitative detection of biomarkers from complex biological media in ex vivo organoids holds great promise for monitoring cell status and the efficacy and toxicity of drugs directly²⁰⁰. Electrochemical sensors, which detect analytes by measuring changes in the electrical properties induced by chemical reactions on the electrode surface, play a key role in assessing the functionality of ex vivo neural organoids and their response to drug compounds or chemicals. Their outstanding characteristics include excellent detection limits, label-free detection capability, wide linear response range, portability, and suitability for continuous monitoring.

The detection of disease biomarkers in neural organoids is crucial for studying corresponding neurological disease models, while neurotransmitter release is essential for understanding synaptic connectivity and activity in neural organoid models. Therefore, electrochemical detection in neural organoids mainly focuses on detecting neurotransmitters or disease biomarkers. Taking Parkinson's disease as an example, α-synuclein is one of its biomarkers. Hung-Yin Lin et al. successfully developed a sensor based on a molecularly imprinted electrochemical polymer to detect increased levels of α-synuclein in brain organoid culture medium from Parkinson's disease patients (Fig. 8a)²⁰¹. They later improved the sensor's sensitivity by optimizing the template peptide sequence²⁰². Dopamine, an important neurotransmitter, and biomarker for Parkinson's disease, was successfully monitored in a human midbrain organoid culture medium using an electrochemical sensor strategy based on self-assembled 3-mercaptopropionic acid monolayer-modified electrodes and cyclic voltammetry signals (Fig. 8b)²⁰³. Serotonin is also a neurotransmitter, and its chemical similarity to dopamine makes its discrimination crucial. Tomi Laurila et al. successfully achieved selective detection of dopamine and serotonin at the nanomolar level in human brain organoid culture medium using a single-walled carbon nanotube-based electrochemical sensor, validating the healthy development of dopamine neurons in the midbrain on this sensor²⁰⁴. Glutamate is closely related to neural development, and Mirella Dottori et al. successfully detected glutamate release from human ESCs-derived brain organoids in different brain regions using a functionalized borosilicate glass capillary-based electrochemical biosensor²⁰⁵.

Fig. 8: Some typical electrochemical biosensors for neural organoids.

a Preparation of Peptide-Imprinted Poly(hydroxy 3,4-ethylenedioxythiophene)-Coated Electrodes and Their Recognition of α-Synuclein²⁰¹. (with permission of American Chemical Society). b Schematic of the improved electrochemical dopamine detection method²⁰³. (with permission of Royal Society of Chemistry)

These electrochemical sensors exhibit distinct sensitivity and selectivity limits in neurotransmitter detection. For instance, a single-walled carbon nanotube sensor further demonstrates nanomolar selectivity for dopamine and serotonin by optimizing detection potentials, addressing their close oxidation potentials²⁰⁴. Glutamate detection via a nanostructured microelectrode shows an LOD of 5.6 μM, with selectivity enhanced by enzyme-specific catalysis and polymeric film barriers, though interference from structural analogs remains²⁰⁵. Sensitivity limitations primarily arise from matrix fouling and competitive adsorption, while selectivity relies on electrochemical potential tuning, surface modification, and molecular imprinting strategies—approaches that have improved α-synuclein detection to 4.0 pM via peptide-imprinted nanotubes²⁰¹. These studies highlight that sensitivity is constrained by matrix fouling and competitive adsorption, while selectivity relies on electrochemical potential tuning, surface modification, and molecular recognition elements, with ongoing challenges in long-term stability and multi-analyte discrimination in complex environments.

Overall, electrochemical sensors for neural organoids are currently mainly used to detect chemical substances in the culture environment. Future directions may involve integrating them into organoid-on-a-chip platforms to achieve real-time monitoring of multiple biomolecules to assess organoid function and their response to drug compounds or chemicals. However, this field still faces various challenges, such as the biocompatibility of electrode materials, ultra-low detection limits, long-term continuous monitoring, simultaneous detection of multiple biomarkers, compatibility with microfluidic technologies, and highly complex culture environments.

Multi-model biosensors for comprehensive evaluation of neural organoids

The utilization of neural interfaces, with diverse morphologies, materials and functionalities, offers a perspective into specific aspects of neural organoids, facilitating a more detailed understanding of their structural and functional attributes. To further achieve a comprehensive and synchronous detection of the diverse attributes of neural organoids, the development of dual-mode and multimodal electronic interfaces and sensors represents an emerging frontier in the field.

The image-based detection method is a classic method for the visualization of internal spatial structure and assessment of developmental stages of organoids. The morphological and compositional similarity between various organoids and their corresponding human organs is one of the most direct evidence supporting their efficacy as a valid model in vitro. The morphological examination accompanied the growth process of neural organoids is instrumental in ascertaining the regulation of their normative growth patterns and in validating the efficacy of supplementary physiological data recorded. Certain widely utilized biocompatible substrate materials, including parylene, SU-8 and PDMS, exhibit inherent optical transparency characteristics, enabling the seamless integration of optical imaging modalities with electrophysiological recording techniques in neural interface systems. This unique material property facilitates the establishment of a novel dual-mode monitoring platform that concurrently captures both optical and electrophysiological neural activity with high spatial and temporal resolution²⁸. For instance, the optically transparent graphene microelectrode developed by Madison N. Wilson et al. serves as a sophisticated tool for chronically monitoring the developmental progression, maturation and functional integration of neural networks between human brain organoids implanted within the retrosplenial cortex of a mouse brain and the host's neural tissue. The transparent electrode can capture the propagation of neural activity at the interface between the organoid and the cortex and it can also provide a detailed visualization of the vascular network's integration of organoids within the host, assisted by two-photon imaging. This approach offers a comprehensive and multimodal assessment of the interaction between implanted organoids and the host brain, enhancing our understanding of the complex dynamics involved in neural transplantation and integration¹⁰⁹. Additionally, with the development of MEMS technology, micro-platform with compact structure has been designed for long-term monitoring on organoids models. Based on this, our team integrates the image detection and impedance detection into an e-plate, achieving the dynamic multimodal spatiotemporal monitoring of olfactory organoids in vitro¹⁰³. In addition, fluorescence imaging using cell-type-specific protein markers is a classic method for revealing the cellular composition within organoids. It facilitates the observation of the developmental process from stem cells to fully mature organoids at the cellular level, as well as subsequent apoptotic processes²⁰⁶. The combination of electrical recording and cell-type-specific imaging can accurately measure the electrophysiological changes on the surface of neural organoids to the single-cell level and relate the electrical signals with cell-type-specific information^207,208.

Neurons, as the fundamental structural and functional units of the nervous system, play a crucial role in sensing stimuli and conducting excitatory signals²⁰⁹. The communication between neurons relies on complex patterns of electrical and chemical signals²¹⁰. Specifically, electrophysiological signals reflect the excitability of neurons and their activity within neural networks, while chemical signals reveal the release and metabolic dynamics of neurotransmitters and other neurochemicals. These two signaling modes are interdependent and jointly regulate the operation of neural networks. Therefore, the development of dual-mode microelectrodes capable of simultaneously detecting electrophysiological and chemical signals is of great significance for understanding the functional mechanisms and pathological processes of the nervous system. In recent years, with the rapid advancement of materials science and microfabrication technologies, dual-mode microelectrodes have been widely applied in neuroscience research^211,212. For instance, Mao et al. successfully achieved simultaneous detection of ascorbic acid (AA) and electrophysiological signals in the rat brain by integrating carbon fiber microelectrodes with glass micropipette electrodes, providing a powerful tool for studying the relationship between neurochemistry and electrical activity²¹³. Additionally, Cai et al. designed a silicon-based microelectrode array (MEA) modified with platinum nanoparticles (PtNPs) and glutamate oxidase (Gluox), enabling the concurrent detection of glutamate (Glu) and electrophysiological signals in the rat hippocampus²¹⁴. This approach offers a novel methodology for exploring the dynamics of neurotransmitters and their correlation with electrophysiological activity. Furthermore, Anita A. Disney et al. developed a multi-channel ceramic-based MEA capable of recording electrophysiological and chemical signals from multiple brain regions simultaneously, providing critical technical support for investigating the spatiotemporal characteristics of neural networks²¹⁵. Although these studies primarily focus on in vivo experiments, their design principles and technical methodologies can also serve as valuable references for research on in vitro neural organoids. However, the application of dual-mode microelectrodes in neural organoid research still faces several challenges. First, the three-dimensional (3D) structure of neural organoids imposes higher demands on the spatial resolution of electrodes. Achieving high-resolution electrochemical detection in such complex 3D environments remains an area requiring further exploration²¹⁶. Second, the long-term stability of electrodes is crucial for ensuring the reliability of experimental data, particularly in the dynamic growth environment of organoids. The biocompatibility and mechanical stability of electrode materials need further optimization²¹⁷. Additionally, the interface design between electrodes and neural organoids requires improvement to reduce interfacial impedance and enhance signal transmission efficiency. By continuously optimizing electrode design and fabrication processes, dual-mode microelectrodes are expected to play a more significant role in functional studies, disease modeling, and drug screening of neural organoids, thereby advancing the field of neuroscience.

In conclusion, the image detection has high compatibility with other sensing techniques. The requirement for material transparency ensures the quality of image detection and the integration will not compromise the integrity of other sensing modalities, thus preserving the accuracy and reliability of the data collected. In the realm of multimodal sensing, the image detection serves as a foundational component, providing a visual interface that complements other analytical methods, enhancing the overall functionality of the system. Electrophysiology, as the most important domain in the study of neural organoids, is often assisted by electrochemical and electrical impedance sensing to enrich the dataset and investigate the fundamental principles governing the dynamics of change. The current research on dual-mode and multimodal electronic interfaces and sensors are predominantly concentrated on in vivo electrodes, which can be easily extended to neural organoids to obtain the inside information. Nevertheless, there are still some challenges. Signal interference presents a significant obstacle, as it can compromise the integrity of the data collected. The complexity of data analysis is also a considerable factor, as it requires sophisticated algorithms and computational resources to discern different signal sources. Additionally, there are inherent limitations in the manufacturing processes and associated costs, which can impede the widespread adoption and scalability of these multimodal sensing platforms. Achieving complete synchronization of signal acquisition and subsequent separate analysis remains an aspirational goal that necessitates further research and development. It is anticipated that with ongoing advancements in materials science, signal processing and manufacturing techniques, these challenges will be progressively addressed, leading to more refined and effective multimodal sensing devices capable of providing comprehensive understanding of the intricate organoid models.

Conclusion and outlook

Neural organoids, as an ideal in vitro model, have shown great potential in deciphering neural circuit connections and understanding the human body. To fully utilize this potential and promote the development of neurobiology, technologies need to be developed to reflect the architectures and functionalities of the organoids. In this review, we have summarized the latest advances in bioelectronic interfaces and sensors, focusing on their innovation in adapting to the 3D structure of organoids. The long-term, real-time, and accurate spatiotemporal description of 3D neural networks ensures the continuously expanding application scope of neural organoids in basic research and clinical treatment in the future.

The emerging bioelectronics has offered a unique biological interface with neural organoids. These technologies, each with its unique focus, encompass a comprehensive range of applications, extending from the surface to the interior of neural organoids and from the detection of fine neuronal activity to the characterization of the entire global physiological network of the organoid. Functionally integrated devices that combine existing technologies, including sequencing, imaging and computing, with electrical recording may further promote the development and fully excavate the potential of organoids model^103,208. Despite substantial advancements, current neural interface technologies still face notable challenges. For instance, the prolonged growth and development period of neural organoids complicates the establishment of a long-term stable neural interface without disrupting their natural developmental processes. The advent of flexible and adaptive materials offers promising solutions, as optimized electrodes can be designed to better mimic the properties of biological tissues, thereby minimizing mechanical mismatches and resultant damage^218,219. However, the long-term introduction of electrodes is inevitably associated with some degree of damage. Strategies such as employing biodegradable materials as electrodes or utilizing highly biocompatible and corrosion-resistant coatings can effectively minimize the adverse effects of foreign substances on biological systems, ensuring the sustained stability and functionality of neural interfaces^220,221,222.

In recent years, numerous researchers have focused their attention on incorporating localized and precisely controllable physical and chemical stimuli into neural interfaces, aiming to manipulate neural activity and investigate the underlying mechanisms of neural circuit dynamics. The closed-loop modulation and sensing system, which integrates both stimulation and sensing capabilities, is capable of not only delivering controllable electrical, optical, magnetic, and chemical signal stimulation to neurons but also detecting the electrical activity of target neurons in real time. Based on the feedback from neuronal responses, the stimulation parameters can be dynamically adjusted to achieve optimal modulation³². Electrical stimulation has garnered widespread attention due to its involvement in various key biological processes, including cell proliferation, differentiation, and migration²²³. For instance, commercially available closed-loop brain stimulation systems have been used for the treatment of neurological disorders such as PD and epilepsy^224,225. With the ongoing miniaturization of devices, these systems show promising potential for applications in safe and effective preclinical research with neural organoids. Optogenetic stimulation enables precise spatial and temporal control of neural activity through the targeted expression of light-sensitive proteins in specific neuronal populations²²⁶. For instance, opto-electronic neural probes have been developed to facilitate the analysis and high-resolution recording of signals from individual target neurons²²⁷. Additionally, chemical modulation, including drug delivery, ultrasound modulation and magnetic modulation techniques, have attracted significant research interest as potential therapeutic approaches for neurological disorders^31,228. This closed-loop regulatory system is beneficial for modulating neural plasticity, replacing damaged neurons and establishing new circuits within the nervous system. While current implementations primarily rely on implantable electrodes, with the increasing maturation of neural organoid models, this stimulus-sensing closed-loop regulatory system is likely to emerge as a prominent trend in the development of next-generation neural interface technologies for organoid applications.

The clinical translational potential of organoid technology has garnered increasing attention as an effective preclinical platform for evaluating and guiding personalized therapeutic strategies. In particular, patient-derived stem cell organoids have emerged as an importance tool in advancing personalized medicine. These organoids can faithfully recapitulate the histological architecture and mutational profiles of the original pathology, and neural interfaces can further visualize these disease-associated features²²⁹. Accumulating evidence from clinical trials demonstrates the successful application in cancer treatment and fetal congenital disorders^230,231. The synergistic combination of organoid platforms with neural interfaces holds significant promise for optimizing therapeutic regimens through high-throughput drug and combination screening. This approach not only enhances clinical decision-making but also contributes to improved patient outcomes and reduced healthcare expenditures. The limited maturity of current organoid systems remains a critical barrier to their broader scalability and clinical translation, particularly due to the absence of key physiological features such as functional vasculature and tissue-specific microenvironmental cues²³². The development of more sophisticated models that recapitulate these essential components will be crucial for achieving clinically relevant phenotypes. The organoid technology also holds significant promise for transplantation therapy through the use of patient-derived stem cell products. A notable advantage of this alternative approach is the absence of immune rejection, a feature that has encouraged ongoing clinical trials with considerable therapeutic potential. Nevertheless, due to technical and ethical limitations, comprehensive preclinical validation is necessary prior to clinical implementation^230,231. In this regard, implantable electrodes have made significant contributions to validating the functional connections formed between organoids and the host following transplantation¹⁰⁹.