This study was supported by institutional funds (DZNE), the Helmholtz Association within the Helmholtz Validation Fund (HVF-0102), and the Dresden International Graduate School for Biomedicine and Bioengineering (DIGS-BB). We would also like to acknowledge the platform for behavioral animal testing at the DZNE-Dresden (Alexander Garthe, Anne Karasinsky, Sandra G&#252;nther, and Jens Bergmann) for their support. We would like to acknowledge that a portion of Figure 1 was created using the platform BioRender.com.

Multimodal Large-Scale Neural Recordings and Analysis on HD-MEA 

The authors declare no competing or financial interests.

The intricate dynamics of spatiotemporal neuronal activity, emerging from interconnected neuronal ensembles, have long been a subject of intrigue in neuroscience. Traditional methodologies, such as patch-clamp, standard MEA, and Ca2+ imaging, have provided valuable insights into brain complexity. However, they often fall short in capturing the comprehensive network-wide computational dynamics21,22,23. The technical protocol of the HD-MEA platform, as detailed in this JoVE study, represents a significant leap forward, offering a panoramic view of neural dynamics across diverse modalities, from cell assemblies to expansive networks (i.e., acute, ex-vivo mouse brain slices and in-vitro human iPSC networks)26,29,30,32.
Acute, ex-vivo mouse brain slices have been a foundational tool in neuronal research, facilitating molecular and circuit-level investigations6,7. However, the challenge of maintaining tissue viability has been a persistent bottleneck. The protocol delineated in this study introduces critical modifications to optimize the quality and longevity of these slices to exploit their benefits on the HD-MEA platform. This protocol underscores the importance of - i) Achieving slice uniformity, for which the use of a vibratome is preferred over a tissue chopper due to its precision and minimized tissue damage, despite the trade-off of longer slicing times. ii) Ensuring constant carbogenation throughout the process, from extraction to recording, to maintain tissue viability. iii) Regulating temperature and allowing adequate recovery time before recording. iv) Utilizing an agarose block or mold to stabilize the brain, prevent tearing, and minimize glue contact. v) Maintaining optimal flow rates of carbogenated aCSF within the HD-MEA reservoir to ensure slice health while avoiding issues like decoupling, noise, and drift (Table 2).
For both mouse brain slices and human iPSC preparations, enhancing electrode-tissue interface coupling is paramount30,46,47. Our protocol underscores the importance of utilizing the adhesion-promoting molecule Poly-dl-ornithine (PDLO). This molecule not only augments the surface area for detecting electrical signals but also boosts electrical conductivity46. By doing so, it promotes cellular adhesion, growth, and the development of functional network properties. Such optimization plays a pivotal role in enhancing the efficacy of the HD-MEA platform. This, in turn, ensures accurate and consistent analysis of microscale ex-vivo and in-vitro connectomes and their spatiotemporal firing sequences. Notably, PDLO has been shown to outperform other substrates like polyethyleneimine (PEI) and poly-l-ornithine (PLO) in promoting spontaneous firing activity and responsiveness to electrical stimuli in neuronal cultures. Additionally, PDLO has been used for surface functionalization on the HD-MEA and shown to enhance the electrode-slice coupling interface and increase the signal-to-noise ratio in both OB and HC slices26,29. The addition of a custom-built platinum anchor further augments the electrode-slice interface coupling, leading to recordings with a higher signal-to-noise ratio.
The utilization of HD-MEA for both ex-vivo mouse brain slices and in-vitro human iPSC networks introduces a method adept at exploring extensive, multiscale, and multimodal dynamics. This innovative approach, however, brings forth considerable challenges, especially in data management48,49,50,51. A single HD-MEA recording acquired at 18 kHz/electrode sampling frequency generates a staggering 155 MB/s of data. The data volume escalates rapidly when factoring in multiple slices, diverse pharmacological conditions, or prolonged recording periods. Such an influx of information calls for robust storage infrastructures and advanced computational tools for streamlined processing. The ability of the HD-MEA platform to simultaneously gather data from thousands of neuronal ensembles is both a boon and a hurdle. It provides supreme insights into the computational dynamics of brain functions, yet it also necessitates a refined analytical framework. In this JoVE protocol, we have provided examples of computational strategies, including large-scale event detection, classification, graph theory, frequency analysis, and machine learning. These methods underscore the intensive efforts made to tackle the challenges of analyzing complex neural data. Nonetheless, there is still considerable room for the development of more advanced computational tools to analyze these multidimensional neural datasets. Armed with the appropriate tools and methodologies, the potential of the HD-MEA platform is magnified, offering profound insights into the intricacies of brain functions in both healthy and pathological conditions.
In essence, the HD-MEA platform, when integrated with the detailed protocols and computational tools discussed, offers a transformative approach to understanding the intricate workings of the brain. By capturing large-scale, multiscale, and multimodal dynamics, it provides invaluable insights into processes such as learning, memory, and information processing. Moreover, its application in in-vitro human iPSC networks has the potential to revolutionize drug screening and personalized medicine. However, while this platform represents a significant advancement in neuroscience research, it is crucial to acknowledge and address the inherent technical challenges. With ongoing refinement and the integration of advanced computational tools, the HD-MEA platform stands poised to usher in a new era of precise diagnostic tools, the identification of specific biomarkers, and targeted therapies for neurological disorders.

Neuronal ensembles, often termed cell assemblies, are pivotal in neural coding, facilitating intricate computations for processing multiscale neural information1,2,3. These ensembles underpin the formation of expansive neuronal networks and their nuanced microcircuits4. Such networks and their oscillatory patterns drive advanced brain functions, including perception and cognition. While extensive research has explored specific neuronal types and synaptic pathways, a deeper understanding of how neurons collaboratively form cell assemblies and influence spatiotemporal information processing across circuits and networks remains elusive5.
Acute, ex-vivo brain slices are pivotal electrophysiological tools for studying intact neural circuits, offering a controlled setting to probe oscillatory activity patterns of neural function, synaptic transmission, and connectivity, with implications in pharmacological testing and disease modeling6,7,8. This study protocol highlights two key brain circuits - the hippocampal-cortical (HC) involved in learning and memory processes9,10, and the olfactory bulb (OB) responsible for odor discrimination11,12,13. In these two regions, new functional neurons are continuously generated by adult neurogenesis throughout life in mammalian brains14. Both circuits demonstrate multidimensional dynamic neural activity patterns and inherent plasticity that participate in rewiring the existing neural network and facilitate alternative information processing strategies when required15,16.
Acute, ex-vivo brain slice models are indispensable for delving into brain functionality and understanding disease mechanisms at the microcircuit level. However, in-vitro cell cultures derived from human induced pluripotent stem cells (iPSCs) neuronal networks offer a promising avenue of translational research, seamlessly connecting findings from animal experiments to potential human clinical treatment17,18. These human-centric in-vitro assays serve as a reliable platform for assessing pharmacological toxicity, enabling precise drug screening, and furthering research into innovative cell-based therapeutic strategies19,20. Recognizing the pivotal role of the iPSC neuronal model, we have dedicated the third module of this protocol study to thoroughly investigate the functional characteristics of its derived networks and to fine-tune the associated cell culture protocols.
These electrogenic neural modules have been commonly studied using techniques like calcium (Ca2+ imaging), patch-clamp recordings, and low-density microelectrode arrays (LD-MEA). While Ca2+ imaging offers single-cell activity mapping, it is a cell-labeling-based method hindered by its low temporal resolution and challenges in long-term recordings. LD-MEAs lack spatial precision, while patch-clamp, being an invasive single-site technique and laborious, often yields a low success rate21,22,23. To address these challenges and effectively probe network-wide activity, large-scale simultaneous neural recordings have emerged as a pivotal approach for understanding the computational principles of neural dynamics underlying brain complexity and their implications in health and disease24,25.
In this JoVE protocol, we demonstrate a large-scale neural recording method based on the high-density MEA (HD-MEA) for capturing spatiotemporal neuronal activity across various brain modalities, including hippocampal and olfactory bulb circuits from ex-vivo mouse brain acute slices (Figures 1A-C) and in-vitro human iPSC-derived neuronal networks (Figures 1D-E), previously reported by our group and other colleagues26,27,28,29,30,31,32,33,34,35. The HD-MEA, built on complementary-metal-oxide-semiconductor (CMOS) technology, boasts on-chip circuitry and amplification, allowing sub-millisecond recordings across a 7mm2 array size36. This non-invasive approach captures multi-site, label-free extracellular firing patterns from thousands of neuronal ensembles simultaneously using 4096 microelectrodes at a high spatiotemporal resolution, revealing the intricate dynamics of local field potentials (LFPs) and multiunit spiking activity (MUA)26,29.
Given the vastness of the data generated by this methodology, a sophisticated analytical framework is essential, yet poses challenges37. We have developed computational tools that encompass automatic event detection, classification, graph theory, machine learning, and other advanced techniques (Figure 1F)26,29,38,39. Integrating the HD-MEA with these analytical tools, a holistic approach is devised to probe the intricate dynamics from individual cell assemblies to broader neural networks across diverse neural modalities. This combined approach deepens our grasp of the computational dynamics in normal brain functions and offers insights into anomalies present in pathological conditions28. Moreover, insights from this approach can propel advancements in brain-inspired modeling, neuromorphic computing, and neural learning algorithms. Ultimately, this method holds promise in uncovering the core mechanisms behind neural network disruptions, potentially identifying biomarkers, and guiding the creation of precise diagnostic tools and targeted treatments for neurological conditions.

<table border="1" fo:keep-together.within-page="1" fo:keep-with-next.within-page="always">
	<tbody>
		<tr>
			<td>150 mm Glass Petri Dish</td>
			<td>generic</td>
			<td>generic</td>
			<td>Brain Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>0.22 &#956;m Sterile Filter Unit&#160;</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Assorted</td>
		</tr>
		<tr>
			<td>90 mm Plastic Culture Dish</td>
			<td>TPP</td>
			<td>93100</td>
			<td>Brain Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Agarose</td>
			<td>Roth</td>
			<td>6351.5</td>
			<td>Brain Preparation Workspace</td>
		</tr>
		<tr>
			<td>Agarose Mold</td>
			<td>CUSTOM</td>
			<td>CUSTOM</td>
			<td>Brain Preparation Workspace; Custom designed 3D Printer Design, available upon request</td>
		</tr>
		<tr>
			<td>Aluminum Foil</td>
			<td>generic</td>
			<td>generic</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>Anesthesia chamber</td>
			<td>generic</td>
			<td>generic</td>
			<td>Brain Extraction Workspace; Assorted Beaker, Bedding etc</td>
		</tr>
		<tr>
			<td>Ascorbic Acid</td>
			<td>Sigma Aldrich</td>
			<td>A4544-25G</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Assorted Beakers</td>
			<td>generic</td>
			<td>generic</td>
			<td>Solution Preparation Workspace; 50 mL</td>
		</tr>
		<tr>
			<td>Assorted Luers</td>
			<td>Cole Parmer</td>
			<td>45511-00</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Assorted Volumetric flasks</td>
			<td>generic</td>
			<td>generic</td>
			<td>Solution Preparation Workspace; 500 mL, 1 L</td>
		</tr>
		<tr>
			<td>B27 Supplement</td>
			<td>Life Technologies</td>
			<td>17504-044</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>BDNF</td>
			<td>Peprotech</td>
			<td>450-02</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Biological Safety Cabinet with UV Lamp</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>HD-MEA Coating, Plating, Mainainance Workspace</td>
		</tr>
		<tr>
			<td>BrainPhys Neuronal Medium&#160;</td>
			<td>STEMCELL Technologies&#160;</td>
			<td>05790&#160;</td>
			<td>CDI, and BrainXell Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Brainwave Software</td>
			<td>3Brain AG</td>
			<td>Version 4</td>
			<td>Brain Slice and Human iPSC Recording Workspace</td>
		</tr>
		<tr>
			<td>BrainXell&#160;Glutamatergic Neuron Assay&#160;</td>
			<td>BrainXell</td>
			<td>&#160;BX-0300&#160;</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>CaCl2</td>
			<td>Sigma Aldrich</td>
			<td>21115-100ML</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Carbogen</td>
			<td>generic</td>
			<td>generic</td>
			<td>All Workspaces; 95%/5% O2 and CO2 mixture</td>
		</tr>
		<tr>
			<td>Cell Culture Incubator&#160;</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Assorted</td>
		</tr>
		<tr>
			<td>CMOS-based HD-MEA chip</td>
			<td>3Brain AG</td>
			<td>CUSTOM</td>
			<td>Brain Slice and Human iPSC Recording Workspace</td>
		</tr>
		<tr>
			<td>Conical Tubes, 50 mL, Falcon (Centrifuge Tubes)&#160;</td>
			<td>STEMCELL Technologies&#160;</td>
			<td>38010&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Crocodile Clip Grounding Cables</td>
			<td>JWQIDI</td>
			<td>B06WGZG17W</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Curved Forceps</td>
			<td>FST</td>
			<td>11052-10</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>DMEM/F12 Medium</td>
			<td>Life Technologies</td>
			<td>&#160;11330-032</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Dulbecco&#8217;s Phosphate Buffered Saline without Ca2+ and Mg2+ (D-PBS)</td>
			<td>STEMCELL Technologies&#160;</td>
			<td>37350&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Filter Paper</td>
			<td>Macherey-Nagel</td>
			<td>531 011</td>
			<td>Brain Preparation Workspace</td>
		</tr>
		<tr>
			<td>Fine Brush</td>
			<td>Leonhardy</td>
			<td>773</td>
			<td>Brain Slice Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Forceps</td>
			<td>VITLAB</td>
			<td>67895</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>GDNF</td>
			<td>Peprotech</td>
			<td>450-10</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Geltrex</td>
			<td>Life Technologies</td>
			<td>A1413201</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Glass pasteur pipette</td>
			<td>Roth</td>
			<td>4518</td>
			<td>Brain Slice Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Glucose</td>
			<td>Sigma Aldrich</td>
			<td>G7021-1KG</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>GlutaMAX</td>
			<td>Life Technologies</td>
			<td>35050-061</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Gravity-based Perfusion System</td>
			<td>ALA</td>
			<td>VC3-8xG</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>HD-MEA Recording platform</td>
			<td>3Brain AG</td>
			<td>CUSTOM</td>
			<td>Brain Slice and Human iPSC Recording Workspace</td>
		</tr>
		<tr>
			<td>Heater</td>
			<td>Warner Instruments</td>
			<td>TC-324C</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Hemocytometer or Automated Cell Counter</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>HD-MEA Coating, Plating, Mainainance Workspace</td>
		</tr>
		<tr>
			<td>Hypo Needles</td>
			<td>Warner Instruments</td>
			<td>641489</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>iCell GlutaNeurons Kit, 01279&#160;</td>
			<td>CDI</td>
			<td>R1061&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Iris Scissors</td>
			<td>Vantage</td>
			<td>V95-304</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>Isoflurane</td>
			<td>Baxter</td>
			<td>HDG9623</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>KCl</td>
			<td>Sigma Aldrich</td>
			<td>P5405-250G</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Laminin&#160;</td>
			<td>Sigma-Aldrich&#160;</td>
			<td>L2020&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Liquid Nitrogen Storage Unit&#160;</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>HD-MEA Coating, Plating, Mainainance Workspace</td>
		</tr>
		<tr>
			<td>Magnetic Stirrer</td>
			<td>generic</td>
			<td>generic</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Metal Screws</td>
			<td>Thorlabs</td>
			<td>HW-KIT2/M</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>MgCl2</td>
			<td>Sigma Aldrich</td>
			<td>M1028-100ML</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>MgSO4</td>
			<td>Sigma Aldrich</td>
			<td>63138-250G</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Microdissection Tool Holder&#160;</td>
			<td>Braun</td>
			<td>4606108V</td>
			<td>Brain Slice Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Microdissection Tool Needle</td>
			<td>Braun&#160;</td>
			<td>9186166</td>
			<td>Brain Slice Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Modular Stereomicroscope</td>
			<td>Leica</td>
			<td>CUSTOM</td>
			<td>Brain Slice Recording Workspace; custom specifications and modifications</td>
		</tr>
		<tr>
			<td>N2 Supplement</td>
			<td>Life Technologies</td>
			<td>17502-048</td>
			<td>CDI, and BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>NaCl</td>
			<td>Sigma Aldrich</td>
			<td>S3014-1KG</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>NaH2PO4</td>
			<td>Sigma Aldrich</td>
			<td>S0751-100G</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>NaHCO3</td>
			<td>Sigma Aldrich</td>
			<td>S5761-500G</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Neurobasal Medium&#160;</td>
			<td>Life Technologies</td>
			<td>21103-049</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Optical Cage System</td>
			<td>Thorlabs</td>
			<td>Assorted</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Optical Table w/Breadboard</td>
			<td>Thorlabs</td>
			<td>SDA7590</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>PDLO</td>
			<td>Sigma Aldrich</td>
			<td>P0671</td>
			<td>HD-MEA Coating, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Penicillin-streptomycin, 100x&#160;</td>
			<td>Thermo Fisher Scientific&#160;</td>
			<td>15140-122&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Pipette tips</td>
			<td>TipONE</td>
			<td>S1120-8810</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Pipettors&#160;</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Assorted</td>
		</tr>
		<tr>
			<td>Platinum Anchor</td>
			<td>CUSTOM</td>
			<td>CUSTOM</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Polyethylene Tubing</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Pump</td>
			<td>MasterFlex</td>
			<td>78018-22</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Razor Blade</td>
			<td>Apollo</td>
			<td>10179960</td>
			<td>Brain Preparation Workspace</td>
		</tr>
		<tr>
			<td>Reference Electrode Cell Culture Cap</td>
			<td>CUSTOM</td>
			<td>CUSTOM</td>
			<td>Human iPSC Recording Workspace; Custom designed 3D Printer Design, available upon request</td>
		</tr>
		<tr>
			<td>Rubber Pipette Bulb</td>
			<td>Duran Wheaton Kimble</td>
			<td>292000205</td>
			<td>Brain Slice Preparation Workspace, Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Serological Pipettes, 1 mL, 2 mL, 5 mL, 10 mL, 25 mL</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Assorted</td>
		</tr>
		<tr>
			<td>Slice Recovery Chamber</td>
			<td>CUSTOM</td>
			<td>CUSTOM</td>
			<td>Brain Slice Recovery Workspace; Custom designed 3D Printer Design, available upon request</td>
		</tr>
		<tr>
			<td>Spatula</td>
			<td>ISOLAB</td>
			<td>047.06.150</td>
			<td>Brain Preparation Workspace</td>
		</tr>
		<tr>
			<td>Sucrose</td>
			<td>Sigma Aldrich</td>
			<td>84100-1KG</td>
			<td>Solution Preparation Workspace</td>
		</tr>
		<tr>
			<td>Super Glue</td>
			<td>UHU</td>
			<td>358221</td>
			<td>Brain Slice Preparation Workspace</td>
		</tr>
		<tr>
			<td>Surgical Scissors</td>
			<td>Peters Instruments</td>
			<td>BC 344</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>Tabletop Centrifuge&#160;</td>
			<td>Assorted</td>
			<td>Assorted</td>
			<td>Assorted</td>
		</tr>
		<tr>
			<td>TGF-&#946;1</td>
			<td>Peprotech</td>
			<td>100-21C</td>
			<td>BrainXell Commercial Supplier Protocol</td>
		</tr>
		<tr>
			<td>Tissue Paper</td>
			<td>generic</td>
			<td>generic</td>
			<td>Brain Extraction Workspace</td>
		</tr>
		<tr>
			<td>Trypan Blue</td>
			<td>STEMCELL Technologies&#160;</td>
			<td>07050&#160;</td>
			<td>CDI Commerical Supplier Protocol</td>
		</tr>
		<tr>
			<td>Upright Microscope</td>
			<td>Olympus</td>
			<td>CUSTOM</td>
			<td>Imaging Workspace; Custom specifications and modifications</td>
		</tr>
		<tr>
			<td>Vacusip</td>
			<td>Integra</td>
			<td>159010</td>
			<td>Brain Slice Recording Workspace</td>
		</tr>
		<tr>
			<td>Vibratome</td>
			<td>Leica</td>
			<td>VT1200s</td>
			<td>Brain Slice Preparation Workspace; Includes: Specimen plate, buffer tray, ice tray, specimen plate holding tool, vibratome blade adjusting tool</td>
		</tr>
		<tr>
			<td>Vibratome Blade</td>
			<td>Personna</td>
			<td>N/A</td>
			<td>Brain Slice Preparation Workspace</td>
		</tr>
		<tr>
			<td>Water Bath</td>
			<td>Lauda</td>
			<td>L000595</td>
			<td>Brain Slice Recovery Workspace</td>
		</tr>
	</tbody>
</table>

recording and analyzing multimodal large-scale neuronal ensemble dynamics on cmos-integrated high-density microelectrode array

Large-scale neuronal networks and their complex distributed microcircuits are essential to generate perception, cognition, and behavior that emerge from patterns of spatiotemporal neuronal activity. These dynamic patterns emerging from functional groups of interconnected neuronal ensembles facilitate precise computations for processing and coding multiscale neural information, thereby driving higher brain functions. To probe the computational principles of neural dynamics underlying this complexity and investigate the multiscale impact of biological processes in health and disease, large-scale simultaneous recordings have become instrumental. Here, a high-density microelectrode array (HD-MEA) is employed to study two modalities of neural dynamics - hippocampal and olfactory bulb circuits from ex-vivo mouse brain slices and neuronal networks from in-vitro cell cultures of human induced pluripotent stem cells (iPSCs). The HD-MEA platform, with 4096 microelectrodes, enables non-invasive, multi-site, label-free recordings of extracellular firing patterns from thousands of neuronal ensembles simultaneously at high spatiotemporal resolution. This approach allows the characterization of several electrophysiological network-wide features, including single/-multi-unit spiking activity patterns and local field potential oscillations. To scrutinize these multidimensional neural data, we have developed several computational tools incorporating machine learning algorithms, automatic event detection and classification, graph theory, and other advanced analyses. By supplementing these computational pipelines with this platform, we provide a methodology for studying the large, multiscale, and multimodal dynamics from cell assemblies to networks. This can potentially advance our understanding of complex brain functions and cognitive processes in health and disease. Commitment to open science and insights into large-scale computational neural dynamics could enhance brain-inspired modeling, neuromorphic computing, and neural learning algorithms. Furthermore, understanding the underlying mechanisms of impaired large-scale neural computations and their interconnected microcircuit dynamics could lead to the identification of specific biomarkers, paving the way for more accurate diagnostic tools and targeted therapies for neurological disorders.

Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology

Recording and Analyzing Multimodal Large-Scale Neuronal Ensemble Dynamics on CMOS-Integrated High-Density Microelectrode Array

Multimodel spatiotemporal mapping and extraction of oscillatory firing features 
To quantify network-wide LFP and spike events that emerged from dynamical neuronal ensembles, we investigated synchronous large-scale firing patterns in HC and OB circuits and human iPSC networks. Recorded brain slice circuits from step 3.2 and recorded iPSC networks from step 3.3 were analyzed according to steps 4.1-4.2 of the protocol. First, event detection and denoising were performed for all recorded datasets and regionally resolved according to circuit specifications. Next, topographical pseudo-color spatial mapping of mean large-scale LFP and spike firing patterns, rastergrams of detected events, and representative 5-s traces of filtered waveforms were plotted (Figures 3A-I). Topographical pseudo-color mapping of large-scale LFP and spike firing rate patterns were overlaid on the respective microscope-captured optical images of HC (Figure 3A), OB (Figure 3B), and human iPSC neuronal network (Figure 3C). This allows the investigation of individual circuit and network-based oscillatory patterns and responses. HC and OB rastergrams contain detected LFP event counts sorted over the DG, Hilus, CA3, CA1, EC, and PC layers of the HC circuit and ONL, OCx, GL, PL, and GCL layers of the OB network over a 60-s time bin (Figures 3D,E). The human iPSC rastergram displays synchronous detected spike events of the interconnected cultured network over a 20 s time bin (Figure 3G). Next, 5s representative event traces from large-scale HD-MEA recording sites show a range of recorded oscillatory frequencies in the HC (i.e., selected electrode in CA3) (Figure 3G) and OB (i.e., selected electrode in GL) (Figure 3H) circuits and multiunit spike bursting activity in the human iPSC network from four selected active electrodes in the array (Figure 3I). These exemplary signals show biosignal signatures, including low-frequency LFP oscillations (1-100 Hz) with bandpass filtered &#948;, &#952;, &#946;, and &#947; frequency bands; sharp wave ripples (SWR) (140-220 Hz); and high-frequency single and MUA (300-3500 Hz). Finally, power spectral density (PSD) analysis was employed to simultaneously quantify a specific oscillatory band&#39;s power magnitude in the interconnected HC and OB circuit recorded from HD-MEA (Figures 3J,K).
Multimodal Network-wide Functional Connectome 
To infer the large-scale connectivity of multilayered neural networks from simultaneously firing patterns of concurrently active neuronal ensembles, the cross-covariance between pairs of active electrodes in detected events was calculated according to step 4.2.6 of the protocol. Here, the correlation coefficient was sorted based on layers in the HC and OB circuit or unsorted in the iPSC network and then stored in a symmetric matrix. Functional connectomes of HC and OB circuitry were generated by applying Multivariate Granger causality and directed transfer function (DTF) to quantify the influence of one time series on another and assess the directional information flow within the correlated links in the distinct networks. Connectome mapping of HC (Figure 4A) and OB (Figure 4B) and network visualization were performed using the Gephi program 9.2 version (https://gephi.org). Similar parameter constraints were placed on the functional links to compare the HC and OB brain slice circuits and illustrated 100 s of the functional connectivity of detected LFP events. Nodes are scaled according to degree strength with nodal color indicating layer and link color identifying the intra- and inter-layer connections. Functional connectomes of human iPSC networks were generated by applying spatial-temporal filters (STF) and distance-dependent latency thresholds (DdLT) to enhance the selection of significant links and refine the identification of meaningful connections by applying filtered and normalized cross-correlation histogram (FNCCH) analysis. Connectome mapping of human iPSC networks on the entire HD-MEA chip (Figure 4C) visualization performed using Gephi. Nodal color indicates excitatory or inhibitory input, and link color identifies connections.
<img alt="Figure 1" class="xfigimg" src="/files/ftp_upload/66473/66473fig01v2.jpg" /> 
Figure 1: Overview of the experimental and computational platform on large-scale HD-MEA. (A) Isometric schematic representation of our multimodal biohybrid neuroelectronic platforms realized with CMOS-based HD-MEA to capture neural dynamics from HC, OB, and human iPSC neuronal circuits and networks. (B) Schematic workflow for mouse brain slicing and its workscape to obtain HC and OB slices. (C) Topographical representations of the large-scale firing patterns recorded simultaneously from the entire HC and OB slices superimposed with the extracted extracellular waveforms to the slice optical images. (D) Schematic representation of iPSC neuronal network obtained from humans. (E) Fluorescence micrographs showing cellular c-fos and somatic/dendritic MAP-2 of the entire human neuronal network on HD-MEA chip (left) matched with the entire average firing activity map (right). (F) Computational framework including advanced data analysis, connectivity mapping, and AI-machine learning tools to analyze multidimensional neural data obtained from large-scale recordings on HD-MEAs. <a href="https://www.jove.com/files/ftp_upload/66473/66473fig01large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 2" class="xfigimg" src="/files/ftp_upload/66473/66473fig02v2.jpg" /> 
Figure 2: Layouts for ex-vivo brain slice and in-vitro human iPSC culture preparation and recording workspaces. (A) Schematic workflow illustrating the setup for preparing HC and OB slices, featuring the requisite tools and equipment in each workspace. (B)&#160;Schematic representation for human iPSC culture preparation, including the necessary tools and devices. A complete list of materials is included in steps 1.2.2, 2.1, 2.2, 3.1.1, 3.3, and the Table of Materials. <a href="https://www.jove.com/files/ftp_upload/66473/66473fig02large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 3" class="xfigimg" src="/files/ftp_upload/66473/66473fig03v2.jpg" /> 
Figure 3: Mapping and extracting spatiotemporal patterns of network dynamics. (A-C)&#160;Mean LFP and spike rate spatial maps, computed over five-minute recordings, superimposed on the microscope light image. (D-F)&#160;Raster plots depicting detected, denoised LFP events in a 60-second data subsample and spikes in a 20-second data subsample. (G-I) Representative waveform trace extraction from a 5-second segment of the raster plot data subsample (highlighted red in the raster plot), displayed as raw LFP oscillatory bands (1-100 Hz); &#948; (1-4 Hz), &#952; (5-12 Hz), &#946; (13-35 Hz), and &#947; (35-100 Hz) frequency bands; SWR (140-220 Hz); and high-frequency single and MUA spiking (300-3500 Hz). (J,K) Power spectral density maps of fast and slow oscillatory LFPs (1-100 Hz) and SWR (140-220 Hz). <a href="https://www.jove.com/files/ftp_upload/66473/66473fig03large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
<img alt="Figure 4" class="xfigimg" src="/files/ftp_upload/66473/66473fig04v2.jpg" /> 
Figure 4: Organization of multimodal network-wide functional connectomes. (A-C) Gephi maps illustrating nodal functional connectivity, where nodes correspond to one of the example color bar legends (below), while the links (or edges) are shaded to match the connecting nodes. Example legends for (A) HC, (B) OB, and (C) iPSC layers are displayed on a 64 x 64 array. HC and OB layers are plotted over a 100-s time bin to effectively reduce the number of visible nodes and links for visualization purposes. <a href="https://www.jove.com/files/ftp_upload/66473/66473fig04large.jpg" target="_blank">Please click here to view a larger version of this figure.</a>
Table 1: Solutions for brain slice preparation and media for iPSC neuronal cultures. (A) High-sucrose cutting solution for ex-vivo brain slice preparation. (B) aCSF recording solution for ex-vivo brain slice preparation and recording. (C-D) Human neuronal iPSC Media Protocol, where (C) is BrainPhys complete media used for cell thawing, HD-MEA chip coating, and cultured HD-MEA maintenance, and (D) the dotting media used for HD-MEA cell plating. <a href="https://www.jove.com/files/ftp_upload/66473/Table 1.xlsx" target="_blank">Please click here to download this Table.</a>
Table 2: Troubleshooting common HD-MEA recording acquisition issues. A list of common problems, their potential causes, and troubleshooting solutions related to HD-MEA chips, recording platform, system noise, and software. <a href="https://www.jove.com/files/ftp_upload/66473/Table 2.xlsx" target="_blank">Please click here to download this Table.</a>

Watch this Scientific Journal Video about Author Spotlight: Advancing Large-Scale Neural Dynamics Through HD-MEA Technology at JoVE.com

Here, we employ HD-MEA to delve into computational dynamics of large-scale neuronal ensembles, particularly in hippocampal, olfactory bulb circuits, and human neuronal networks. Capturing spatiotemporal activity, combined with computational tools, provides insights into neuronal ensemble complexity. The method enhances understanding of brain functions, potentially identifying biomarkers and treatments for neurological disorders.

Large-scale neuronal networks and their complex distributed microcircuits are essential to generate perception, cognition, and behavior that emerge from ...

recording-analyzing-multimodal-large-scale-neuronal-ensemble-dynamics

Research

JoVE Journal

Neuroscience

2.5K Views.  German Center for Neurodegenerative Diseases (DZNE). Here, we employ HD-MEA to delve into computational dynamics of large-scale neuronal ensembles, particularly in hippocampal, olfactory bulb circuits, and human neuronal networks. Capturing spatiotemporal activity, combined with computational tools, provides insights into neuronal ensemble complexity. The method enhances understanding of brain functions, potentially identifying biomarkers and treatments for neurological disorders.