Temporal Dynamics of Neocortical Development in Organotypic Mouse Cultures: A Comprehensive Analysis

Murine organotypic brain slice cultures have been widely used in neuroscientific research and are offering the opportunity to study neuronal function under normal and disease conditions. Despite the brought application, the mechanisms governing the maturation of immature cortical circuits in vitro are not well understood. In this study, we present a detailed investigation into the development of the neocortex in vitro. Utilizing a holistic approach, we studied organotypic whole-hemisphere brain slice cultures from postnatal mice and tracked the development of the somatosensory area over a five-week period. Our analysis revealed the maturation of passive and active intrinsic properties of pyramidal cells together with their morphology, closely resembling in vivo development. Detailed Multi-electrode array (MEA) electrophysiological assessments and RNA expression profiling demonstrated stable network properties by two weeks in culture, followed by the transition of spontaneous activity towards more complex patterns including high-frequency oscillations. However, weeks 4 and 5 exhibited increased variability and initial signs of neuronal loss, highlighting the importance of considering developmental stages in experimental design. This comprehensive characterization is vital for understanding the temporal dynamics of the neocortical development in vitro, with implications for neuroscientific research methodologies, particularly in the investigation of diseases such as epilepsy and other neurodevelopmental disorders. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Introduc.on Due to the limita@ons of experimental animal models, which are cost s, labor and ethical concerns on animal numbers and suffering, many advances have been made during recent years to improve in vitro neurological models to be?er mimic the in vivo situa@on. Although induced pluripotent stem cells (iPSCs) and organoids have been a big step in the right direc@on and provide the advantage of recapitula@ng human development in simplified neural networks1, par@ally preserve gene@c pa@ent characteris@cs 2 and eliminate the ethical concerns @ed to the use of embryonic stem cells 3, they unfortunately s@ll encounter the problem of impaired matura@on and lack of cellular diversity and structural complexity 4,5. While acute brain slices are a viable op@on to inves@gate the rodent brain at different ages rela@vely close to the in vivo situa@on, they unfortunately do not enable gene@c modifica@on, Figure 1, Graphical abstract. Schema'c overview of experimental setup. (A) Cor9cal organotypic brain slice cultures were created from P5-6 wildtype mice and cultured on semiporous membrane inserts in controlled medium. Experiments were performed con9nuously over a 9me course of 5 weeks. Development of single-cell firing proper9es and network events displayed along the 9meline. (B) Electrophysiological data was obtained by performing whole-cell Patch-Clamp recordings and Mul9-electrode array (MEA) recordings. (C) Real-9me PCR was performed on cDNA samples created from isolated RNA from cultures to monitor RNA expression levels. (D) Na9ve cultures were stained for cellular expression markers and neurons filled during Patch-Clamp recordings were stained and reconstructed for morphological analysis. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint long-term s@mula@on or pharmacological manipula@on. Although these methods can be applied in the living animal, in vitro prepara@ons offer more precisely controlled experimental condi@ons. Organotypic brain slice cultures (oBSCs) close this gap by providing an experimental @me frame of mul@ple weeks along with intact three-dimensional cytoarchitecture as well as inclusion of all cellular subtypes of the brain . oBSCs have been widely adapted in neuroscien@fic research since they were first established by Gähwiler6 with the ini@al roller tube culturing technique. Though this method is robust and yields stable cultures for mul@ple weeks7–11, the thinning to 2-3 cell layers and therefore loss of the original spa@al organiza@on might be preferable for applica@ons where visual accessibility is favored, but not for electrophysiological, morphological and developmental studies that require int act three - dimensional architecture of certain brain areas. As a consequence, the method was further adapted by Stoppini et al. 12, introducing a semiporous membrane culturing method at air - liquid-interface to preserve the original structure of brain prepara@ons. Both methods have been shown to maintain the organotypic structure, their normal cell morphology and to generally correlate well with development observed in vivo 7,9,13–17 and have been used successfully in the past, inves@ga@ng , for example, neurodegenera@on18,19, ischemia 20–23, interneuron development7,10, axon forma@on and dynamics24,25 and network dynamics15,26,27. In addi@on, oBSCs have been used as a screening plaborm for novel therapy targets and therapeu@cs14. Especially prepara@ons of the rat and mouse hippocampus have been used and characterized widely 11,28–33 with only a limited number of studies repor@ng the use of other brain areas such as cerebellum 13,15,34,35, striatum 10,36 or cortex 7,9,16,37–39. Neocor@cal oBSCs are of special interest when it comes to the inves@ga@on of inhibi@on and excita@on balancing, developmental network dynamics, cell -cell rela@onships as well as pathogenesis in the maturing cortex. Unfortunately, to date, informa@on on the @ssue development ex vivo is scarce and has only been reported in cultures created by the ini@al roller tube technique7– 10,25,40–42 or in studies only focusing on specific parameters24,37–39,43. This study aimed to close this gap of knowledge by providing a holis@c characteriza@on of the development of membrane insert cultured mouse neocor@cal oBSCs over a @me course of five weeks, inves@ga@ng single cell and network electrophysiology, RNA expression levels and morphology of pyramidal cor@cal neurons, as this informa@on will hold vital indica@ons on which @meframes of culturing are suitable for different types of experimental approaches. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint

& Methods Animals All procedures involving animals were performed in accordance with the guidelines of the Federa@on of European Laboratory Animal Science Associa@on, the EU Direc@ve 2010/63/EU, and the German animal welfare law. For harves@ng of brain @ssue for organotypi c cultures, wildtype C57BL/6J mice of either sex aged 5-7 postnatal days were used. Mice were obtained from the local Ins@tute of Laboratory Animal Science RWTH University Clinic Aachen and housed at a 12h light-dark cycle with food and water ad libitum. Tissue harvest was performed aher decapita@on of mice. Slice prepara.on Aher mouse decapita@on, the brain was carefully extracted from the skull using microdissec@on tools. Hemispheres were separated and glued on solidified agarose blocks with rostral side facing upwards. Brain @ssue was sliced at 350 µm and whole-hemisphere slices were used for cul@va@on. Brain slices were generated using a vibratome (Leica VT1200S) filled with ice -cold ar@ficial cerebrospinal fluid (aCSF), containing: 125 mM NaCl, 25 mM NaHCO3, 2.5 mM KCl, 1.25 NaH2PO4, 1 mM MgCl2 * 6 H2O, 2 mM CaCl2, 25 mM glucose and 1% penicillin/streptomycin (Gibco) dissolved in Millipore water. aCSF was constantly perfused with carbogen gas (95% O 2 and 5% CO 2) and kept cold by reusable ice cubes for appropriate hypothermia and oxygena@on during prepara@on. For all experiments except RT -PCR, somatosensory cor@cal regions of whole mouse brain slices were used and analyzed. Organotypic brain slice cultures Slices were placed on semipermeable membrane inserts (pore size 0.4 µm, Millipore), adhered by removing excess liquid and placed in 6 -well plates at air -liquid interface on 1.2 ml of controlled medium containing: 20% horse serum, 1 mM L-glutamine, 0.00125% ascorbic acid, 0.001 mg/ml insulin, 1 mM CaCl 2, 2 mM MgSO 4, 13 mM glucose and 1% penicillin/streptomycin in minimum essen@al medium at an osmolarity of 320 mOsm/l and pH of 7.28. Cultures were maintained in an incubator with controlled atmosphere (37°C, 5% CO2, 100% humidity) and full medium change was performed every two days. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Time periods All experiments on organotypic brain slice cultures were performed in a @me frame of 1 -37 days in vitro. For Patch-Clamp experiments, periods are defined as ini@al (1-2 DIV), week 1 (6- 9 DIV), week 2 (11 -16 DIV), week 3 (17 -23 DIV), week 4 (28 -30 DIV) and week 5 (34 -37 DIV). Remaining experiments were performed on the exact corresponding day, meaning ini@al (2 DIV), week 1 (7 DIV), week 2 (14 DIV), week 3 (21 DIV), week 4 (28DIV) and week 5 (35 DIV), +/- 1 DIV. Patch-Clamp recordings Whole-cell Patch-Clamp recordings on mouse organotypic brain slice cultures were performed at various @me points with cultures being in vitro between 2 – 37 days. For this cause, brain slices were excised from the cul@va@on membrane and transferred to the recording bath. During recordings, slices were con@nuously perfused with carbogenated aCSF at 30°C at a perfusion rate of approximately 7 ml/min. Recording pipe?es were pulled from borosilicate glass capillaries (World Precision Instruments) with a resistance of 3 -7 MΩ. For recordings, pipe?es were filled with intracellular solu@on containing: 140 mM K-gluconic acid, 1 mM CaCl2, 10 mM EGTA, 2 mM MgCl 2, 4 mM Na 2ATP , 10 mM HEPES and 5 mg/ml biocy@n (Sigma). Neurons were patched within the somatosensory cor@cal region of whole mouse brain slices, approximately between 100 – 500 µm from the cor@cal border. Pyramidal cells were iden@fied by their soma@c shape and firing proper@es. Whole-cell recordings were performed using an EPC10 amplifier (HEKA) and Patchmaster sohware (HEKA). Signals were sampled at 10 kHz, filtered at 1 kHz and analyzed using Fitmaster sohware (HEKA) and custom wri?en MatLab (MathWorks) scripts. Aher recordings, slices were fixed in 4% PFA (Morphisto) for 10-24 h. Mul.-electrode array recordings Mul@-electrode array (MEA) recordings on mouse organotypic brain slice cultures were performed at @me points 2 DIV, 7 DIV, 14 DIV, 21 DIV, 28 DIV and 35 DIV. Brain slices were excised from the cul@va@on membrane and transferred onto the MEA chip (256MEA30/8iR- ITO-pr, 16x16 electrodes, electrode size 30 μm, 200 μm spacing, Mul@ Channel Systems MCS GmbH) with the electrode grid covering most of the slice, especially the somatosensory .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint cor@cal region. Slices were weight down by a meshed harp and con@nuously perfused with carbogenated aCSF at a temperature of 30°C. Recordings were performed using a 256 electrode MEA system (USB-MEA 256, Mul@ Channel Systems MCS GmbH) with an electrode size of 30 µm and in-between electrodes spacing of 200 µm. Signals were sampled at 10 kHz using Mul@ Channel Experimenter sohware ( Mul@ Channel Systems MCS GmbH). Aher an equilibra@on period of 20 min, slices were recorded for 2 x 5 min and recordings were analyzed using custom scripts in MatLab (MathWorks). Parvalbumin staining Free-floa@ng organotypic slices were washed in PBS-T (0.1% Tween, Invitrogen; PBS 1x) for 10 minutes and then incubated in 0.1M glycine in PBS for 30 minutes. Slices were washed in PBS- T (3 @mes, 15 minutes each) and incubated in blocking solu@on (10% horse serum, 1% triton in PBS) for 1 hour, aher which sec@ons were incubated at 4 °C with NeuN (1:500; Millipore, ABN78) and parvalbumin (1:1000; Sigma, P3088) an@bodies for 48 hours. Following this, slices were thoroughly washed in PBS -T (10 @mes, 15 minutes each), and then incubated in the secondary an@body mix, consis@ng of AlexaFlour secondary an@bodies (1:500, ThermoFisher) and DAPI (1:500) in blocking solu@on, for 1 hour at room temperature. Slices were then washed twice more in PBS -T, and mounted on slides with Fluoromount (Invitrogen). Images were acquired using either a Leica Upright Microscope or Zeiss LSM710 laser scanning confocal microscope. RNA isola.on and cDNA genera.on For RNA isola@on, two whole hemisphere slices from the same culturing insert were collected into one tube to obtain sufficient amounts of RNA. Tissue was homogenized by addi@on of one metal bead (steel, Qiagen) and 1 ml trizole (Invitrogen) per tube and inser@ng them into a SpeedMill (Analy@k Jena) for 2 minutes. Subsequently, samples were centrifuged for 1 min at 1000 rpm and 4°C, leh on ice for 30 min, homogenized and then leh at room temperature for 5 min. 200 µl of Chloroform/Trichloromethane (Sigma) were added to each tube, vortexed for 60 seconds and leh at room temperature for 3 min. Samples were then centrifuged at 12.000 g and 4°C for 30 min and aqueous phase was pipe?ed into new tubes. 400 µl isopropanol was added to each tube, mixed and leh at room temperature for 10 minutes. Subsequently, .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint samples were centrifuged for 10 min at 12.000 g and 4°C, supernatant was discarded and 800 µl 75% ethanol were added. This step was repeated for a second @me, then tubes were leh upside down on a clean @ssue to dry at room temperature. The pellet was finally taken up in 25 µl nuclease free water each and the amount of RNA was measured on a NanoDrop. All samples showing sufficient RNA concentra@on (> 100 ng/ul) and purity (260/280 ra@o of 2.0 +/ - 0.2) were then used to generate cDNA by normalizing the sample with the lowest concentra@on to 15 µl volume and calcula@on of the amount of RNA for remaining samples accordingly. To each sample RNA, 4 µl of iScript reac@on mix (Bio-Rad), 1 µl of iScript reverse transcriptase (Bio-Rad) and nuclease free water to fill total volume to 20 µl were added. cDNA was finally generated in a Thermocycler (Biometra) by priming at 25°C for 5 min, reverse transcrip@on at 46°C for 20 min and reverse transcriptase inac@va@on step at 95° for 1 min. RT-PCR mRNA quantification was performed by real -time RT -PCR using the ΔΔCt -method. Quantitative PCR was performed in an QuantStudio TM I apparatus (Applied Biosystems) containing 1x iQ SYBR Green supermix (Bio -Rad Laboratories Inc.), 5 pm of each oligonucleotide primer ( Actb: 5ʹ -CATTGCTGACAGGATGCAGAAGG-3ʹ and 5ʹ - TGCTGGAAGGTGGACAGTGAGG-3ʹ; Syp: 5ʹ -TTCAGGACTCAACACCTCGGT-3 and 5ʹ - CACGAACCATAGGTTGCCAAC-3ʹ; Scn1a: 5ʹ -CTTGAGCCCGAAGCTTGCT-3ʹ and 5ʹ - TCCTTCTTCCACGCTGATTTG-3ʹ; Scn2a: 5ʹ -CATCGTCTTCATGATTCTGCTCA-3ʹ and 5ʹ - GGTTTTTCGCTGCTCGATGTA-3ʹ; Scn8a: 5ʹ -CATCTTTGACTTTGTGGTGGTCAT-3ʹ and 5ʹ - TGACGCGGAATAGGGTCG-3ʹ; Pvalb: 5ʹ -CTGGGGTCCATTCTGAAGGG-3ʹ and 5ʹ - TTCAACCCCAATCTTGCCGT-3ʹ; Hcn1: 5ʹ- TGAAGCTGACAGATGGCTCTT-3ʹ and 5ʹ- GAGTCGGTCAATAGCAACTGTCT-3ʹ) and 2.5-fold diluted synthesized cDNA (total volume of qPCR reaction 20 μl), by incubating 3 min at 95°C, 40 cycles of 15s at 95°C and 1 min at 60°C. Quantification was based on b-actin. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Biocy.n stainings Cells previously filled with biocy@n during Patch-Clamp recordings were washed three @mes for 20 minutes in Phosphate buffered saline (PBS-T-T) containing 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2HPO4, 1.8 mM KH 2PO4, 0.05 % Triton X-100 (Sigma) and 0.2 % Tween-20 (Sigma- Aldrich). Subsequently, slices were incubated in a staining solu@on containing 0.01 mg/ml Streptavidin-Cy3 (Sigma-Aldrich) in PBS-T-T overnight at 4°C. On the following day, slices were washed three @mes with PBS for 10 minutes, incubated in DAPI (Sigma, 1.43 µM) solu@on for 5 minutes and again washed three @mes with PBS for 10 minutes. Finally, slices were mounted on Superfrost Plus adhesion slides (Epredia) with Fluoromount G (ThermoFisher Scien@fic). Confocal imaging and morphological reconstruc.ons Neurons stained with Streptavidin -Cy3 were selected for confocal imaging if they exhibited sufficient fluorescent signal and overall morphology of a pyramidal cell. Cells were then scanned as a whole in x, y and z axis on an inverted confocal microscope (Zeiss LSM710 and Zeiss LSM980 Airyscan 2) with a 40x oil immersion objec@ve. Single stacks were imaged at a resolu@on of 1024x1024 with mul@ple @les in x and y direc@on and z -stacks at 0.66 µm thickness. Fluorophore Cy3 was imaged with a 561 nm laser and a 569-712 filter. Using the complete confocal neuron scans, 3D models of corresponding neurons were generated using the neuron tracing tool in Arivis sohware (Zeiss). Axon, dendrites and som a were segmented and obtained 3D models of neurons were further used for morphological analysis. Data analysis and sta.s.cs Patch-Clamp recording analysis Patch-Clamp data was analyzed using Fitmaster sohware (HEKA) and custom wri?en scripts for MatLab (MathWorks). Cells were included in analysis, when membrane poten@al was <= - 60 mV, they were able to generate mul@ple ac@on poten@als upon current injec@on and ac@on poten@al amplitude was larger than 40 mV. In Fitmaster, data was filtered digitally at 1 kHz to remove artefacts. Ac@on poten@als were analyzed using the corresponding sohware func@on. The membrane poten@al was measured at the start points of every protocol and a mean value .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint was calculated for every cell. Input resistance was determined by crea@ng an I -V curve of current input responses in whole-cell mode and calcula@ng the reverse slope value delta V/I. Spontaneous events in current -clamp mode were characterized as star@ng by reaching rheobase (at least one spike) and ending if membrane poten@al returned to baseline for at least 500 ms. Spontaneous events in voltage clamp mode were characterized by reaching an amplitude of at least 200 pA and ending if current returned to baseline for at least 500 ms. Spontaneous events were analyzed using the corresponding func@ons in Fitmaster sohware. Several integral values that could not be calculated in Fitmaster due to baseline shih issues were subsequently calculated with a custom -wri?en MatLab script. For all values corresponding to a single event (dura@on, integral, spikes, pos./neg. amplitude), mean values were created for every cell. If Gaussian distribu@on was assumed (membrane poten@al, input resistance, AP Half -width, v oltage sag), ROUT method outlier test was performed and significance in differences between @me periods was calculated per GraphPad Prism 10 sohware using One -Way ANOVA or mixed sta@s@cal test with post -hoc Tukey’s mul@ple comparisons test. Remaining data was assumed to be non -parametrical and therefore analyzed with Kruskal-Wallis sta@s@cal test and Dunn’s mul@ple comparison correc@on. MEA analysis MEA recordings were analyzed using custom -wri?en Matlab (MathWorks) scripts. For each slice, two recording epochs of 5 minutes and (if present) three network-driven local field poten@als (LFPs) per recording were analyzed. The data was first filtered with a low pass filter of 50 Hz and , if present, noisy channels were removed from the analysis. LFP events were detected at individually selected threshold per recording in a manual ly chosen reference channel. An individual @me window was subsequently set for the analysis of an LFP event. In a first step, we qualita@vely judged the complexity of the LFP as a) very low b) low c) medium d) complex or d) very complex. Categoriza@on criteria were as follows: one single spike (very low), 2-3 spikes (low), >3 spikes (medium), ini@al spike followed by high-frequency oscilla@on at a low amplitude not reaching the ini@al amplitude (complex), high-frequency oscilla@on at a high amplitude reaching the full ini@al amplitude (very complex) or not applicable (n. a.), when the waveform did not fall into one of the categories. Subsequently, the average of the minimum and maximum amplitudes over all channels over the dura@on of the recording was calculated. The threshold was set for posi@ve and nega@ve LFP components as a value between .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 5-20% above or below the average value. Aherwards , the electrodes that crossed th e threshold values at least once within the selected recording period were counted as ac@ve electrodes. From the number of ac@ve electrodes we further calculated 1) the area on the electrode grid covered by LFP ac@vity, 2) a ra@o of posi@ve vs. nega@ve electrodes (note that an electrode could be both posi@ve and nega@ve if the signal crossed both thresholds during the period analyzed), 3) the absolute summa@on of LFP amplitudes (cumula@ve most nega@ve and posi@ve values) and average LF P amplitude (averaged over all ac@ve electrodes) of the posi@ve and nega@ve LFP, 4) dura@on of the LFP ( averaged over all channels). Significance in differences between @me periods was calculated using Kruskal-Wallis sta@s@cal test with post- hoc Dunn’s mul@ple comparisons test per GraphPad Prism 10 sohware. Staining analysis (Parvalbumin) Per slice, two regions of interest were selected, one in the prefrontal cortex/secondary motor cortex and one in somatosensory cortex. As no difference was found between those two regions, both ROIs were counted and included in analysis. Cell coun@ng was automated and NeuN count was used as normalizer at 100% to calculate overall percentage of Parvalbumin - posi@ve cells. Mean values were generated for every slice and subsequently, for every @me period. Significance in differences between @me periods was calculated using One -Way ANOVA sta@s@cal test with post-hoc Tukey’s mul@ple comparisons test per GraphPad Prism 10 sohware. Morphological analysis Neurons were reconstructed with the neuron tracing tool in Arivis sohware (Zeiss). Dendrites were reconstructed from the soma to terminal end points and 3D models of neurons were subsequently segmented into apical and basal dendrites. Values for dendri@c length, branch points and terminal end points were ascertained i n Arivis sohware and plo?ed in GraphPad Prism 10 sohware. The axon was not included in the analysis. Significance in differences between @me periods was calculated using One -Way ANOVA sta@s@cal test with post -hoc Tukey’s mul@ple comparisons test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint

Passive and firing proper.es For this study, we used organotypic brain slice cultures (oBSCs) generated from P5-6 mice and targeted pyramidal cells of the somatosensory area of the cortex, located at a depth of 100 – 500 µm from the cor@cal surface. Pyramidal cells were iden@fied during Patch -Clamp recordings by their soma@c shape and general firing behavior, as well as post -hoc by their morphology (figure 2A). Recordings were performed con@nuously from 2 days in vitro (DIV) up to 37 days in vitro with six @me periods characterized as: ini@al (1-2 DIV), week 1 (6 -9 DIV), week 2 (11-16 DIV), week 3 (17-23 DIV), week 4 (28-30 DIV) and week 5 (34-37 DIV). Overall, 179 neurons were recorded and included in our analysis for @me periods ini@al (n=21), week 1 (n=28), week 2 (n=36), week 3 (n=45), week 4 (n=30) and week 5 (n=19) . In response to depolarizing current injec@on, we observed that neurons of the ini@al culturing period exhibited immature firing behavior and depolariza@on block a her several ac@on poten@als, thus not being able to fire con@nuous trains of ac@on poten@als when depolarized into firing satura@on (figure 2B, ini@al). Interes@ngly, this observa@on resolved already during the first week in culture with neurons gaining and retaining the ability to fire con@nuous trains of ac@on poten@als up to 5 weeks in culture (figure 2B, week 1 – week 5). While nearly two thirds of recorded neurons displayed a de polariza@on block during the ini@al culturing period, this phenomenon was only found occasionally during the later @mepoints (figure 2C). Ac@on poten@als were broad during the ini@al culturing period and the first week with an ac@on poten@al half-width of 2.62 ± 0,15 ms and 2.44 ± 0,11 ms, respec@vely and decreased con@nuously during the following weeks to reach mature steady-state values of 1,86 ± 0,10 ms aher 3 weeks (figure 2D). The instantaneous firing frequency was elevated during the ini@al culturing period with an area under the curve (AUC) value of 269,3 ± 31,04 but decreased and stabilized during culturing weeks 1 and 2 to a value of 124,5 ± 16,07. Interes@ngly, this value increased again at the later culturing periods to a final value of 528,8 ± 65,28 in week 5 (figure 2E). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Figure 2. Firing proper'es of recorded neurons. (A) Exemplary pyramidal cell, biocy9n filled during Patch -Clamp recordings and stained with Streptavidin-Cy3, 16 DIV. (B) Examples of induced spike trains from pyramidal cells of every 9me period, approximately 100 pA above rheobase. (C) Illustra9on of propor9on of neurons exhibi9ng depolariza9on block upon current injec9on un9l satura9on of number of ac9on poten9als. (D) Development of ac9on poten9al half - width over a 9me course of five weeks, **** = p<0.0001 for One-way ANOVA test. Visual comparison of ac9on poten9al width between ini9al culturing period and five weeks in vitro. (E)Development of instantaneous firing frequency over five weeks of cul9va9on, area under the curve illustrated in the graph below, * = p<0.05 for Kruskal-Wallis test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Another parameter for neuronal matura@on in organotypic cultures was found in the voltage sag, defined as the ra@o between the steady-state decrease and the largest decrease in voltage following a hyperpolarizing current, in this case - 200 pA. The sag is caused by hyperpolariza@on-ac@vated ca@on current ( Ih) and thought to be mainly controlled by ac@va@on of HCN channels 44. The voltage sag was significantly elevated during the ini@al culturing at 9,07 ± 1,04 mV, but decreased significantly and reached physiological steady-state levels at week 1 with 3,14 ± 0,47 mV (figures 3A and 3C). The cellular input resistance displayed a similar development over @me with values elevated during ini@al culturing at 338,2 ± 29,19 MW, followed by a significant decrease and stabiliza@on aher one week to 168 ± 9,91 MW (figures 3B and 3C). The res@ng membrane poten@al resided at - 73,65 ± 0,96 mV in the initial culturing phase and underwent significant hyperpolariza@on by week 2, dropping to - 81,19 ± 1,72 mV (figure 3D). All individual values are given in supplementary table 1. Overall, our data suggest ed a generally immature neuronal phenotype shortly aher culture produc@on from mice aged 5-7 postnatal days, but also con@nuing matura@on during the first three weeks in vitro to a more stabilized state with increasingly physiological firing behavior and intrinsic excitability. Figure 3. Passive proper'es and voltage sag of recorded neurons. (A) Development of voltage sag over five weeks. (B) Development of input resistance over five weeks. (C) Comparison of input resistance and voltage sag between neurons of ini9al, week 1 and week 2 culturing periods. (D) Development of res9ng membrane poten9al over a 9me course of five weeks. Significance levels for all graphs: * = p<0.05, **** = p<0.0001 for One-way ANOVA test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Development of spontaneous ac.vity in organotypic neocor.cal cultures We first wanted to observe spontaneous ac@vity in organotypic cultures on a single-cell level and therefore monitored events in current -clamp mode during whole -cell Patch -Clamp recordings over our culturing periods ini@al (n=17), week 1 (n=25), week 2 (n =30), week 3 (n=36), week 4 (n=23) and week 5 (n=17). This data set contained less cells than the previous one, as recordings of spontaneous ac@vity were not performed for every cell. Spontaneous ac@vity in current -clamp mode was characterized as an “event” when it reached ac@on poten@al threshold at least once and was considered as finished when it returned to baseline aherwards for at least 500 ms (figure 4A). We evaluated a recording @me of three minutes for every cell and calculated mean values for every parameter corresponding to single events, such as dura@on, ac@on poten@als per event and integral. We observed a striking increase in number of events and event dura@on over the cul@va@on period of five weeks, with neurons exhibi@ng nearly no spontaneous ac@vity during the ini@al culturing phase and becoming gradually more ac@ve over @me. The mean dura@on of spontaneous events increased from short bouts of ac@vity to more complex and elaborate ac@vity pa?erns which s@ll contained single spikes and short events, but also cellular up-states. This resulted in a significant increase of the mean dura@on of events between the ini@al culturing period, where events displayed a mean dura@on of 0,27 ± 0,16 s, and following @me periods, with a mean dura@on of 2,37 ± 0,43 s at week 2 and 7,97 ± 4,72 s at week 5 (figure 4B). The development of spontaneous ac@vity is best illustrated by the exemplary tra ces shown (figure 4C), showing the mostly silent state of ini@al cultures (figure 4C, ini@al) progressing to up-states exhibi@ng only few ac@on poten@als and par@al depolariza@on block (figure 4C, week 1) to elaborate spontaneous events accompanied by prolonged bursts of ac@on poten@als during later cul@va@on periods (figure 4C, weeks 2, 3 and 4). Interes@ngly, several neurons started displaying abnormal forms of excitatory discharges at week 5 in culture (figure 4C, week 5), however, it has to be stressed that these high frequency events were not the exclusive or domina@ng form of spontaneous ac@vity and were only rarely observed in cultures aged 4 -5 weeks (supplement figure 1) . This is also highlighted by the fact that spontaneous ac@vity did not seem to con@nuously accelerate into a high-frequency phenotype during culture as shown by the overall number of events per cell , with on average 0,8824 ± 0,46 events observed per 180 s in the initial culturing period and stable values of 8,967 ± 1,52 events by week 2 (figure 4D). This was also reflected by the ac@ve @me spent above res@ng .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Figure 4. Spontaneous ac'vity as recorded in current-clamp mode. (A) Defini9on of events during current-clamp recordings, event star9ng at reaching ac9on poten9al threshold and ending when returned to baseline for at least 500 ms. (B) Mean dura9on of spontaneous events. (C) Exemplary traces of spontaneous up -states recorded. (D) Number of spontaneous events recorded per cell over a recording 9me of 3 minutes. (E) Percentage of recording cell spent in ac9ve state, meaning above baseline in ac9ve event as defined in A. (F) Overall number of spontaneous ac9on poten9als recorded in one cell over a recording 9me of 3 minutes . (G) Number of ac9on poten9als per single event recorded. (H) Visual representa9on of defini9on of integral as area under the curve of a spontaneous event. Graph illustra9ng numeric evalua9on depicted on the right. Significance levels for all graphs: **** = p<0.0001, *** = p<0.001, ** = p<0.01, * = p<0.05 for Kruskal-Wallis-Test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint membrane poten@al, increasing from 0,51 ± 0,27% at ini@al culturing periods to a steady-state value of 11,81 ± 2,48 % by week 2 (figure 4E) . These ac@vity parameters only showed a significant increase when comparing the ini@al culturing phase to later culturing periods, but no con@nuous increase in-between later periods. Instead, spontaneous ac@vity shihed to more complex pa?erns composed of single spikes, bursts and long -las@ng up-states. The absolute ac@on poten@als recorded per cell significantly increased aher the ini@al culturing from 1,235 ± 0,63 to 63,7 ± 18,58 at week 2, stayed stable un@l week 4 and subsequently showed a trend towards increasing further in week 5 to 108,2 ± 64,22 (figure 4F). The change between the propor@ons of different types of spontaneous ac@vity was especially highlighted by the number of spikes per event, which saw a significant increase from ini@al values at 0,35 ± 0,18 to 4,88 ± 1,10 at week 2, but also a clear trend in late @me periods towards spontaneous events encompassing larger numbers of ac@on poten@als with a final value of 25,14 ± 20,41 at week 5 (figure 4G). This was further illustrated by calcula@ng the integral (the area under the curve) for every event and comparing mean values between individual cells. The integral value increased significantly from 1,23 ± 0,93 Vs at the ini@al culturing phase compared to 66,19 ± 20,38 Vs at week 3, but the difference ceased to be significant in week 4 and 5 despite values of 79,85 ± 50,43 Vs and 95,11 ± 75,48 Vs, respectively. This was due to the ac@vity of the neurons becoming more heterogenous, resul@ng in a rise of variance and a clustering into cells exhibi@ng predominantly single spikes and short bursts versus cells that tended to display long, elaborate up-states, some of them poten@ally drihing into either high-frequency oscilla@ons or runaway excita@on and pathological ac@vity. All individual values are given in supplementary table 1. In summary, when looking at the intracellular spontaneous ac@vity recorded at the plasma membrane of the cell, cor@cal pyramidal neurons were near silent in the ini@al phase of cul@va@on, but started to progressively develop physiological spontaneous ac@vity during the first two weeks in culture (figure 5, week 1). For later cul@va@on @me points, especially week 4 and 5, heterogeneity between individual cells as well as cultures rose, including all kinds of spontaneous ac@vity pa?erns from near-silent, single spiking and burs@ng ac@vity to complex up-states and even discharge pa?erns that could be characterized as high-frequency oscilla@ons or epilep@form ac@vity (figure 5, week 4). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint To get a be?er idea of how these pa?erns of spontaneous ac@vity develop, we next analyzed voltage-clamp recordings of whole -cell Patch -Clamp and therefore the synap@c input individual cells were receiving. Again, we included cells over the whole @me course of our culturing periods, namely ini@al (n=14), week 1 (n=17), week 2 (n=21), week 3 (n=34), week 4 (n=23) and week 5 (n=17). Spontaneous ac@vity in voltage-clamp mode was characterized as an “event” when it reached a threshold of at least 200 pA and considered as finished when it returned to baseline aherwards for at least 500 ms. Events were recorded at a holding poten@al of -70 mV and nega@ve as well as posi@ve amplitudes were calculated from the baseline (figure 6A). As expected, similar to the events recorded in current-clamp, the number of spontaneous synap@c events increased significantly from 1.0 ± 0,46 at the ini@al culturing period to a steady value of 11,1 ± 1,39 at week 2 (figure 6B), but rose again in week 5 to 16,12 ± 6,15. Also matching the trend seen in current -clamp recordings was the mean dura@on of events occurring in voltage -clamp mode, which increased significantly from 0,32 ± 0,14 s at the ini@al culturing period to 3,40 ± 0,90 s in week 1 and 8,57 ± 2,48 s in week 4 (figure 6C). Interes@ngly, we found a trend towards decreasing mean event dura@on in week 5 to 4,4 ± 1,56 s, which is in contrast to the increase observed in current -clamp recordings. Regarding the mean nega@ve amplitude of synap@c events, which reflected the excitatory input the neuron was receiving, we observed a significant increase from -267,9 ± 193,2 pA at the ini@al culturing period to -1337 ± 324,8 pA in week 1, followed by a con@nuous decrease of amplitude with the decline being significant between week 1 and week 5 with a final value of Figure 5. Spontaneous network ac'vity of organotypic cor'cal cultures a>er 1 week vs. 4 weeks . Depic9on of three minutes of current-clamp whole-cell Patch-Clamp recordings of pyramidal neurons in mouse cortex during week 1 vs. week 4 in culture. Up -states were found in both periods, however, up -states found during week 1 of culturing were lower in frequency and displayed par9al depolariza9on block, which was resolved in later culturing periods. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint -394,9 ± 101,8 pA (figure 6D). This result might seem counterintui@ve regarding the rising spontaneous ac@vity and excita@on profiles of the current -clamp recordings, however, the development of the mean nega@ve amplitude of synap@c input was well explained by the mean posi@ve amplitude of synap@c events, reflec@ng inhibitory inputs, which was found to con@nuously and significantly rise from 12,43 ± 5,53 pA at the ini@al culturing period to 179,5 ± 26,13 pA at week 3 (figure 6E), counterac@ng the excitatory input and therefore decreasing the mean nega@ve current amplitude. Interes@ngly, we observed a drop of the mean posi@ve amplitude to 91,76 ± 24,08 pA in week 5 of culturing. As the amplitudes only illustrate d the single most posi@ve and nega@ve peak points of events, we addi@onally again calculated the integral for each event and accounted for propor@ons below the baseline with nega@ve values and propor@ons above the baseline with posi@ve values to mirror inhibitory and excitatory components in the final value. Strikingly, while synap@c input was virtually non-existent in the ini@al culturing period, posi@ve and nega@ve synap@c input seemed to mostly balance each other with slightly bigger values for excitatory synap@c input un@l week 3, reflected by small nega@ve values. We observed a sharp increase of the excitatory component from -9,61 ± 7,18 pAs in week 1 to -1569 ± 747 pAs in week 4, followed by a dis@nct drop to -234 ± 106,7 pAs in week 5 (figure 6F). The increase of synap@c input in week 4 and decrease in week 5 was in accordance with the observed development of the event length, which substan@ally influences the integral value. When we took the distribu@on of the single values into account, we could again iden@fy clusters of small values as well as several cells di splaying huge integral values due to excessive excitatory ac@vity. All individual values are given in supplementary table 1. In conclusion , the development of synap@c input was coherent with the findings made in current-clamp recordings. Events gradually increased in frequency, length and complexity over the culturing period, mirroring the development of spontaneous neuronal ac@vity. The rise of inhibitory input was observed with a slight delay to excitatory input and values became more heterogeneous and clustered into low and high ac@vity profiles during culturing periods week 4 and 5. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Figure 6. Spontaneous ac'vity as recorded in voltage -clamp mode. (A) Defini9on of events during voltage -clamp recordings, event star9ng at reaching a threshold of 200 pA and ending when returned to baseline for at least 500 ms. Events were recorded at a holding poten9al of -70 mV, nega9ve and posi9ve amplitudes were calculated from the baseline. (B) Number of spontaneous events per cell as recorded over a recording 9me of three minutes. (C) Mean dura9on of spontaneous events. (D) Mean maximum nega9 ve amplitude. (E) Mean maximum posi9ve amplitude . (F) Mean integral calculated for every 9me period, calcula9ng area under the baseline as nega9ve values and areas above baseline as posi9ve values to account for excitatory/inhibitory propor9ons. Significance levels for all graphs: **** = p<0.0001, *** = p<0.001, ** = p<0.01, * = p<0.05 for Kruskal-Wallis-Test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Morphological development of neocor.cal pyramidal neurons in culture Next, we inves@gated the cellular morphology and its development over the culturing period in cells that were filled with biocy@n during Patch-Clamp recordings. Cells were stained with Streptavidin, imaged on a confocal microscope and reconstructed. Per @me period, we included five cells into the analysis (supplement figure 2). Figure 7. Development of pyramidal cell morphology. (A) Representa9ve reconstruc9ons of neurons for each culturing period, depicted at their corresponding depth as measured from the cor9cal surface. Apical dendrite is color-coded in orange, basal dendrites in blue . (B) Graph depic9ng development of total dendri9c length over the culturing periods. (C) Graph depic9ng development of the number of total branch points over the culturing periods. (D) Graph depic9ng development of the number of total dendri9c end terminals over t he culturing periods. Significance levels for all graphs: * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001 for One-Way ANOVA test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint All cells exhibited the typical morphological form of pyramidal cells (figure 7A) and were analyzed for their total dendri@c length, number of branch points and number of terminal end points. The total dendri@c length of neurons increased significantly from 1511 ± 269,6 µm at the ini@al culturing phase up to 5030 ± 143,2 µm in week 4, but significantly dropped to 3310 ± 286,4 µm between week 4 and week 5 (figure 7B). A similar development was observed for the number of total branch points, increasing from 24,4 ± 2,4 at initial culturing to 55 ± 3,90 by week 2, but dropping again to 37,8 ± 3,53 in week 5 (figure 7C) and number of terminal end points, which significantly increased aher the ini@al culturing phase from 34,2 ± 3,71 to 67,4 ± 4,01 by week 2 , but decreased again to 49,0 ± 3,65 in week 5 (figure 7D). A coherent development was observed when apical and basal dendrites were analyzed separately (supplement figure 3). All individual values are given in the supplement, table 1. In conclusion, these findings mirror the ongoing matura@on of neurons aher the ini@al culturing phase up to week 2 of culturing and stability thereaher, but also again highlight the occurring changes in week 5 of culturing. Network dynamics of neocor.cal organotypic brain slice cultures show transi.on to high-frequency oscillatory phenotype Considering our findings on the single-cell level, we were wondering how the development of spontaneous ac@vity would reflect on a network level. Therefore, to get more insights into the overall network ac@vity of our cor@cal cultured slices over @me in vitro, we performed mul@- electrode array (MEA) recordings for every @me period, namely n=17(9) for ini@al, n=17(9) for week 1, n=13(7) for week 2, n=15(8) for week 3, n=16(8) for week 4 and n=14(7) for week 5, n = number of recordings with number of slices indicated in brackets. Recordings were performed on a 16x16 (256) electrode grid and the slice was posi@oned in a way that would cover the somatosensory cor@cal area. Aherwards, recordings were analyzed using custom wri?en Matlab scripts (see Methods for Details) detec@ng every electrode that reached the individual threshold for posi@ve or nega@ve components of local field poten@als (LFPs) and calcula@ng their corresponding amplitudes (figure 8A, upper row). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint As all slices displayed robust network-driven LFP ac@vity covering all neocor@cal por@ons, we chose reference channels with the highest amplitude for every recording to analyze a Figure 8. Mul'-electrode array recordings of organotypic cultures. (A) Schema9c representa9on of MEA recording and analysis pipeline. Slices were posi9oned on 16x16 (256) electrode grid to cover the somatosensory cor9cal area. Every slice was recorded for 2x five minutes and from each recording, three LFPs were analyzed at most. During analysis, electrodes reaching threshold for either posi9ve or nega9ve component of LFP were detected and maximum amplitude was calculated. The specific LFP was analyzed from one reference channel, representa9ve for ac9vity seen in the remaining electrodes as illustrated in the overview of the recording. Onset and determina9on points of posi9ve and nega9ve components were detected and waveform of the individual LFP generated. (B) Exemplary waveforms of LFPs over a 9me course of five weeks. (C) Development of LFP complexity over five weeks with LFPs categorized into very low (single spike), low (2 -3 spikes), medium (4-5 spikes), complex (1-3 ini9al full sized spikes followed by low amplitude oscilla9on) and very complex (ini9al spike followed by full-sized amplitude oscilla9on). (D) Mean dura9on of LFPs for every recording over five weeks, ** = p<0.01 and **** = p<0.0001 for Kruskal-Wallis-Test. (E) Ra9o of electrodes reaching posi9ve vs. nega9ve threshold for every recording obtained. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint maximum of three LFPs per recording. Our analysis scripts detected onset and termina@on points for each LFP and calculated their individual amplitudes, @meframes and waveforms (figure 8A, lower row). Subsequently we analyzed which waveforms of LFPs were the most domina@ng throughout the slices at different @me-points. We found mostly simple waveforms for the ini@al culturing phase, consis@ng of merely one posi@ve and one nega@ve component, followed by maturing of the ac@vity pa?ern into more elaborate waveforms encompassing a higher number of different components throughout culturing week 1, week 2 and week 3, but shihing into mul@ple seconds long, oscilla@ng events in week 4 and week 5 (figure 8B). Though LFP waveforms were more of a spectrum rather than clearly dis@nguishable subtypes , it appeared like pa?erns of certain complexi@es were domina@ng each culturing period. Hence, to give a be?er overview on which type of LFP is domina@ng which culturing period, we classified LFP waveform complexity by spike number and pa?ern, with one spike defined as consis@ng of both posi@ve and nega@ve component. LFPs were classified into complexi@es very low (single spike), low (2 -3 spikes), medium (> 3 spikes), complex (ini@al spike(s) followed by low amplitude oscilla@on , not reaching amplitude of ini@al spike ) and very complex (ini@al spike(s) followed by high amplitude oscilla@on, reaching full amplitude of ini@al spike). While in the ini@al culturing phase, exclusively LFPs of very low complexity were observed, LFP waveforms seemed to follow a similar matura@on course as observed in the other experiments. LFPs became more elaborate up un@l week 3 of culturing, however, complex (37.5%) and very complex ( 31.25%) oscilla@ng events were taking over and dominated over half of the recordings performed by week 4 (figure 8C). Interes@ngly, while events of very low complexity only made up small propor@ons from week 2 to week 4, they suddenly reappeared as one of the domina@ng components in week 5 (28.57%) along with very complex oscilla@ng events (35.71%), again illustra @ng the increasing heterogeneity of the cultures and their clustering into groups of low/medium basal ac@vity and high ac@vity, especially during week 5. This was also reflected in the dura@on of events, which significantly increased from 1,02 ± 0,04 s at the ini@al culturing phase to 15,59 ± 4,55 s in week 3 and 26,96 ± 4,75 s in week 4, but also showed dis@nct clustering into shorter and longer dura@ons from week 3 to week 5 (figure 8D), resul@ng in a decreased mean dura@on of 15,72 ± 3,63 s in week 5 because of a cluster of slices exhibi@ng simple, short waveform LFPs as seen before in figure 8C. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint Comparison of the ra@o between electrodes reaching threshold for either posi@ve or nega@ve components of the LFP remained stable with no significant differences (figure 8E), indica@ng there were always both elements to every LFP , however, when we analyzed the quan@ta@ve data of LFP development, we found clear changes in LFP characteris@cs over @me. For the posi@ve component of the LFPs, it progressively covered significantly larger propor@ons of the electrode grid from 858,6 ± 48,77 µm2 at the ini@al phase to 1179 ± 48,65 µm2 in week 2 (figure 9A). Also, the average posi@ve LFP amplitude increased from 0,27 ± 0,45 mV to 1,1 ± 0,14 mV by week 2 (figure 9B). T he summa@on of posi@ve LFPs over the grid displayed a similar development with values increasing from 20,71 ± 3,97 mV to 105,6 ± 14,89 mV by week 2 (figure 9C). A very similar trend was observed for the nega@ve LFP component, also covering significantly more area over @me in culture with values increasing from 896,1 ± 57,51 µm2 to 1236 ± 28,51 µm2 by week 2 (figure 9D). The average nega@ve LFP amplitude increased from - 0,22 ± 0,04 mV to -0,53 ± 0,06 mV by week 4 (figure 9E) and the sum of nega@ve LFPs from - Figure 9. Evalua'on of mul'-electrode array recordings of organotypic cultures. Graphs depic9ng value s for posi9ve component of LFP , namely (A) area of the array covered, (B) average posi9ve LFP and (C) sum of posi9ve LFP s. Graphs depic9ng value for nega9ve component of LFP , namely (D) area of the array covered, (E) average posi9ve LFP and (F) sum of posi9ve LFPs. Significance levels for all graphs: **** = p<0.0001, *** = p<0.001, ** = p<0.01 for Kruskal-Wallis-Test. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 17,56 ± 3,52 mV to -51,07 ± 5,80 mV by week 4 (figure 9F). Interes@ngly, changes in average and summarized nega@ve LFP were only significant between the ini@al culturing phase and week 4, which indicates that similar to the excitat ory components in single -cell recordings, nega@ve LFP components seem ed to be present from the start of the culture with rela@ve stability and only suddenly increas ed in late culturing weeks 4 and 5, while posi@ve components gradually increase d over the whole culturing period. All individual values are given in supplementary table 2. Summarized, the MEA data confirm ed the matura@on of spontaneous network ac@vity into more elaborate pa?erns from week 1 to week 3 of organotypic neocor@cal cultures, but also the increas e of heterogeneity and variability accompanied by the appearance of high- frequency oscilla@ons in week 4 and 5. Development of the expression of Syp, Pvalb, Hcn1, Scn1a, Scn2a, and Scn8a in neocor.cal organotypic cultures To gain more insight on the cellular development of our cor@cal oBSCs, we extracted RNA from whole hemispheres cultured as organotypic brain slices. We examined expression levels of synaptophysin as general neuronal marker, as well as parvalbumin as marker for fast-spiking interneurons. Addi@onally, we analyzed expression levels of Hcn1, Scn1a, Scn2a and Scn8a to inves@gate the development of ion conductance. Tissue was collected at 2 DIV (n=7), 7 DIV (n=7), 14 DIV (n=6), 21 DIV (n=8) and 28 DIV (n=6) respec@vely. Quan@ta@ve RT-PCR revealed that synaptophysin expression levels were slightly increased aher the ini@al culturing period up un@l week 2 and showed a decline between week 2 and week 4 (figure 10A), indica@ng first signs of neuronal loss, though not significant. Parvalbumin expression increased steeply between the ini@al culturing period and week 3, but also decreased towards later culturing periods (figure 10B). Hcn1 showed a trend of increasing expression between the ini@al culturing and week 3, decreasing again at later @me points (figure 10C). Scn1a and Scn8a expression increased significantly between ini@al culturing and week 3, but dropped aherwards (figure 10D and F). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint The expression curve for Scn2a showed high variability and was inconclusive, however, it appeared like Scn2a expression peaked in week 1 of culturing, but subsequently decreased and stabilized at lower levels during week 2 to week 4 (figure 10E). Especially Parvalbumin showed a clear eleva@on in expression between the ini@al culturing phase and week 3, with a near 20 -fold increase in expression levels. We therefore decided to perform histological stainings on @ssue sec@ons of acute samples and cortex cultured for 7, 14 and 21 days respec@vely and label for NeuN and Parvalbumin. Figure 10. RT-PCR of markers Synaptophysin, Parvalbumin, Hcn1, Scn1a, Scn2a, Scn8a. Graphs depic9ng values for RT-PCR to evaluate development of RNA expression, namely (A) Synaptophysin, (B) Parvalbumin, (C) Hcn1, (D) Scn1a, (E) Scn2a and (F) Scn8a, all values for ini9al culturing period normalized to 1, expression levels of following periods all normalized in rela9on to values of ini9al phase. Significance levels for all graphs: **** = p<0.0001, *** = p<0.001, ** = p<0.01, * = p<0.05 for One-way ANOVA. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint While the general cellular and neuronal content of slices largely persisted as demonstrated by DAPI and NeuN labeling, we observed a clear upregula@on of Parvalbumin-posi@ve cells during week 1 and week 2 (figure 11A). This development was followed by a decline of expression as seen in RT-PCR, but interes@ngly, already seemed to appear during week 3 (figure 1 1B). To ensure specificity of the staining, we obtained confocal images of the stained @ssue sec@on (figure 11C). All individual values are given in supplementary table 3. Figure 11. Development of parvalbumin-posi've fast-spiking interneurons in organotypic cultures. (A) Stainings of cultures fixed at ini9al, week 1, week 2 and week 3 9me periods, namely DAPI, NeuN and Parvalbumin. (B) Graphs depic9ng development of parvalbumin-posi9ve neuronal popula9on over 9me, ** = p<0.01, *** = p<0.001, **** = p<0.0001 for One- Way ANOVA test. (C) Specificity of Parvalbumin staining as shown by confocal images of parvalbumin-posi9ve neurons, which were simultaneously posi9ve for NeuN. .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint In conclusion, RT -PCR and staining results indicated a clear upregula@on of the markers Parvalbumin, Hcn1, Scn1a and Scn8a between the ini@al culturing phase and week 3. Synaptophysin expression remained mostly stable, while Scn2a expression peaked in week 2 and declined aherwards. Nevertheless, nearly all markers dropped in expression by week 4, poten@ally indica@ng a general neuronal loss in late culturing phases week 4 and week 5.

The development of the early postnatal mouse neocortex in vitro concerning electrophysiological proper@es, morphology and RNA expression levels holds important implica@ons for planning of experiments involving mouse cor@cal oBSCs. We observed that when genera@ng these cultures from postnatal P5-6 mice, neuronal and network proper@es undergo significant changes over five weeks in culture. During the ini@al culturing phase, neurons displayed immature firing behavior characterized by a depolariza@on block and incomplete trains of ac@on poten@als. In the developing cortex, an increased input resistance along with a slow membrane @me constant allows for the summa@on of excitatory postsynap@c poten@als despite low network firing frequencies during early developmental stages of the cortex 45. During the first week in culture, neurons matured and intrinsic proper@es as well as firing behavior became increasingly physiological as demonstrated by significantly decreasing res@ng membrane poten@al, voltage sag, input resistance and AP half- width, closely mimicking the development in vivo 46,47. This matura@on was also previously described in hippocampal cultures, accompanied by an increasing expression of synaptophysin37. Intrinsic and firing parameters were probably hugely influenced by the ion conductance dynamics of the neurons, as RNA expression levels showed a con@nuous increase of Hcn1, Scn1a and Scn8a expression levels between the ini@al culturing phase and week 3 , which is also observed in vivo 48–50. This matura@on of ion conductance was furthermore reflected by the development of spontaneous ac@vity, which was very rare on single-cell level in the ini@al culturing phase and increased over @me in culture. Neurons gradually shihed from simple single spikes and short bursts to complex pa?erns of ac@vity, containing elaborate up- states. This development of spontaneous ac@vity in cultures of the somatosensory cortex along with the significant variability was described previously 39. It is also coherent with the .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint rise in Hcn1 expression levels, as H cn channels have been found to influence burs@ng behavior51. We found that spontaneous synap@c ac@vity, while s@ll non-existent in the ini@al culturing period, developed over the first 2 -3 weeks in cultures with inhibitory components apparently rising with some delay to excitatory inputs. While excita@on and inhibi@on seemed to balance each other well during the first weeks in culture, excitatory synap@c input values became more dominant during week 4 in culture, however, this likely seemed to be caused by few cells displaying heightened ac@vity and thus, manipula@ng the mean value, resul@ng in no significance in sta@s@cal tests albeit the obvious differences. These results were coherent with the phenomenon found in current-clamp events, where cells clustered into groups of medium or low ac@vity and high ac@vity, again highligh@ng the heterogeneity of organotypic neocor@cal cultures in culturing week 4 and week 5. The variability of spontaneous ac@vity on single-cell level could be a?ributed to the fact that network ac@vity has been found to be ini@ated in L5 52 with informa@on in L2/3 being more sparse, variable and informa@ve 53. Network ac@vity was present in all culturing phases, however, developed from a simple and short LFP waveform to high-frequency oscilla@ons las@ng mul@ple second by week 4 and 5. It is ques@onable if these events can be categorized as pathological and epileptogenic, or if they are physiological oscilla@ons in the high-frequency range. Considering the steep increase of Parvalbumin expression levels by 20-fold as compared between the ini@al culturing phase and week 3, equivalent to the development seen in hippocampal cultures54 and in vivo55, and the retained regularity of these events (supplement figure 4), it is likely that they were not caused by runaway excita@on. The oscilla@ons observed during MEA recordings had a controlled and rhythmic character, different from the aberrant ac@vity found in few single-cell recordings or the epilep@form phenotype described aher four weeks in hippocampal cultures11. Up-states and network oscilla@ons both rely on a delicate balance between excita@on and inhibi@on, which only developed over @me in culture, raising the possibility that the observed oscilla@ons could be physiological. It has been shown that parvalbumin-posi@ve neurons are crucial for the ini@a@on of gamma oscilla@ons 56 and phase-lock to different oscilla@ons in the hippocampus57, while their loss leads to a decrease in overall oscilla@on and especially gamma rhythm58. As high-frequency network events were terminated in a controlled manner, this indicates balanced excitatory and inhibitory forces. The matura@on of single-cell and network proper@es happened independent from sensory input and therefore ha d to be driven intrinsically. This was also observed in sensory deprived animals, where matura@on onset .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint happened with a delay of several days, but with the same @me course 59. External electrical s@mula@on influences the frequency and distribu@on of spontaneous ac@vity, as well as the propor@on of up-states26. Hence, it is likely that the depriva@on of external input in oBSCs is balanced by an increase in spontaneous network ac@vity and overall excitability as observed in visually deprived animals 60, which would explain the overall increase in excitability in culturing weeks 4 and 5 despite the network reaching a mature state by week 2. This altera@on of network proper@es to promote excitability as a response to the loss of external input is a form of homeosta@c plas@city , needed to maintain network stability and variability and depends on intact E/I balancing mechanisms61. Nevertheless, aberrant heightened ac@vity and runaway excita@on was observed in intracellular recordings, however, only in very few cells. The clustering into cellular and network phenotypes with either very basal or very high ac@vity, especially observed in week 5 of culturing, could be due to a general heterogeneity in slice viability or poten@ally even differen@al changes in transcripto mic profiles influencing homeosta@c plas@city and ac@vity, such as the transcrip@on factor family PARbZip62. A certain degree of neuronal loss definitely takes place during weeks 4 and 5, evidenced by the collec@ve drop of RNA expression levels in these culturing periods. Moreover, an indica@on of poten@al degrada@on in week 5 was also found in morphological analysis by the significant loss of neuronal complexity. However, the general dendri@c arboriza@on of pyramidal neurons in culture closely mimics the in vivo matura@on process during the first postnatal week 63 and both in vitro and in vivo results indicate that the final dendri@c configura@on is established aher this @me frame. This is highlighted by the fact that coherent morphological changes could not be found in studies inves@ga@ng mice from P10 onward46.

In this study, we provided a holis@c characteriza@on of the development of the early postnatal mouse neocortex over five weeks in vitro . In summary, the immature cortex underwent a matura@on similar as observed in vivo in terms of intrinsic and firing proper@es, network ac@vity and RNA expression levels. Cor@cal oBSCs therefore approach a mature state aher approximately two weeks in vitro . In conclusion, experiments aiming to inves@gate developmental phenomena of the cortex are best performed from 0 -14 days in vitro . In contrast, inves@ga@ons in need of an adult configura@on of the network, e.g. simula@ons of .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint diseases of the adult brain, will find the cor@cal network to be most stable and physiologically mature in weeks 2 and 3 of culturing. Beyond this @me frame, rising variability, heterogeneity, excitability and poten@al neuronal loss have to be considered, however, cor@cal cultures aged 4 or 5 weeks could poten@ally serve as models for network oscilla@ons.

This work was funded by the Chan Zuckerberg Ini@a@ve Collabora@ve Pairs Pilot Project Awards (Phase 1 & Phase 2) and by the German Research Founda@on (DFG/FNR INTER research unit FOR2715). This work was supported by the "Confocal Microscopy Facility", a core facility of the Interdisciplinary Center for Clinical Research (IZKF) Aachen within the Faculty of Medicine at RWTH Aachen University.

1. Giandomenico, S. L. et al. Cerebral organoids at the air–liquid interface generate diverse nerve tracts with func@onal output. Nat Neurosci 22, 669–679 (2019). 2. Marche?o, M. C. N., Winner, B. & Gage, F. H. Pluripotent stem cells in neurodegenera@ve and neurodevelopmental diseases. Hum Mol Genet 19, (2010). 3. Louit, A., Galbraith, T. & Berthod, F. In Vitro 3D Modeling of Neurodegenera@ve Diseases. Bioengineering vol. 10 Preprint at h?ps://doi.org/10.3390/bioengineering10010093 (2023). 4. Arber, C., Lovejoy, C. & Wray, S. Stem cell models of Alzheimer’s disease: Progress and challenges. Alzheimer’s Research and Therapy vol. 9 Preprint at h?ps://doi.org/10.1186/s13195-017-0268-4 (2017). 5. Bhaduri, A. et al. Cell stress in cor@cal organoids impairs molecular subtype specifica@on. Nature 578, 142–148 (2020). 6. Gähwiler, B. H. Organotypic cultures of neural @ssue. Trends Neurosci 11, 484–489 (1988). 7. Klostermann, O. & Wahle, P . Pa?erns of spontaneous ac@vity and morphology of interneuron types in organotypic cortex and thalamus-cortex cultures. Neuroscience 92, 1243–1259 (1999). 8. Annis, C. M., K. O’Dowd, D. & Robertson, R. T. Ac@vity-dependent regula@on of dendri@c spine density on cor@cal pyramidal neurons in organotypic slice cultures. J Neurobiol 25, 1483–1493 (1994). 9. Götz, M. & Bolz, J. Forma@on and preserva@on of cor@cal layers in slice cultures. J Neurobiol 23, 783–802 (1992). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 10. Plenz, D. & Kitai, S. T. Up and Down States in Striatal Medium Spiny Neurons Simultaneously Recorded with Spontaneous AcQvity in Fast-Spiking Interneurons Studied in Cortex-Striatum-SubstanQa Nigra Organotypic Cultures. (1997). 11. McBain, C. J., Boden, P . & Hill, R. G. Rat hippocampal slices ‘in vitro’ display spontaneous epilep@form ac@vity following long-term organotypic culture. J Neurosci

27, 35–49 (1989). 12. Stoppini, L., Buchs, P .-A. & Muller, D. A Simple Method for Organotypic Cultures of Nervous Tissue. Journal of Neuroscience Methods’ vol. 37 (1991). 13. Bara?a, J., Marienhagen, J. W. A., Ha, D., Yut, J. & Robertson, R. T. CHOLINERGIC INNERVATION OF CEREBRAL CORTEX IN ORGANOTYPIC SLICE CULTURES: SUSTAINED BASAL FOREBRAIN AND TRANSIENT STRIATAL CHOLINERGIC PROJECTIONS. Neuroscience vol. 72 (1996). 14. Cho, S., Wood, A. & Bowlby, M. R. Brain Slices as Models for NeurodegeneraQve Disease and Screening Pla[orms to IdenQfy Novel TherapeuQcs. Current Neuropharmacology vol. 5 (2007). 15. Baker, R. E., Corner, M. A. & van Pelt, J. Spontaneous neuronal discharge pa?erns in developing organotypic mega-co-cultures of neonatal rat cerebral cortex. Brain Res 1101, 29–35 (2006). 16. Dong, H. W. & Buonomano, D. V. A technique for repeated recordings in cor@cal organotypic slices. J Neurosci Methods 146, 69–75 (2005). 17. Gähwiler, B. H., Capogna, M., Debanne, D., McKinney, R. A. & Thompson, S. M. Organotypic slice cultures: A technique has come of age. Trends Neurosci 20, 471–477 (1997). 18. Mewes, A., Franke, H. & Singer, D. Organotypic Brain Slice Cultures of Adult Transgenic P301S Mice-A Model for Tauopathy Studies. PLoS One 7, (2012). 19. Croh, C. L. & Noble, W. Prepara@on of organotypic brain slice cultures for the study of Alzheimer’s disease [version 1; referees: 2 approved]. F1000Res 7, (2018). 20. Dong, W.-Q., Schurr, A., Reid, K. H., Shields, C. B. & West, C. A. The Rat Hippocampal Slice PreparaQon as an In Vitro Model of Ischemia. h?p://ahajournals.org (1988). 21. Pedersen, J. Z. et al. Hypoglycemia, Hypoxia, and Ischemia in a CorQcostriatal Slice PreparaQon: Electrophysiologic Changes and Ascorbyl Radical FormaQon. Journal of Cerebral Blood Flow and Metabolism vol. 18 (1998). 22. Laake, J. H., Haug, F.-M., Wieloch, T. & O?ersen, O. P . A Simple in Vitro Model of Ischemia Based on Hippocampal Slice Cultures and Propidium Iodide Fluorescence. Brain Research Protocols vol. 4 www.elsevier.comrlocaterbresProtocol (1999). 23. Arias, R. L., Rene, J., Tasse, P . & Bowlby´cns Bowlby´, M. R. NeuroprotecQve InteracQon Effects of NMDA and AMPA Receptor Antagonists in an in Vitro Model of Cerebral Ischemia. Brain Research vol. 816 (1999). 24. Uesaka, N., Hirai, S., Maruyama, T., Ruthazer, E. S. & Yamamoto, N. Ac@vity dependence of cor@cal axon branch forma@on: A morphological and electrophysiological study using organotypic slice cultures. Journal of Neuroscience 25, 1–9 (2005). 25. Jürgen Bolz, N. N. M. G. & T. B. FormaQon of Target-Specific Neuronal ProjecQons in Organotypic Slice Cultures from Rat Visual Cortex. (1990). 26. Goel, A. & Buonomano, D. V. Chronic electrical s@mula@on homeosta@cally decreases spontaneous ac@vity, but paradoxically increases evoked network ac@vity. J Neurophysiol 109, 1824–1836 (2013). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 27. Rutherford, L. C., Dewan, A., Lauer, H. M. & Turrigiano, G. G. Brain-Derived Neurotrophic Factor Mediates the AcQvity-Dependent RegulaQon of InhibiQon in NeocorQcal Cultures. (1997). 28. Raineteau, O., Rietschin, L., Gradwohl, G., Guillemot, F. & Gähwiler, B. H. Neurogenesis in hippocampal slice cultures. Molecular and Cellular Neuroscience 26, 241–250 (2004). 29. Muller, D., Buchs, P .-A. & Stoppini, L. Time Course of SynapQc Development in Hippocampal Organotypic Cultures. Brain Research vol. 71 (1993). 30. Noraberg, J. et al. Organotypic hippocampal slice cultures for studies of brain damage, neuroprotec@on and neurorepair. Curr Drug Targets CNS Neurol Disord 4, 435–452 (2005). 31. Schurr, A., Payne, R. S., Heine, M. F. & Rigor, B. M. Hypoxia, Excitotoxicity, and NeuroprotecQon in the Hippocampal Slice PreparaQon. Journal of Neuroscience

vol. 59 (1995). 32. Buchs, P .-A., Stoppini, L. & Muller, D. Structural ModificaQons Associated with SynapQc Development in Area CA1 of Rat Hippocampal Organotypic Cultures. Brain Research (1993). 33. Mielke, J. G. et al. Cytoskeletal, synap@c, and nuclear protein changes associated with rat interface organotypic hippocampal slice culture development. Developmental Brain Research 160, 275–286 (2005). 34. Dupont, J. L., Fourcaudot, E., Beekenkamp, H., Poulain, B. & Bossu, J. L. Synap@c organiza@on of the mouse cerebellar cortex in organotypic slice cultures. Cerebellum 5, 243–256 (2006). 35. Birgbauer, E., Rao, T. S. & Webb, M. Lysolecithin induces demyelina@on in vitro in a cerebellar slice culture system. J Neurosci Res 78, 157–166 (2004). 36. Ostergaard -B Finsen, K. et al. Organotypic Slice Cultures of the Rat Striatum: An Immunocytochemical, Histochemical and in Situ HybridizaQon Study of SomatostaQn, NeuropepQde Y, NicoQnamide Adenine DinucleoQde Phosphate-Diaphorase, and Enkephalin. Exp Brain Res vol. 103 (1995). 37. Staal, J. A., Alexander, S. R., Liu, Y ., Dickson, T. D. & Vickers, J. C. Characteriza@on of cor@cal neuronal and glial altera@ons during culture of organotypic whole brain slices from neonatal and mature mice. PLoS One 6, (2011). 38. Buonomano, D. V. Timing of Neural Responses in CorQcal Organotypic Slices. www.pnas.orgcgidoi10.1073pnas.0736909100 (2003). 39. Johnson, H. A. & Buonomano, D. V. Development and plas@city of spontaneous ac@vity and up states in cor@cal organotypic slices. Journal of Neuroscience 27, 5915–5925 (2007). 40. Caeser, M., Bonhoeffer, T. & Bolz, J. Cellular OrganizaQon and Development of Slice Cultures from Rat Visual Cortex. Exp Brain Res vol. 77 (1989). 41. Echevarrıa, D. & Albus¨department, K. AcQvity-Dependent Development of Spontaneous Bioelectric AcQvity in Organotypic Cultures of Rat Occipital Cortex. Developmental Brain Research vol. 123 www.elsevier.com/locate/bres (2000). 42. Annis, C. M., Robertson,’, R. T. & O’dowd’, D. K. Aspects of Early Postnatal Development of CorQcal Neurons That Proceed Independently of Normally Present Extrinsic Influences. 43. Oishi, Y ., Bara?a, J., Robertson, R. T. & Steward, O. Assessment of Factors RegulaQng Axon Growth between the Cortex and Spinal Cord in Organotypic Co-Cultures: Effects of Age and Neurotrophic Factors. JOURNAL OF NEUROTRAUMA vol. 21 (2004). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 44. Fan, J. et al. Reduced hyperpolariza@on-ac@vated current contributes to enhanced intrinsic excitability in cultured hippocampal neurons from PrP−/− mice. Front Cell Neurosci 10, (2016). 45. Feldmeyer, D. & Radnikow, G. Developmental altera@ons in the func@onal proper@es of excitatory neocor@cal synapses. in Journal of Physiology vol. 587 1889–1896 (2009). 46. Ciganok-Hückels, N. et al. Postnatal development of electrophysiological and morphological proper@es in layer 2/3 and layer 5 pyramidal neurons in the mouse primary visual cortex. Cerebral Cortex (2022) doi:10.1093/cercor/bhac467. 47. Etherington, S. J. & Williams, S. R. Postnatal development of intrinsic and synap@c proper@es transforms signaling in the layer 5 excitatory neural network of the visual cortex. Journal of Neuroscience 31, 9526–9537 (2011). 48. Allen Ins@tute for Brain Science (2024). Allen Developing Mouse Brain Atlas [SCN1A, ID 20265]. Available from developing mouse.brain-map.org. h?ps://developingmouse.brain-map.org/gene/show/20028. 49. Allen Ins@tute for Brain Science (2024). Allen Developing Mouse Brain Atlas [HCN1, ID 15165]. Available from developing mouse.brain-map.org. h?ps://developingmouse.brain-map.org/gene/show/14941. 50. Liang, L. et al. Developmental dynamics of voltage-gated sodium channel isoform expression in the human and mouse brain. Genome Med 13, 1–14 (2021). 51. Rátkai, A. et al. Homeosta@c plas@city and burst ac@vity are mediated by hyperpolariza@on-ac@vated ca@on currents and T-type calcium channels in neuronal cultures. Sci Rep 11, (2021). 52. Sanchez-Vives, M. V. & McCormick, D. A. Cellular and network mechanisms of rhytmic recurrent ac@vity in neocortex. Nat Neurosci 3, 1027–1034 (2000). 53. Sakata, S. & Harris, K. D. Laminar Structure of Spontaneous and Sensory-Evoked Popula@on Ac@vity in Auditory Cortex. Neuron 64, 404–418 (2009). 54. Bergold, P . J. & Casaccia-Bonnefil, P . PreparaQon of Organotypic Hippocampal Slice Cultures Using the Membrane Filter Method. (1997). 55. Allen Ins@tute for Brain Science (2024). Allen Developing Mouse Brain Atlas [PVALB, ID 19293]. Available from developing mouse.brain-map.org. h?ps://developingmouse.brain-map.org/gene/show/19056. 56. Sohal, V. S., Zhang, F., Yizhar, O. & Deisseroth, K. Parvalbumin neurons and gamma rhythms enhance cor@cal circuit performance. Nature 459, 698–702 (2009). 57. Klausberger, T. et al. Brain-state-and cell-type-specific firing of hippocampal interneurons in vivo. (2003). 58. Ruthazer, E. S. et al. Mice With Decreased Number of Interneurons Exhibit Aberrant Spontaneous and Oscillatory Ac@vity in the Cortex. Front. Neural Circuits 12, 96 (2018). 59. Rochefort, N. L. et al. SparsificaQon of Neuronal AcQvity in the Visual Cortex at Eye- Opening. www.pnas.org/cgi/content/full/. 60. Brown, A. P . Y ., Cossell, L. & Margrie, T. W. Visual Experience Regulates the Intrinsic Excitability of Visual Cor@cal Neurons to Maintain Sensory Func@on. Cell Rep 27, 685- 689.e4 (2019). 61. Turrigiano, G. G. & Nelson, S. B. Homeosta@c plas@city in the developing nervous system. Nature Reviews Neuroscience vol. 5 97–107 Preprint at h?ps://doi.org/10.1038/nrn1327 (2004). 62. Valakh, V. et al. A transcrip@onal constraint mechanism limits the homeosta@c response to ac@vity depriva@on in mammalian neocortex Abbreviated Title: HLF and TEF regulate homeosta@c plas@city. Elife 12, (2023). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint 63. Kroon, T., van Hugte, E., van Linge, L., Mansvelder, H. D. & Meredith, R. M. Early postnatal development of pyramidal neurons across layers of the mouse medial prefrontal cortex. Sci Rep 9, (2019). .CC-BY-NC 4.0 International licensemade available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprintthis version posted April 5, 2024. ; https://doi.org/10.1101/2024.04.05.588217doi: bioRxiv preprint