Modeling Noise and Artifacts in Brain Organoids Spike Measurements for Robust 2-D MEA Analysis

Understanding brain function and dysfunction requires accurate measurement of neuronal activity. Microelectrode arrays (MEAs) provide a mature technology for capturing high-resolution recordings of spikes and network bursts (BRs) in neuronal cultures. However, when translating to brain organoids (BOs), i.e., 3-D self-organized structures containing many cell types and cyto-architectures typical of the human brain, at early development stage, standard MEA’s recordings currently lack three-dimensionality needed to properly analyze BO activity, limiting the analysis to the edge of the organoids. These lead to an amplitude of the extracellular recordings up to 100–1000 times smaller than intracellular measurements (range 10 µV–200 µV) and A signal-to-noise ratio (SNR) of less than 1 dB. The presence of noncontacting electrodes and a synaptic density and connectivity still under development amplifies the presence of thermal fluctuations, low-frequency drift, commonmode noise, and ghost spikes, which can seriously hinder reliable spike detection and BR analysis. In this work, we introduce a benchmark 2-D MEA platform that systematically examines the effect of different sources of noise using BO extracellular recordings. Our findings show that noise and artifacts can have highly variable effects, underscoring key factors that must be considered in both methodological development and experimental design and offering practical guidance for improving MEA-based investigations reliability in 3-D neuronal architectures.