Complexity is a hallmark of the human brain, emerging from the interaction of billions of neurons over time. Among various measures, the Hurst exponent ($H$) has been used to characterize temporal correlations in neuronal activity. Here, we introduce a method to estimate $H$ directly from binary spike train series and propose that the coefficient of variation (CV) of $H$, rather than $H$ itself, more accurately reflects neural complexity during organoid maturation. After validating the method on simulated data with embedded short- and long-range correlations, we applied it to cortical organoids derived from human-induced pluripotent stem cells, recorded via multielectrode arrays over 250 day. The analysis revealed a peak in complexity, as indicated by the CV of $H$, around the 100th day. At this stage, the organoids exhibited a wide repertoire of dynamic patterns, including both isolated spiking and highly synchronized network events. These results suggest that organoid activity undergoes a structured evolution in complexity and that CV of $H$ provides a robust metric to capture this transition. Our approach offers a scalable tool for quantifying neural complexity in the developing brain.

Close

PrevNext

View figure in article

• Cellular organization, physiology & dynamics
• Fluctuations & noise
• Neuroscience, neural computation & artificial intelligence
• Brownian dynamics
• Information theory
• Stochastic analysis methods