Poster Z13 - Neuronal regulation of brain organoids via electrical stimulation for induced learning

Introduction: Brain organoids, three-dimensional self-aggregates of neural cells, can recapitulate some of the functions and developmental characteristics of the human brain in vitro. However, the advancement of replicating cognition or learning on these organoid models remains challenging due to a lack of 3D microelectrode array (MEA) interfaces, establishing appropriate learning protocols, and supportive glial cells showcasing spontaneous electrophysiological activity. We previously reported 3D shell MEAs designed to match the unique architecture of brain organoids.[1] Unlike 2D MEAs, mapping the three-dimensional functionality of a human brain or brain organoid has yet to be realized. Here, we present a self-folded MEA capable of interfacing with submillimeter-sized brain organoids capable of both stimulation and recording for open-loop and closed-loop learning experiments over weeks. In open-loop learning, unidirectional instructional stimuli without feedback can be used to monitor brain organoid activity, while closed-loop learning incorporates an adaptive framework based on continuous feedback. We have identified electrical sensory inputs that elicit a change in firing patterns, indicating their ability for stimulation-triggered learning. A parametric study showed an increase in firing rate in the instance following the stimulation. The capability of this physical embodiment of a recurrent neural network to respond to an electrical sensory input is being explored for reservoir-based open-loop learning by using modulated inputs at various time series. This study is an important step in advancing the previously described concept of “Organoid Intelligence” (OI),[2] which seeks to develop neuromorphic computational devices using brain cells.

Materials and

Methods: We fabricated our four-petal MEA design under traditional microfabrication practices with oxide-coated silicon wafers. The design was composed of a bilayer of photocrosslinkable epoxy SU8 with metal (gold, Au) wiring, and the electrodes were coated with a high electrode conductive polymer matrix Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: PSS). This careful construction ensured the reliability and functionality of our MEA design. We next developed a Brain Organoid Microchip (BOM) packaging for greater versatility and ease of functionality. We differentiated brain organoids from the induced Pluripotent Stem Cell (iPSC) NIBSC-8 cell lines from 400-600 μm in size, ensuring the consistency and quality of our experimental samples.

Results, Conclusions, and Discussions: With a standard seven-layer CAD mask design, we’ve developed a novel polymer matrix with epoxy SU8 resin with soft gold electrodes (Fig. 1A). The inherent biocompatibility and amorphous textile structure of SU8 resin offer an unpreceded foundation for longer-term electrophysiology. In a controlled isolated environment, we can simultaneously measure and stimulate a single brain organoid at varying electrode points (Fig. 1B). Using a spike-storing algorithm, we could detect single units (Fig. 1C) at high temporal resolution. Via external stimulation, we performed an extensive current and frequency sweep a found high induced activity from 20 – 30 µA (Fig. 1D). Stimulation and recording data from the proposed platform could inform human brain development in utero and understand how to design neuromorphic and non-von Neumann computational architecture.

In summary, we have successfully monitored the electrical activity of the brain
organoids inside shell electrodes for weeks. Importantly, we observed that electrical stimulation shows evoked spiking, a prerequisite for demonstrating reinforcement learning, neuronal plasticity, and memory. In the future, we seek to create closed-loop learning paradigms, investigate 3D spatiotemporal patterns, and incorporate neurochemical stimulation.

References: [1] Huang et al., Sci Adv (2022). [2] Smirnova et al., Frontiers in Science (2023).

Acknowledgements (Optional): Acknowledgments: We acknowledge funding from the JHU Surpass Program