Encoding Tactile Stimuli for Braille Recognition with Organoids

TL;DR

This work tackles open-loop tactile Braille recognition using neural organoids as energy-efficient bio-hybrid processors. It introduces a transferable spatio-temporal-amplitude encoding that maps event-based tactile sensor data (from an Evetac sensor) into electrical stimulation for organoids on a low-density MEA, enabling classification with an SVM. Key findings show $61\%$ accuracy with a single organoid and $83\%$ with a three-organoid ensemble, with enhanced robustness to noise in the multi-organoid configuration, illustrating the potential of organoids for scalable, low-power computation. The study provides a concrete encoding framework and demonstrates the feasibility of bio-hybrid computation for tactile processing, offering a foundation for scalable organoid-based sensing and control architectures.

Abstract

This study proposes a transferable encoding strategy that maps tactile sensor data to electrical stimulation patterns, enabling neural organoids to perform an open-loop artificial tactile Braille classification task. Human forebrain organoids cultured on a low-density microelectrode array (MEA) are systematically stimulated to characterize the relationship between electrical stimulation parameters (number of pulse, phase amplitude, phase duration, and trigger delay) and organoid responses, measured as spike activity and spatial displacement of the center of activity. Implemented on event-based tactile inputs recorded from the Evetac sensor, our system achieved an average Braille letter classification accuracy of 61% with a single organoid, which increased significantly to 83% when responses from a three-organoid ensemble were combined. Additionally, the multi-organoid configuration demonstrated enhanced robustness against various types of artificially introduced noise. This research demonstrates the potential of organoids as low-power, adaptive bio-hybrid computational elements and provides a foundational encoding framework for future scalable bio-hybrid computing architectures.

Figures (8)

• Figure 1: Interaction between a forebrain organoid (FO) cultured on a multi-electrode array (MEA) and the Neuroplatform. The 8 recording electrodes are arranged with 200 $\mu$m centre-to-centre spacing, providing a distributed interface across the organoid. Neural activity is recorded from the FO and electrical stimulation is delivered via a remotely accessible software interface.
• Figure 2: Illustration of adjustable parameters in the Neuroplatform.
• Figure 3: Experimental procedure for collecting Braille using a tactile sensor, encoding it into electrical stimulation, delivering it to organoids, and classifying 26 Braille letters based on organoid outputs. The first row outlines the experimental steps, and the second row shows the type of data associated with each step.
• Figure 4: Illustration of the encoding strategy. (a) Schematic showing the Braille character (red dots) sliding across the marker-embedded Evetac surface along the direction of the red arrow. (b) The output of the neuromorphic tactile sensor in the form of a heatmap, where the colorbar indicates the number of events and brighter colors represent higher event counts. (c) The corresponding electrical stimulation encoding for different regions of the original data. The table below shows an example of the encoding results from sensor outputs in Area 3 and 7 to electrical stimulation parameters.
• Figure 5: Relationship between stimulation parameter values and the total spike count across all channels of the organoid. The x-axis represents the parameter value, and the y-axis indicates the average spike count across all channels. Lines represent the index of the stimulated electrode.