From Neural Activity to Computation: Biological Reservoirs for Pattern Recognition in Digit Classification

TL;DR

The paper investigates a biologically grounded reservoir computing framework (BRC) in which a cultured neural network on a high-density MEA acts as the reservoir to transform spatial input patterns into rich, high-dimensional representations. Inputs are encoded as spatial electrical stimulation patterns, and evoked spiking responses across the array are converted into feature vectors that feed a simple linear readout for digit-pattern classification, with a synthetic artificial reservoir used as a benchmarking reference. The biological reservoir achieves performance on the order of ~75% accuracy on a 10-class digit task, comparable to the artificial reservoir, but exhibits notable cross-day drift in encoding due to intrinsic network dynamics and plasticity. This work demonstrates the viability of biologically instantiated reservoirs for pattern recognition, highlights potential energy efficiency and interpretability advantages, and suggests future work to generalize across tasks and account for temporal evolution in living neural substrates.

Abstract

In this paper, we present a biologically grounded approach to reservoir computing (RC), in which a network of cultured biological neurons serves as the reservoir substrate. This system, referred to as biological reservoir computing (BRC), replaces artificial recurrent units with the spontaneous and evoked activity of living neurons. A multi-electrode array (MEA) enables simultaneous stimulation and readout across multiple sites: inputs are delivered through a subset of electrodes, while the remaining ones capture the resulting neural responses, mapping input patterns into a high-dimensional biological feature space. We evaluate the system through a case study on digit classification using a custom dataset. Input images are encoded and delivered to the biological reservoir via electrical stimulation, and the corresponding neural activity is used to train a simple linear classifier. To contextualize the performance of the biological system, we also include a comparison with a standard artificial reservoir trained on the same task. The results indicate that the biological reservoir can effectively support classification, highlighting its potential as a viable and interpretable computational substrate. We believe this work contributes to the broader effort of integrating biological principles into machine learning and aligns with the goals of human-inspired vision by exploring how living neural systems can inform the design of efficient and biologically plausible models.

Figures (7)

• Figure 1: Framework of our Biological Reservoir Computing (BRC) paradigm. In this approach, a multi-electrode array (MEA) functions as a bidirectional interface, enabling both the stimulation of and the recording from a cultured biological neural network. Discrete inputs are encoded by selectively activating specific subsets of MEA electrodes, which deliver targeted stimuli to the network. The evoked spiking responses are captured via a separate set of electrodes and transformed into high-dimensional vectors that encode the input within a latent computational space. Due to the inherent complexity and rich dynamics of the biological reservoir, this transformation is intrinsically nonlinear. Subsequently, a linear classifier is trained to infer the category of the original input from its corresponding latent representation.
• Figure 2: Illustration of a cultured biological network plated on a MEA device. Each square represents a single MEA electrode. Input patterns are mapped onto the MEA by assigning elements of the patterns to specific electrodes. Electrical pulses are delivered based on the corresponding input intensities, and the evoked activity of the network is recorded via the remaining electrodes. The resulting spiking responses are used to construct a high-dimensional representation of the input in the feature space of the biological reservoir. Neuron scale in the figure is not to scale and is adjusted for visibility.
• Figure 3: Visual representation of the input patterns. Digit recognition: input patterns represent the digits from 0 to 9.
• Figure 4: Heatmaps of neural activity following stimulation: The left panel shows the recorded response within a $10\,ms$ time window following stimulation with the input pattern "0", while the right panel depicts the corresponding response for the input pattern "1". The heatmaps illustrate the spatial distribution of spiking activity across the MEA electrodes.
• Figure 5: Accuracy variation across different readout windows. For each stimulation session (n=9), classification accuracy was computed using neural responses extracted from different post-stimulus time windows, ranging from $5$ to $50\,ms$. The plot reports the mean accuracy $\pm$ standard error of the mean (SEM). This analysis highlights how the temporal integration window influences the effectiveness of the reservoir readout.