Neural temporal code-driven stimulation in real-time using the Victor–Purpura distance

• DOI: 10.1007/s12530-025-09670-4
• Publication Date: 2025-03-30T04:52:39Z
• Authors: Alberto Ayala; Angel Lareo; Pablo Varona; Francisco B. Rodr...

Abstract Numerous neural systems encode information through the generation of specific sequences of action potentials, creating temporal activity patterns that form neural codes. These systems produce functionally equivalent activity patterns, albeit with some variability in their temporal structure. In this work, we have implemented a real-time closed-loop stimulation protocol to study temporal coding in neural systems. The protocol identifies codes in neural signals acquired in real-time and delivers a stimulus when a predefined code is identified allowing a degree of variability in the detection through the use of the Victor–Purpura metric. The goal of the stimulation is to induce a new state in the system and to study the equivalence between codes with intrinsic variability. The real-time performance of the protocol was validated in closed-loop experiments with an electronic neuron by characterizing the latencies. Moreover, its functionality was corroborated through two proof of concept scenarios where we modulated the activity of a neural model to induce a new dynamic state. In the first scenario, when tested in a regular bursting model state with Gaussian stochastic inputs to induce temporal variability, our protocol robustly generated short bursts amidst consecutive bursts produced by the model without stimulation. In the second scenario, tested in an operating chaotic model state, the protocol drove the model to show regular bursting activity. The findings indicated that the dynamic state was consistently induced by closed-loop stimulation in contrast to the less effective open-loop stimulation, i.e., without precise activity-dependent stimulation of the system. The reproducibility of these results in both test scenarios is supported by the statistical analysis carried out. This protocol allows inferring of equivalence between different matching patterns when closed-loop stimulation, driven by these patterns, elicits comparable responses. This new real-time protocol implementation is available as open-source software.