It is demonstrated that neuromorphic-triggered optogenetic inhibition significantly alters ripple dynamics and reduces oscillatory energy, establishing a practical and accessible neuromorphic framework for low-latency closed-loop control of fast brain dynamics in vivo.
Abstract
Real-time closed-loop neuromodulation, in which stimulation is precisely timed to ongoing brain dynamics, holds transformative potential for treating neurological disorders and probing neural circuit function. However, it requires low-latency, energy-efficient processing of high-bandwidth neural signals that conventional computing architectures struggle to deliver. Neuromorphic computing, which emulates the event-driven and massively parallel operation of biological neural circuits, offers a compelling alternative. Yet, its integration into closed-loop frameworks validated in vivo for fast, transient oscillations has not been demonstrated. Here, we present a fully integrated neuromorphic framework for real-time detection and manipulation of hippocampal ripples: brief (30-100 ms), high-frequency (100-250 Hz) oscillations that are critical for memory consolidation and implicated in neurological disorders. We train compact spiking neural networks comprising 41 neurons and 530 parameters using surrogate-gradient backpropagation, achieving detection performance competitive with deep learning models across 23 recording sessions while consuming up to 200-fold less energy when deployed on SpiNNaker neuromorphic hardware. Integration with the open-source Open Ephys platform yields total closed-loop latencies of approximately 50 ms, enabling intra-event stimulation in up to 80% of ripples. Validating the complete sensing-processing-stimulation pipeline in awake, head-fixed mice, we demonstrate that neuromorphic-triggered optogenetic inhibition significantly alters ripple dynamics and reduces oscillatory energy. This work establishes a practical and accessible neuromorphic framework for low-latency closed-loop control of fast brain dynamics in vivo.
A high-performance hardware implementation of large-scale neuromorphic system to investigate anesthetic-induced neural dynamics and paves the way for advanced brain–machine interfaces and closed-loop anesthetic delivery systems.
Chuan-Guang Wang, Xiaotian Pan, Si Chen et al.· Frontiers in Systems Neurosc...· 0 citations
This work introduces their symmetric counterpart by replacing adaptation with slow self-excitation, motivated by intrinsic calcium-mediated membrane currents, and derives and validate a mean-field neural mass model that remains stable while retaining working-memory functionality.
D. Depannemaecker, Adrien D’hollande, G. Casagrande et al.· Nature Communications· 0 citations
BACKGROUND
High-frequency deep-brain stimulation can reduce seizure burden but imposes a large stimulation load on the tissue, and its effects depend on both frequency and temporal patterning. We tested whether a compact reservoir model could generate a hippocampal-inspired event schedule and whether that schedule could modulate neuronal activity in vitro with fewer nominal pulses than 50-Hz stimulation.
METHODS
A 10-unit leaky reservoir with a one-dimensional affine ridge readout transformed cortical input recordings into CA3-inspired outputs. Retrospective computational validation used paired cortical-CA3 recordings, training-only scaling, removal of at least the first 10 s of every generated trace, date-grouped held-out evaluation, multiple spectral and temporal metrics, objective-function ablation and surrogate signals. One offline biomimetic schedule was then delivered open loop to primary rat cortical cultures on multielectrode arrays and compared with periodic 0.16-Hz and 50-Hz stimulation.
RESULTS
17 of 19 paired recordings passed the pre-specified transient diagnostic. In 11 scorable date-grouped held-out recordings, the reservoir had modestly lower median log-PSD RMSE and autocorrelation error than linear and lagged-ridge baselines. When applied to primary cortical cultures, the primary pooled analysis of all 11 biomimetic-stimulated MEAs yielded a mean firing-rate slope of 0.7655 relative to the null value of 1 (95% CI for the mean slope, 0.491-1.040; p = 0.0865). In separate MEA-level fold-change analyses of all 11 biomimetic-stimulated MEAs, normalized firing rate was 0.8458 and normalized burst rate was 0.7792, both significantly reduced relative to baseline. A post hoc exploratory classification identified reduced firing-rate slopes in 4 of 11 MEAs.
CONCLUSION
These results provide an open-loop in vitro proof of concept that a low-event, irregular stimulation schedule can be associated with heterogeneous modulation of cortical-network activity. Periodic stimulation at the same nominal mean event rate did not produce a statistically significant modulation under the conditions tested. The irregular schedule may contain timing-related properties that contribute to the observed response, but the specific contribution of ESN-derived temporal ordering remains unresolved. Future work should test appropriate temporal controls and evaluate the approach in closed-loop and epilepsy-model experiments.
Ángel Canal-Alonso, Adam Armada-Moreira, Alessio Di Clemente et al.· Brain Stimulation· 0 citations
Controlling specific neuronal dynamics with electrical stimulation is critical for therapeutic neuromodulation, yet deriving optimal control policies remains challenging due to the complex and non-stationary nature of biological neuronal networks. While reinforcement learning (RL) offers a powerful closed-loop control framework, its reliance on prolonged stimulus-driven exploration is difficult to reconcile with the physiological limits of living tissue. Here, we demonstrate an in silico-to-in vitro transfer strategy that achieves efficient state-dependent control of network bursting in cultured neurons. The transferred policy outperforms heuristic controls, while maintaining constrained stimulation usage. Concurrent calcium imaging reveals the mechanistic basis of the learned policy: the agent optimizes stimulation spatially and temporally, exploiting local network topology and intrinsic physiological temporal dynamics. These results establish in vitro brain-on-chip cultures as a tractable stepping stone for RL-based neuromodulation and demonstrate that effective control policies can be derived in biophysically calibrated digital twins and transferred directly to living networks.
Eduardo Carvalho, José C. Mateus, Ricardo Pinto et al.· bioRxiv· 0 citations
It is aimed at proving that SNNs have potential in such areas as computer vision, robotics, and speech recognition, and their role in overcoming the barrier between artificial and biological neural systems is proved.
Mesala Sravani, K. Kumari, S. M. Reddy· International Journal of Unc...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
