Neuronal Avalanches in the Neocortex

National Institute Of Mental Health

Investigators

Linked publications, trials & patents

Abstract

1. Using holographic (laser-guided) activation to stimulate single nerve cells in the brain, we discovered that even a few action potentials generated independently of ongoing brain activity are communicated highly efficiently within the network. This remarkable responsiveness to novel, unexpected stimuli supports the theory that the brain operates in a 'critical state,' where activity manifests as neuronal avalanches. Our findings reveal an extraordinary sensitivity of the brain in processing and communicating new information and provide a theoretical framework that explains this heightened sensitivity. Critical Scaling of Novelty Critical Scaling of Novelty in the Cortex Tiago L. Ribeiro, Ali Vakili, Bridgette Gifford, Raiyyan Siddiqui, Vincent Sinfuego, Sinisa Pajevic, & Dietmar Plenz doi: https://doi.org/10.1101/2024.12.23.630084 The ability to detect and transmit novel events is essential for adaptive behavior in uncertain environments. Here, we investigate how holographically triggered, unanticipated action potentials propagate through the primary visual cortex of resting mice, focusing on pyramidal neuron communication. We find that these novel spikes — uncorrelated with ongoing activity — exert a disproportionately large influence on neighboring neurons, whose response scales as a power law (exponent ~0.2–0.3). Even a few such spikes can recruit a large fraction of the local network, enabling robust decoding of perturbation origin despite high trial-by-trial variability and ongoing activity dominated by large activity fluctuations in the form of scale-invariant, parabolic neuronal avalanches. Simulations confirm this scaling to small, local perturbations aligns with the high susceptibility of complex systems near criticality. These results suggest that critical dynamics facilitate efficient transmission of novel signals, revealing a fundamental mechanism for cortical novelty detection. 2. Recent advancements in Artificial Intelligence have highlighted the significant benefits of designing computational processes based on neuronal network principles. In our recent study, we show how incorporating brain-specific mechanisms—such as balanced inhibition and local synaptic plasticity—can enhance the performance of reservoir computers. Our research offers profound mechanistic insights into improving reservoir computing, which is recognized for its universal learning capabilities. Boosting Reservoir Computing with Brain-inspired Adaptive Dynamics Keshav Srinivasan, Dietmar Plenz, Michelle Girvan arXiv:2504.12480 Abstract: Reservoir computers (RCs) provide a computationally efficient alternative to deep learning while also offering a framework for incorporating brain-inspired computational principles. By using an internal neural network with random, fixed connections - the 'reservoir'- and training only the output weights, RCs simplify the training process but remain sensitive to the choice of hyperparameters that govern activation functions and network architecture. Moreover, typical RC implementations overlook a critical aspect of neuronal dynamics: the balance between excitatory and inhibitory (E-I) signals, which is essential for robust brain function. We show that RCs characteristically perform best in balanced or slightly over-inhibited regimes, outperforming excitation-dominated ones. To reduce the need for precise hyperparameter tuning, we introduce a self-adapting mechanism that locally adjusts E/I balance to achieve target neuronal firing rates, improving performance by up to 130% in tasks like memory capacity and time series prediction compared with globally tuned RCs. Incorporating brain-inspired heterogeneity in target neuronal firing rates further reduces the need for fine-tuning hyperparameters and enables RCs to excel across linear and non-linear tasks. These results support a shift from static optimization to dynamic adaptation in reservoir design, demonstrating how brain-inspired mechanisms improve RC performance and robustness while deepening our understanding of neural computation.