Predictive Routing Emerges from Self-supervised Stochastic Neural Plasticity

Overview

Journal bioRxiv

Date 2025 Jan 13

PMID 39803482

Authors

Hamed Nejat

Jason Sherfey

Andre M Bastos

Affiliations

Soon will be listed here.

Abstract

Neurophysiology studies propose that predictive coding is implemented via alpha/beta (8-30 Hz) rhythms preparing specific pathways to process predicted inputs. This leads to a state of relative inhibition, reducing feedforward gamma (40-90 Hz) rhythms and spiking for predictable inputs. This is called predictive routing model. It is unclear which circuit mechanisms implement this push-pull interaction between alpha/beta and gamma rhythms. To explore how predictive routing is implemented, we developed a self-supervised learning algorithm we call generalized Stochastic Delta Rule (gSDR). Development of this algorithm was necessary because manual tuning of parameters (frequently used in computational modeling) is inefficient to search through a non-linear parameter space that leads to emergence of neuronal rhythms. We used gSDR to train biophysical neural circuits and validated the algorithm on simple objectives. Then we applied gSDR to model observed neurophysiology. We asked the model to reproduce a shift from baseline oscillatory dynamics (~<20Hz) to stimulus induced gamma (~40-90Hz) dynamics recorded in the macaque visual cortex. This gamma oscillation during stimulation emerged by self-modulation of synaptic weights via gSDR. We further showed that the gamma-beta push-pull interactions implied by predictive routing could emerge via stochastic modulation of the local circuitry as well as top-down modulatory inputs to a network. To summarize, gSDR succeeded in training biophysical neural circuits to satisfy a series of neuronal objectives. This revealed the inhibitory neuron mechanisms underlying the gamma-beta push-pull dynamics that are observed during predictive processing tasks in systems and cognitive neuroscience.

References

Zhao D, Zeng Y, Li Y . BackEISNN: A deep spiking neural network with adaptive self-feedback and balanced excitatory-inhibitory neurons. Neural Netw. 2022; 154:68-77. DOI: 10.1016/j.neunet.2022.06.036. View

Rimehaug A, Stasik A, Hagen E, Billeh Y, Siegle J, Dai K . Uncovering circuit mechanisms of current sinks and sources with biophysical simulations of primary visual cortex. Elife. 2023; 12. PMC: 10393295. DOI: 10.7554/eLife.87169. View

Frazier-Logue N, Hanson S . The Stochastic Delta Rule: Faster and More Accurate Deep Learning Through Adaptive Weight Noise. Neural Comput. 2020; 32(5):1018-1032. DOI: 10.1162/neco_a_01276. View

Chien V, Maess B, Knosche T . A generic deviance detection principle for cortical On/Off responses, omission response, and mismatch negativity. Biol Cybern. 2019; 113(5-6):475-494. PMC: 6848254. DOI: 10.1007/s00422-019-00804-x. View

Adesnik H, Bruns W, Taniguchi H, Huang Z, Scanziani M . A neural circuit for spatial summation in visual cortex. Nature. 2012; 490(7419):226-31. PMC: 3621107. DOI: 10.1038/nature11526. View

Kirihara K, Tada M, Koshiyama D, Fujioka M, Usui K, Araki T . A Predictive Coding Perspective on Mismatch Negativity Impairment in Schizophrenia. Front Psychiatry. 2020; 11:660. PMC: 7360815. DOI: 10.3389/fpsyt.2020.00660. View

Sherfey J, Ardid S, Hass J, Hasselmo M, Kopell N . Flexible resonance in prefrontal networks with strong feedback inhibition. PLoS Comput Biol. 2018; 14(8):e1006357. PMC: 6103521. DOI: 10.1371/journal.pcbi.1006357. View

Borgers C, Kopell N . Synchronization in networks of excitatory and inhibitory neurons with sparse, random connectivity. Neural Comput. 2003; 15(3):509-38. DOI: 10.1162/089976603321192059. View

Haegens S, Barczak A, Musacchia G, Lipton M, Mehta A, Lakatos P . Laminar Profile and Physiology of the α Rhythm in Primary Visual, Auditory, and Somatosensory Regions of Neocortex. J Neurosci. 2015; 35(42):14341-52. PMC: 4683691. DOI: 10.1523/JNEUROSCI.0600-15.2015. View

10.

Kass R, Ventura V, Brown E . Statistical issues in the analysis of neuronal data. J Neurophysiol. 2005; 94(1):8-25. DOI: 10.1152/jn.00648.2004. View

11.

van Kerkoerle T, Self M, Dagnino B, Gariel-Mathis M, Poort J, van der Togt C . Alpha and gamma oscillations characterize feedback and feedforward processing in monkey visual cortex. Proc Natl Acad Sci U S A. 2014; 111(40):14332-41. PMC: 4210002. DOI: 10.1073/pnas.1402773111. View

12.

Roopun A, Kramer M, Carracedo L, Kaiser M, Davies C, Traub R . Period concatenation underlies interactions between gamma and beta rhythms in neocortex. Front Cell Neurosci. 2008; 2:1. PMC: 2525927. DOI: 10.3389/neuro.03.001.2008. View

13.

Dallerac G, Li X, Lecouflet P, Morisot N, Sacchi S, Asselot R . Dopaminergic neuromodulation of prefrontal cortex activity requires the NMDA receptor coagonist d-serine. Proc Natl Acad Sci U S A. 2021; 118(23). PMC: 8201892. DOI: 10.1073/pnas.2023750118. View

14.

Lee J, Whittington M, Kopell N . Top-down beta rhythms support selective attention via interlaminar interaction: a model. PLoS Comput Biol. 2013; 9(8):e1003164. PMC: 3738471. DOI: 10.1371/journal.pcbi.1003164. View

15.

Billeh Y, Cai B, Gratiy S, Dai K, Iyer R, Gouwens N . Systematic Integration of Structural and Functional Data into Multi-scale Models of Mouse Primary Visual Cortex. Neuron. 2020; 106(3):388-403.e18. DOI: 10.1016/j.neuron.2020.01.040. View

16.

Mendoza-Halliday D, Major A, Lee N, Lichtenfeld M, Carlson B, Mitchell B . A ubiquitous spectrolaminar motif of local field potential power across the primate cortex. Nat Neurosci. 2024; 27(3):547-560. PMC: 10917659. DOI: 10.1038/s41593-023-01554-7. View

17.

Katsanevaki C, Bastos A, Cagnan H, Bosman C, Friston K, Fries P . Attentional effects on local V1 microcircuits explain selective V1-V4 communication. Neuroimage. 2023; 281:120375. DOI: 10.1016/j.neuroimage.2023.120375. View

18.

Peterson E, Rosen B, Belger A, Voytek B, Campbell A . Aperiodic Neural Activity is a Better Predictor of Schizophrenia than Neural Oscillations. Clin EEG Neurosci. 2023; 54(4):434-445. PMC: 11905180. DOI: 10.1177/15500594231165589. View

19.

Eisen A, Kozachkov L, Bastos A, Donoghue J, Mahnke M, Brincat S . Propofol anesthesia destabilizes neural dynamics across cortex. Neuron. 2024; 112(16):2799-2813.e9. DOI: 10.1016/j.neuron.2024.06.011. View

19.

Xiong Y, Donoghue J, Lundqvist M, Mahnke M, Major A, Brown E . Propofol-mediated loss of consciousness disrupts predictive routing and local field phase modulation of neural activity. Proc Natl Acad Sci U S A. 2024; 121(42):e2315160121. PMC: 11494327. DOI: 10.1073/pnas.2315160121. View