P094 Two coupled networks with a tricritical phase boundary between awake, unconscious and dead states capture cortical spontaneous activity patterns
Maryam Ghorbani1,2*, Negar Jalili Mallak3, Mayank R. Mehta4,5,6
1Department Electrical Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
2Rayan Center for Neuroscience and Behavior, Ferdowsi University of Mashhad, Mashhad, Iran
3School of Computing and Augmented Intelligence, Arizona State University, Tempe, AZ
4UCLA, W.M. Keck Center for Neurophysics, Department of Physics and Astronomy, Los Angeles
5UCLA, Department of Neurology, Los Angeles, CA, United States of America
6UCLA, Department of Electrical and Computer Engineering, Los Angeles
*Email: maryamgh@um.ac.ir
Introduction
A major goal in systems neuroscience is to develop biophysical yet minimal theories that can explain diverse aspects of in vivo data accurately to reveal the underlying mechanisms. Under a variety of conditions, cortical activity shows spontaneous Up- and Down-state (UDS) fluctuations (1, 2). They are synchronous across vast neural ensembles, yet quite noisy, with highly variable amplitudes and durations (3). Here we tested the hypothesis that this complex pattern can be captured by just two weakly coupled, noiseless, excitatory-inhibitory (E-I) networks.
Methods
The model consisted of two mean-field E-I networks, with recurrent, long-range excitatory connections. The LFP and single unit responses were measured from various parts of the parietal and frontal cortices of 8 naturally resting rats using tetrodes. Parietal cortical LFP in anesthetized mice was measured from 116 animals from the deeper parts of the neocortex. The animals were anesthetized using urethane only once during this recording session.
Results
The model could reproduce recently observed periodic versus highly variable UDS in strongly versus weakly coupled organoids respectively. The same model could quantitatively capture the differential patterns of UDS in vivo during anesthesia and natural NREM sleep. Further, by varying just two free parameters, the strength of adaptation and of recurrent connection between the two networks, we made 18 quantitative predictions about the complex properties of UDS. These not only matched experimental data in vivo, but could reproduce and explain the systematic differences across electrodes and animals.
Discussion
The model revealed that, the cortex remains close to the awake-UDS phase boundary in all the sleep sessions but near awake-UDS-dead tricritical phase boundary during anesthesia. Thus, just two weakly coupled mean-field networks, with only two biophysical parameters, can accurately capture cortical spontaneous activity patterns under a variety of conditions. This has several applications, from understanding stimulus response variability, to anesthesia and cortical state transitions between awake, asleep and unconscious.
Acknowledgements
None
References
1.https://doi.org/10.1523/JNEUROSCI.13-08-03252.1993.
2.https://doi.org/10.1523/JNEUROSCI.19-11-04595.1999.
3.https://doi.org/10.1523/JNEUROSCI.0279-06.2006.
