Thalamic Control Over Laminar Cortical Dynamics Across Conscious States
Poster No:
2018
Submission Type:
Abstract Submission
Authors:
Eli Müller1, James Shine2, Brandon Munn1, Ben Fulcher3, Bing Brunton4, Steve Brunton4, Vicente Medel5, Michelle Redinbaugh6, Yuri Saalman7, Giulia Baracchini8
Institutions:
1University of Sydney, Sydney, NSW, 2The University of Sydney, Sydney, NSW, 3University of Sydney, Sydney, Australia, 4University of Washington, Seattle, WA, 5Universidad Adolfo Ibáñez, Santiago, RM, 6Stanford University, San Francisco, CA, 7University of Wiscosin-Madison, Madison, WI, 8The University of Sydney, Sydney, New South Wales
First Author:
Co-Author(s):
Introduction:
The human brain supports a wide-ranging behavioural repertoire, with dynamic, context-dependant recruitment of specialized neuronal assemblies emerging from coordinated interactions amongst tens of billions of cells. The computational capacities that support these behaviours are in large part related to this cellular complexity, however, this high-dimensional neural activity is of minimal adaptive benefit unless the behaviours that emerge are both stable and yet flexible. In other words, neural dynamics must be reliable for repeated precise execution, but also rapidly adaptable to changing contingencies. How the intricate organization of the brain supports these opposing capacities remains to be understood.
Methods:
In this paper, we repurpose an approach from fluid dynamics for application to functional neuroimaging data. Specifically, we develop a novel measure of laminarity by tracking the ability of a linear dynamic model to accurately predict successive time-points. That is, we applied a simple linear model estimation technique to a high spatiotemporal resolution human 7T fMRI dataset captured during the resting state. First, a linear propagator tracking moment-to-moment shifts in a recording is generated for each subject's timeseries by utilizing the singular-value-decomposition. Then, for every timepoint in the recording, the linear propagator is used to forecast state dynamics at a set of future timepoints (τ; 1-10 TR). These predicted laminar dynamics are then compared to the ground-truth dynamics via the mean-squared-error, providing a time-resolved read-out of how effective (or not) the linear model was at predicting the upcoming brain state.
Results:
We show that neural recordings transition through periods of increased laminarity that irregularly fluctuate across time. The spatial and temporal structure of this laminarity reveals coherence maps and intrinsic timescales that are not captured by common summary measures, including covariance and Fourier power spectrum. We then reveal how pharmacological manipulation of arousal, via propofol-induced anesthesia, results in increased laminar dynamics in human neuroimaging data. Next, we demonstrate how targeted and diffuse thalamocortical subtypes can both promote or diminish laminar cortical dynamics, respectively. We underscore the importance of thalamocortical interactions in a detailed biophysical model which further predicts matrix thalamic stimulation re-instantiates non-laminar dynamics concomitant with conscious arousal, which we then confirm by applying our approach to multielectrode cortical recordings from an anesthetized macaque that was awoken by electrical stimulation of the diffuse thalamic projections within the central lateral thalamus.
Conclusions:
In conclusion, we have demonstrated that fluctuations in laminar flow in multimodal neuroimaging data that differentiate conscious arousal states are controlled by the organization of the thalamus. In this way, we argue that this key feature of the brain's neurobiological architecture helps to shape the robust, yet flexible, adaptive neural dynamics required for effective cognitive function that define our waking lives.
Brain Stimulation:
Invasive Stimulation Methods Other
Modeling and Analysis Methods:
Methods Development 2
Perception, Attention and Motor Behavior:
Consciousness and Awareness 1
Physiology, Metabolism and Neurotransmission:
Pharmacology and Neurotransmission
Keywords:
Computational Neuroscience
Consciousness
Thalamus
1|2Indicates the priority used for review