INS Seminar: 10 Jan. 2019 – 14:00 – “Deciphering basal ganglia function using computational models”

From Thursday 10th January 2019 at 14:00
To Thursday 10th January 2019 at 15:00

Location : INS Seminar Room, Campus Timone, Red Wing, 5th Floor

Speaker Jyotika Bahuguna (Forschungszentrum Julich, Germany)
 
Abstract :  Basal ganglia function can be addressed with computational modeling at different levels of abstractions. Here I demonstrate two such examples and an attempt to combine these models. Firsty, we address the question of what is the striatal representation of an motor action.  In order to investigate this issue, we designed a distance dependent spiking neuronal network model of the striatum consisting of D1 and D2 medium spiny neurons (MSNs) and interfaced it to a simulated robot moving in an environment. We demonstrate that this model is able to reproduce key behavioral features (freezing, ambulation and rotation) of 6 out of 7 optogenetic experiments that involved the manipulation of the striatum. The main result of this model was that D1 and D2-MSNs of an action co-operate whereas D1 and D2-MSNs of competing actions inhibit each other during action selection. Basal ganglia being a set of interacting nuclei and forming many functional pathways form a good substrate for degeneracy. This degeneracy might also explain the variability seen in the data in healthy as well as pathological conditions such as Parkinson's disease. In order to investigate this issue, we model the basal ganglia as a firing rate model and perform a parameter search for effective connectivities between its nuclei for healthy and dopamine depleted conditions. The cost function used for constraining this system was derived from empirical firing rates and phase relationships as observed in healthy and dopamine depleted rats. We were able to generate more than 1000 physiological and pathological firing rate models that met the constraints and showed ample variability in the values of effective connectivities. We then projected these models onto a lower dimensional space of dynamical features such as : a) GPi suppression b) Susceptibiity to oscillations. Despite the large variability in effective connectivities, the models clustered together in this space and showed a clear separation between physiological and pathological conditions. This suggests that rather than absolute values of the effective connectivities, it might be their relative values that determine the dynamical state and projecting them on a lower dimensional space of sensible dynamical features might give a better chance at understanding complex pathologies such as Parkinson's disease than a pure structural analysis. And lastly, we use these firing rate models to deconstruct basal ganglia transfer function in response to striatal optogenetic stimulation in order to explain one of the optogenetic experiments that we failed to explain in the first study.
 
For any question, feel free to contact:
Hiba Sheheitli (hiba.sheheitli@univ-amu.fr) or Sophie Chen (sophie.chen@univ-amu.fr)


INS Seminar: 13 Dec. 2018 – 14:00 – “What information dynamics can tell us about brains”

From Thursday 13th December 2018 at 14:00
To Thursday 13th December 2018 at 15:00

Location : INS Seminar Room, Campus Timone, Red Wing, 5th Floor

Speaker: Dr. Joseph T. Lizier, The University of Sydney

Abstract: The space-time dynamics of interactions in neural systems are often described using terminology of information processing, or computation, in particular with reference to information being stored, transferred and modified in these systems. In this talk, we describe an information-theoretic framework -- information dynamics --  that we have used to quantify each of these operations on information, and their dynamics in space and time. Not only does this framework quantitatively align with natural qualitative descriptions of neural information processing, it provides multiple complementary perspectives on how, where and why a system is exhibiting complexity. We will review the application of this framework in computational neuroscience, describing what it can and indeed has revealed in this domain. First, we discuss examples of characterizing behavioral regimes and responses in terms of information processing, including under different neural conditions and around critical states. Next, we show how the space-time dynamics of information storage, transfer and modification directly reveal how distributed computation is implemented in a system, highlighting information processing hot-spots and emergent computational structures, and providing evidence for conjectures on neural information processing such as predictive coding theory. Finally, via applications to several models of dynamical networks and human brain images, we demonstrate how information dynamics relates the structure of complex networks to their function, and how it can invert such analysis to infer structure from dynamics.

For any question, feel free to contact:
Hiba Sheheitli (hiba.sheheitli@univ-amu.fr) or Sophie Chen (sophie.chen@univ-amu.fr)



INS Seminar – 5 Dec. 2018 – 15:00 : “Resonance of Local Field Potentials in the Connectome”

From Wednesday 5th December 2018 at 15:00
To Wednesday 5th December 2018 at 16:00

Location : INS Seminar Room, Campus Timone, Red Wing, 5th Floor

 Speaker : Joana Cabral (University of Oxford, UK & University of Minho, Portugal)

"Resonance of Local Field Potentials in the Connectome "

 Abstract :  I will describe a mechanistic theory for the transient emergence of macroscopic brain rhythms as collective oscillatory modes emerging transiently from reciprocal interactions between local field potentials in the structural skeleton of the Connectome. This mechanism is grounded on theoretical principles governing the formation of frequency-specific coherent attractors in delay-coupled oscillatory systems. Using a reduced phenomenological network model representing interactions between voltage fluctuations generated locally by neuronal ensembles at 40Hz with realistic wiring and propagation times, numerical simulations reveal the transient emergence of spatially-organized collective oscillatory modes peaking between 0.5-30Hz in line with analytic predictions, matching spectral, spatial and temporal signatures of multimodal neuroimaging data.

For any question, feel free to contact:
Hiba Sheheitli (hiba.sheheitli@univ-amu.fr) or Sophie Chen (sophie.chen@univ-amu.fr)