Opendata, web and dolomites

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 2 - EngineeringPercepts (Reverse engineering sensory perception and decision making: bridging physiology, anatomy and behavior)

Teaser

Unraveling the mechanistic principles that underlie perceptual decision making is extremely challenging, because even the simplest sensory-motor task activates hundred thousands of neurons distributed throughout the entire brain. Moreover, the data provided by the sensory...

Summary

Unraveling the mechanistic principles that underlie perceptual decision making is extremely challenging, because even the simplest sensory-motor task activates hundred thousands of neurons distributed throughout the entire brain. Moreover, the data provided by the sensory systems, representing the state of the world, is noisy. Yet, the brain is able to transform the noisy sensory input across the hierarchy of cortical processing stages, triggering flexible and nuanced decisions – a hallmark of higher cognition. How the brain is able to accomplish this is still unknown. This is why understanding the neuronal mechanisms underlying decision making is one of the major challenges in systems neuroscience and a crucial step to derive general principles that will influence such varied fields as psychology, economics and political science.

The leading approach to probe the neuronal basis of perceptual decisions is ‘top-down’, where statistical theories (e.g. Bayesian inference) are formulated in terms of computational decision elements and neuronal recording/imaging during sensory-motor tasks seeks to identify neural substrates of these elements. This led to identification of neuronal correlates (e.g. changes in firing rate) of sensation, memory, and action in primates during tactile-based decision tasks, suggesting a hierarchically organized cortical system, in which sensory information in primary somatosensory cortex (S1) is gradually transformed into choice in the frontal cortical areas. Similar serial information flow has also been observed in the rodent vibrissal system during a whisker-based decisions task. Now, the crucial question is how these correlates are implemented within the underlying neural networks.
The most promising approach to answer this question is monitoring the sensory-evoked signal-flow throughout the brain, at subcellular resolution and millisecond precision. To do so, the ERC project “EngineeringPercepts” proposes reverse engineering the sensory-motor pathways involved in a well-controlled decision task, by reconstructing the 3D structure and synaptic wiring of the underlying neural networks, transforming the reconstructed circuits into a model that is populated with sensory-evoked responses and simulating signal flow at cellular/network levels. This ‘bottom-up’ approach will provide unmatched understanding of how the interplay between cellular (e.g. biophysics, morphology) and network properties (e.g. synaptic innervation, synchrony) gives rise to neural substrates of perceptual decisions, and it will function as a showcase of how to derive generalizable principles across sensory modalities and species.

More specifically, the ‘bottom-up’ approach will be used to reverse engineer the circuits involved in a whisker-based sensory-motor decision. Rodents can cross a gap by detecting its far side with just a single facial whisker. This suggests that the percept of touching the gap’s side during exploratory movement of the whisker (active whisker touch) triggers the decision making. Evidence of such whisker touch-based correlates of sensation/action have been identified in the vibrissal part of primary somatosensory cortex (vS1) and anterior lateral motor cortex (vM1), respectively. We hypothesize a cellular/network mechanism that encodes object location by active whisker touch. The hypothesis is centered around the major output cell type of the cortex, so called thick-tufted pyramidal tract neurons in layer 5 (L5tt PTs). L5tt PTs in vS1 receive whisker touch input primarily within their proximal dendrites via whisker-specific neurons located in the thalamic ventral posterior medial nucleus (VPM). In contrast, information of whisker movement primarily impinges onto distal dendrites of L5tt PTs. Moreover, L5tt PTs are able to change their activity patterns when inputs into proximal and distal dendrites coincide within a narrow temporal window. L5tt PTs are thus the ideal candidate to encode ob

Work performed

The proposal is based on three main project categories (A-C), which are subdivided into eight parts. Each part addresses one key question about the nature of the synaptic inputs that shape sensory-evoked spiking responses of L5tt PTs in vS1. In the following, I will describe the progress and achievements for four of these subprojects:

In Part A1, we wanted to investigate relationships between structural and whisker-evoked functional parameters of inhibitory interneurons (INs), and use this data to constrain simulations. In collaboration with the department of Dr. Kerr at the Center of Advanced European Studies and Research in Bonn (Germany), we have started this project by recording activity patterns from inhibitory interneurons (INs) in Layer 1 (L1) of vS1. We found that L1 INs have reliable responses during passive whisker touches, and that these responses are as fast as those of the excitatory neurons in L2-6. Moreover, we were able to recover the morphologies of the in vivo recorded neurons, which revealed that axons of L1 INs largely innervate distal dendrites of excitatory neurons. We used the structural and functional L1 IN data, and performed physiologically, anatomically and biophysically constrained simulations. The simulations predicted that L1 INs regulate the robustness of sensory-evoked excitatory responses across stimulation trials, without affecting the magnitude of the response itself. These predictions were confirmed by in vivo pharmacology experiments. The study, which was published in the Proceedings of the National Academy of Sciences (USA), can thus be regarded as a showcase that illustrates the feasibility of our reverse-engineering strategy, and how this approach can be used to provide insight into across-scale mechanisms that shape neuronal responses during sensory stimulation.

In Part B1, we wanted to investigate relationships between whisker-evoked responses and the specific target areas throughout the brain that receive input from L5tt PTs in vS1.We have accomplished the primary goal of this subproject. In a study, which was published in Nature Communications, we showed how to combine injections of retrograde tracer agents into several subcortical target structures of vS1 with recording and labeling of individual L5tt PTs in vivo. Our study revealed several surprising structure-function relationships. First, we found that somata of L5tt PTs with different subcortical targets form sublayers in vS1. Second, dendrite distributions of L5tt PTs within each cortical layer correlated with the respective long-range target. Third, activity patterns of L5tt PTs during both, periods of ongoing activity and whisker stimulation, correlated with the respective long-range target. We showed that these cellular properties result in a structure-function parameter space that allows predicting the L5tt PTs’ subcortical target regions, without the need to inject multiple retrograde tracers. These findings indicated that stimulus features are differentially extracted by L5tt PTs via long-range target-specific subnetworks.

In Part B2, we wanted to reconstruct the whisker-related motor networks throughout the brain. In collaboration with the department of Prof. Strick at the University of Pittsburgh (USA), we have accomplished one of the primary goals of this subproject. In a study, which was published in Neuroscience, we showed how to combine injections of recombination competent rabies virus into individual whisker muscles with brain-wide quantifications of trans-synaptically connected networks. We established the rabies injection methodology for individual facial muscles in rodents, generated a digital 3D model of the motoneuron networks within the brain stem, and identified the synaptic input populations of these whisker muscle-innervating motoneurons. In line with previous reports, we found that these motoneurons receive no direct synaptic input from cortical neurons. Instead, their synaptic input arises fr

Final results

As described above, we have accomplished several goals of the project. Towards the end of the project we thus expect to show how the structural and functional data – as acquired in Parts A1-B3 – can be used to constrain simulations that mimic in vivo conditions during whisker touch. Similar to approaches in other complex dynamical systems (e.g. climate research), we are currently developing a multi-scale model that can be regarded as the most comprehensive digital representation of the cortical circuitry to date. The model allows performing predictive simulations of neuronal processing – with subcellular and millisecond resolution – that mimic in vivo conditions, and that can directly be tested empirically. We will use the model uncover the mechanisms that activate L5tt PTs during sensory stimulation – insight that will be a necessity for understanding how cortical computations orchestrate sensory-guided behaviors (e.g. perceptual decision making).

Moreover, my group has developed several novel experimental, simulation and analysis approaches. First, we established trans-synaptic tracing via injections of recombination competent rabies virus into individual whisker muscles in rats. Second, we established in vivo recording and labeling of individual cortical neurons, whose long-range target structures were unambiguously identified by injections of multiple retrograde tracer agents into different subcortical brain areas. Third, we developed multi-scale models that allow performing simulations of neuronal activity that mimic in vivo conditions.

Furthermore, to illustrate that our reverse engineering approach can be regarded as a general strategy to investigate the interplay between biophysical, cellular and network properties, we have started a collaboration with the laboratory of Prof. Long at NYU School of Medicine in New York (USA). Prof. Long investigates mechanisms that underlie vocalization in the song bird. As illustrated by two scientific articles in Elife and the Journal of Comparative Neurology, we have successfully transferred our approaches for reconstructing neuronal networks from the rodent to the bird brain.