Periodic Reporting for period 1 - BRAINCIP (Brain-wide Criticality and Information Processing)
Periodo di rendicontazione: 2023-09-01 al 2025-08-31
The brain is an adaptive system that efficiently performs unique functions in complex and continuously changing environments. Through our brain we experience, learn, progress, adapt, and create memories that make each of us unique. In fact, although we all share the same brain architecture, each of us develops his own representation of objects and surrounding environments. Why is it so? How does our brain elaborate sensory inputs to produce responses and ultimately create our very own representation of the world? Questions of this sort puzzled philosophers and scientists for centuries, and still represent a major challenge for modern science. Recent studies suggest that our sensory experience is crucially shaped by the ongoing activity of the brain, which is thought to represent the brains internal state. However, such a state remains, to a large extent, an abstract entity, difficult to capture and quantify.
Empirical observations suggest that the brain self-organizes to operate near “criticality”, a peculiar state located at the border between order and disorder. An ordered state is rigid and the information flow among units of the system is low. Conversely, a disordered state is highly erratic with unbounded spreading of perturbations and noise that prevents effective information processing. Criticality interpolates between these two states, and is characterized by unique features that could be advantageous for brain function: long-range correlations, maximal variability of spatio-temporal patterns, and wide response range to external perturbations. With these properties, criticality would provide the brain with the ability to efficiently integrate information across spatial and temporal scales, and promptly and adequately respond to external stimuli by generating a rich variety of coordinated collective behaviors, thus optimizing brain function. This is the very core of the “critical brain hypothesis”. First proposed about 40 years ago, this hypothesis basically states that the brain “lives” at criticality. Theoretical arguments and models indicate that criticality could support optimal computational capabilities by providing a favorable trade-off between reliability and flexibility needed to support complex brain functions. Despite signatures of criticality in brain activity across species (including humans), the alleged role of criticality in brain functions currently lacks direct in vivo empirical confirmation.
This project aims to test the key assumptions of the critical brain hypothesis by investigating the relation between brain criticality and functions primarily in zebrafish larvae. Why zebrafish larvae? First, because the zebrafish is a vertebrate with a relatively high genetic homology to mammals and humans, and a nervous system that shares basic architecture with other vertebrates. This often allows to generalize discoveries in zebrafish to other vertebrates, including mammals. Second, in the larval stadium the zebrafish is small and transparent, and powerful imaging and genetic methods can be applied to follow whole-brain neural activity with cellular resolution during behavioral study. To assess the functional role of criticality in systems that are located at different stages of the vertebrate evolutionary phylogenetic tree, the analysis was extended to rats and humans. This does not only inform about the universal character of self-organization to criticality in neural networks of vertebrates, but also provides first insights into its evolutionary relevance.
Empirical observations suggest that the brain self-organizes to operate near “criticality”, a peculiar state located at the border between order and disorder. An ordered state is rigid and the information flow among units of the system is low. Conversely, a disordered state is highly erratic with unbounded spreading of perturbations and noise that prevents effective information processing. Criticality interpolates between these two states, and is characterized by unique features that could be advantageous for brain function: long-range correlations, maximal variability of spatio-temporal patterns, and wide response range to external perturbations. With these properties, criticality would provide the brain with the ability to efficiently integrate information across spatial and temporal scales, and promptly and adequately respond to external stimuli by generating a rich variety of coordinated collective behaviors, thus optimizing brain function. This is the very core of the “critical brain hypothesis”. First proposed about 40 years ago, this hypothesis basically states that the brain “lives” at criticality. Theoretical arguments and models indicate that criticality could support optimal computational capabilities by providing a favorable trade-off between reliability and flexibility needed to support complex brain functions. Despite signatures of criticality in brain activity across species (including humans), the alleged role of criticality in brain functions currently lacks direct in vivo empirical confirmation.
This project aims to test the key assumptions of the critical brain hypothesis by investigating the relation between brain criticality and functions primarily in zebrafish larvae. Why zebrafish larvae? First, because the zebrafish is a vertebrate with a relatively high genetic homology to mammals and humans, and a nervous system that shares basic architecture with other vertebrates. This often allows to generalize discoveries in zebrafish to other vertebrates, including mammals. Second, in the larval stadium the zebrafish is small and transparent, and powerful imaging and genetic methods can be applied to follow whole-brain neural activity with cellular resolution during behavioral study. To assess the functional role of criticality in systems that are located at different stages of the vertebrate evolutionary phylogenetic tree, the analysis was extended to rats and humans. This does not only inform about the universal character of self-organization to criticality in neural networks of vertebrates, but also provides first insights into its evolutionary relevance.
To probe the functional role of brain criticality in vivo, we first studied brain activity in absence of external (sensory) stimuli and specific tasks. This is the “baseline state” of the brain and should correspond to criticality. Activity of nearly 40000 neurons in the zebrafish brain was recorded using 2-photon calcium imaging, an optogenetic technique that exploits genetic tools and light microscopy to detect the activity of individual neurons. What to look at to identify “criticality”? The most striking property of systems at criticality is scale invariance, or self-similarity, which is the defining attribute of fractals. Quantifying scale-invariance in brain dynamics was key to assess the functional role of criticality. After verifying that the brain of zebrafish operates close to criticality in absence of stimuli and tasks, we asked how the brain state influences the behavioral performance of the fish in response to certain visual stimuli. To address this question, we had to alter the “baseline state” of the brain and move it away from criticality using light, genetic, or pharmacological “perturbations”. Then, we compared the performance of the “critical” and “non-critical” brains. We found that, in fishes away from criticality, the response—e.g. swimming behavior—significantly deviated from the normal, expected behavior observed in “critical” fishes. We quantified the impact of tuning to criticality on processing of sensory stimuli and behavior (optomotor response) in Zebrafish with CDKL5 deficiency disorder (a model of autism associated with impairments in learning, memory, and social interaction) and in fish exposed to pentylenetetrazole (PTZ), a model for epilepsy.
In humans, we investigated tuning to criticality in adults and infants. In particular, we found evidence of tuning to criticality already in the first few days after birth. In particular, we demonstrated that during the early stage of development specific external stimuli, i.e. speech in the native language, drive the brain closer to criticality. Experience of familiar, native language—but not of other languages—significantly modulate the internal state of the brain towards criticality. This represents first evidence of a stimulus-driven tuning to criticality that is associated with functional needs in humans, e.g. learning.
Sleep plays a key role in preserving brain function, keeping brain networks in a state that ensures optimal computation. We found that this state is consistent with criticality, and identified the control circuitry of tuning to criticality in rats across the sleep-wake cycle. By silencing key sleep and wake promoting brain areas, we showed that criticality across the sleep-wake cycle of rats is supported by a balanced control network whose key sub-cortical nodes are the wake-promoting locus coeruleus (LC) and the sleep-promoting ventrolateral preoptic nucleus (VLPO). We then identified a cortical node—the anterior cingulate network (ACC), densely connected to LC—of this network.
In humans, we investigated tuning to criticality in adults and infants. In particular, we found evidence of tuning to criticality already in the first few days after birth. In particular, we demonstrated that during the early stage of development specific external stimuli, i.e. speech in the native language, drive the brain closer to criticality. Experience of familiar, native language—but not of other languages—significantly modulate the internal state of the brain towards criticality. This represents first evidence of a stimulus-driven tuning to criticality that is associated with functional needs in humans, e.g. learning.
Sleep plays a key role in preserving brain function, keeping brain networks in a state that ensures optimal computation. We found that this state is consistent with criticality, and identified the control circuitry of tuning to criticality in rats across the sleep-wake cycle. By silencing key sleep and wake promoting brain areas, we showed that criticality across the sleep-wake cycle of rats is supported by a balanced control network whose key sub-cortical nodes are the wake-promoting locus coeruleus (LC) and the sleep-promoting ventrolateral preoptic nucleus (VLPO). We then identified a cortical node—the anterior cingulate network (ACC), densely connected to LC—of this network.
The results represent a first causal link between criticality and cognitive performance, a major step forward in brain research and a key support for the brain criticality hypothesis. While supporting the critical brain hypothesis, our results represent an important advancement in the quantification and perturbation of the “brain computational function”, which is essential to neuroscience research and may help building more efficient, brain-like neural networks. Measuring and perturbing a collective signal, such as distance to criticality, is a novel way of experimentally interacting with the nervous system, which may have a significant impact on neuroscience research. Establishing criticality as a solid measure of the network state is a key step toward a new generation of biomarkers for early detection of neurological disorders based on a rigorous, quantitative assessment of brain computational abilities through the analysis of neural dynamics. Current outcomes have the potential to further develop on a measure of brain activity that universally correlates with task performance, which would be a major scientific innovation.