Cognition and consciousness arise from analog computations, says new theory

A new theory, published in The Journal of Neuroscience by three scientists in The Picower Institute for Learning and Memory at MIT, offers an explanation of how the brain produces cognition and consciousness: It uses traveling waves of rhythmic neural activity to coordinate nimble neural networks with analog computations. 

The metaphor that the brain operates with “circuits” is incomplete, says Picower Professor Earl K. Miller, the paper’s senior author. Indubitably, the brain’s physically connected circuits provide the infrastructure to store our memories and represent our ongoing needs and goals. But when we need to make improvised use of that knowledge in the rapid-fire, anything-goes sensory context the world constantly throws our way, we can’t just depend on the relatively slow chemical process of rewiring those circuit connections called “synapses,” he says. 

Instead, the brain needs a control system that can coordinate millions of neurons to process information in a fraction of a second. Brain waves, long understood to be the synchronized rhythmic fluctuations of large groups of neurons, turn out to be performing that crucial service, Miller and his colleagues argue, citing years of experimental evidence from his lab and many others.

“Circuits and synapses are important and fundamental, that’s the start. But there is more going on,” says Miller, a member of MIT’s Department of Brain and Cognitive Sciences faculty. “The brain generates waves, and wave dynamics are a highly efficient way to coordinate and perform computation.”

While digital circuits make calculations one step at a time through sequential switches and gates, analog computation, which can be performed via the interference of waves, processes multiple calculations in parallel. That’s not only more efficient, but also locally focused traveling waves happen to be a ubiquitous feature of the brain, the scientists note.

“The brain exploits its own physics,” wrote Miller and co-authors Scott L. Brincat and Jefferson E. Roy, who are research scientists in Miller’s lab.

The new theory is important not only because it provides an explanation of cognition and consciousness, but also because it asserts the potential importance of considering waves in clinical treatment. Conveniently, waves can be manipulated non-invasively.

“Developing treatments based on brain wave dynamics is not just an opportunity, but also an obligation,” says Miller, whose lab is part of a collaboration studying brain waves in autism.

Building the analog argument

To make the case that the brain uses waves to coordinate neurons to produce cognition and consciousness, the scientists begin with the now well-established observation that many neurons don’t just do one job. Instead, they respond to multiple cues and contexts, essentially participating in multiple functional networks at once, a property called “mixed selectivity.” Miller and colleagues have argued for years that this gives the brain immense computational horsepower, but it also initially raised the question of how the brain organizes these multiple overlapping networks with such speed and flexibility to produce the nimble thought we all depend on.

After numerous studies, the answer that has emerged for Miller and many other neuroscientists is that brain waves organize neural ensembles to process information. Miller has shown that brain waves of different frequencies govern cognitive processes such as working memory and predictive coding. Relatively slow “alpha” and “beta” frequency waves, representing memories and goals, regulate faster frequency “gamma” waves, which represent and report incoming sensory information.

The new theory posits that these alpha/beta control waves emerge from the coordinated spiking of neurons in circuits (connected at junctions called “synapses”) that encode stored memories and goals. 

“Synapses store representations, while wave dynamics help determine which representations are active at any given time,” the authors wrote.

In some of the Miller lab’s newer research, the team has found evidence that even as waves emerge from neural spiking, the waves can rapidly grow to directly influence and coordinate spiking via an electric field-mediated process called ephaptic coupling. Importantly, electric fields can exert this coordinating influence very rapidly.

Another essential component of the theory, which Miller’s lab has also shown experimentally, is that alpha/beta waves are capable of exerting their control spatially, by affecting local areas of the cortex, and temporally, by traveling along the cortex. Essentially, the beta waves act as mobile stencils that govern where and when gamma waves can process sensory information and which ensembles of neurons will participate. Taken together, this suggests that the brain engages in “spatiotemporal computing,” the authors write. And where the waves intersect, they can add and subtract, enabling analog computations.

Miller acknowledges that his lab’s next step should be to provide direct evidence that the analog computations are taking place.

“This is a theory. We aim to test it by looking for signatures of analog computation in brain wave patterns,” Miller says. 

Connection to consciousness

The article asserts that consciousness “emerges when these dynamic wave patterns bring the cortex in an organized, globally integrated state, one that naturally links and influences widespread activity.”

Some of the most compelling evidence linking wave dynamics to consciousness comes from studies of general anesthesia that Miller has conducted with Picower Institute colleague Emery N. Brown, who is an Institute professor at MIT, an anesthesiologist at Massachusetts General Hospital, and a professor in Harvard Medical School. Their labs have shown that three different drugs, each with different molecular mechanisms of action, all similarly disrupt brain wave dynamics to produce unconsciousness.

“Consciousness depends less on specific receptors or cell types and more on the integrity of large-scale wave organization,” the authors write in the review.

In other words, much like cognition, consciousness depends on how the brain efficiently organizes itself with brain waves.

“Electric field dynamics offer a low-overhead substrate for organizing and coordinating information across cortical networks,” they conclude. “Given strong evolutionary pressure to maximize computation per unit energy, it would be surprising if evolution did not exploit such a built-in analog computing substrate.”

The Freedom Together Foundation, The Picower Institute for Learning and Memory, the U.S. Army Research Office, the U.S. Office of Naval Research, a MURI grant, the National Institutes of Health, and the Simons Center for the Social Brain supported the research.

de MIT News https://ift.tt/aYdlKFt

Atlas of the brain’s striatum could guide researchers to new drug treatments

A region of the brain called the striatum is critical for many cognitive and motor functions, including decision-making, control of movement, habit formation, and processing of reward. It also plays a role in addiction and is significantly affected by Huntington’s disease, schizophrenia, and other disorders.

In work that could help scientists devise new treatments for those diseases, MIT researchers have generated a new atlas of the neurons found within the striatum. Using single-cell RNA sequencing and other techniques, they were able to identify 31 subgroups of neurons based on which genes they express.

These groups include neurons that are involved in addiction, depression, and schizophrenia. The researchers also discovered why some neurons of the striatum are more vulnerable to Huntington’s disease. All of these results, the researchers say, could help scientists develop new drugs to combat these conditions.

“We see this as the foundation that will allow more studies in our Huntington’s disease and opioid use disorder projects. We needed a roadmap of what is there,” says Myriam Heiman, the Picower Professor of Neuroscience and director of MIT’s Picower Institute for Learning and Memory.

Heiman; Manolis Kellis, a professor of computer science in MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and a member of the Broad Institute of MIT and Harvard; and Dana Gabuzda, a principal investigator at Dana-Farber Cancer Institute and a professor of neurology at Brigham and Women’s Hospital and Harvard Medical School, are the senior authors of the study, which appears today in Cell. MIT postdoc Raleigh Linville and MIT graduate student Benjamin James are the paper’s lead authors.

Mapping the striatum

The striatum, located deep within the brain, receives diverse inputs from the cortex, midbrain, hippocampus, and other regions, which it uses to coordinate planning, movement, and decision-making, as well as processing reward. In this study, the researchers focused on the most populous cell type in the striatum, a type of inhibitory neuron called the medium spiny neuron, which responds to dopamine.

Most of these medium spiny neurons belong to either the direct pathway, which helps to promote movement, or the indirect pathway, which suppresses unwanted movements. These pathways are distinguishable by what type of dopamine receptor they express — dopamine receptor 1 (D1) or dopamine receptor 2 (D2).

Beyond these two divisions, scientists knew that there were many subpopulations performing different roles, especially in the anatomically ventral (lower) regions of the striatum. However, it has been difficult to generate a consensus on how to classify these cells, in part because prior studies focused on specific subregions, meaning that overarching principles of striatal cellular organization were lacking.

To overcome that challenge, the researchers worked closely with brain banks in the United States and Canada to collect postmortem striatal samples representing diverse anatomical regions. 

Then, they used three different techniques to analyze the samples, including single-cell RNA sequencing — a method that can measure RNA molecules within individual cells to reveal which genes are being expressed. Two additional techniques — multiplexed fluorescent in situ hybridization and spatial transcriptomics — allowed the researchers to identify spatial principles of organization within the tissue.

Using these techniques, the researchers were able to identify 31 different subpopulations of neurons, including nine types of medium spiny neurons. Among their medium spiny neuron types are two “outlier” populations that appear to play important roles in schizophrenia, substance use disorder, and depression.

One of those populations, known as D1 outliers, showed high expression of genes involved in addiction and substance use disorder, especially genes related to opioid response. Another population, called D2 outliers, showed high expression of genes that respond to antidepressants. And, both populations appeared to respond strongly to clozapine, an antipsychotic drug used to treat schizophrenia.

Clozapine is among the most effective antipsychotics available, but it’s not widely used in the United States because it can cause a fatal blood disorder in a small percentage of patients. Now that researchers know which cells the drug acts on, they may be able to design more targeted therapeutics to overcome psychosis, but without the harmful side effects, Heiman says.

Huntington’s vulnerability

Another key finding of the paper helps to shed light on why the dorsal (upper) part of the striatum is more vulnerable to Huntington’s disease. The disease is caused by an inherited version of the huntingtin gene that carries too many repetitive DNA segments, called CAG repeats. 

The researchers found that dorsal populations of medium spiny neurons express higher levels of the genes MSH2 and MSH3, which play a role in increasing the number of CAG repeats found in the huntingtin gene. As more of those repeats accumulate, the mutated version of the huntingtin protein becomes more harmful to cells.

The researchers also found that a rare population of medium spiny neurons that forms island-like structures in the ventral striatum was more resistant to the accumulation of CAG repeats. Further study of this class of cells might help researchers learn how to induce other medium spiny neurons to become more resistant to the disease, Heiman says.

“Looking at the genes that these neurons express or don’t express might give us some clues as to how to make other medium spiny neurons resilient like them,” she says. 

Insights into substance use disorders

The researchers also compared their findings from human tissue samples to samples from mice and found several differences, especially in the expression of genes related to drug response and substance use disorders. One such gene, which encodes the mu opioid receptor (OPRM1), is highly expressed in the human D1 outlier population, but not in the corresponding population of neurons in mice. 

This means that standard mouse models may not fully capture the biology of opioid responses, and that engineering mice to express this receptor in a similar manner to humans could make those models significantly more accurate.

“Some of the diversity we’re seeing in the human ventral striatum is species-specific and has implications for modeling substance use disorder in rodents,” Heiman says. “Now that we understand better the species differences, we can use the rodent models for specific questions that apply for conserved genes, but we could also think about humanizing some models.”

The researchers hope that this map, built from tissue contributions by brain donors and their families, and assembled across disciplines and institutions, will provide an important starting point for researchers pursuing new treatments for some of the most difficult-to-treat brain disorders.

The research was funded, in part, by the National Institutes of Health, the G. Harold and Leila Y. Mathers Charitable Foundation, the Freedom Together Foundation, the Natalia Mental Health Foundation, the Biswas Family Foundation, and the Milken Institute.

de MIT News https://ift.tt/Gr2gfEQ