All day long, your brain cells are sending and receiving messages through electrical and chemical signals. These messages help you do things like move your muscles and use your senses — as you taste your food, feel the heat coming off a stove, or read the words on this page.
If we could better understand how those messages are sent and received, we’d gain powerful insights into the brain-body connection and shed light on what’s happening when those connections aren’t working — as with brain diseases such as Alzheimer’s and Parkinson’s.
To that end, neuroscientists at Cedars-Sinai in Los Angeles have built computer models of individual brain cells — the most complex models to date, they say. Using high-performance computing and artificial intelligence, or AI, the models, as described in the journal Cell Reports , capture the shape, timing, and speed of the electrical signals that brain cells called neurons fire.
The new research is part of a decades-long pursuit among scientists to understand the inner workings of the brain, not just cognitively but biologically, genetically, and electrically.
The most famous early researchers were Alan Lloyd Hodgkin, Andrew Fielding Huxley, and John Carew Eccles, who won the 1963 Nobel Prize in Medicine for their discoveries about nerve cell membranes.
“Today is a unique moment when detailed, single-neuron data sets are available in large quantities and for many cells,” says study author, Costas Anastassiou, PhD, a research scientist in the Department of Neurosurgery at Cedars-Sinai. “The size and speed of today‘s computers allows us to explore [detailed] mechanisms at a single-cell level — for every cell.”
How Do You Model Brain Cell Activity Using a Computer?
Turns out, the electrical pulses neurons use to communicate can be replicated using computer code. Each cell’s “action potentials” — or “spikes” in electrical activity — can be represented by zeros and ones. The “one” is when the cell’s voltage rises above the “spike threshold,” and the “zero” is when it’s below that threshold.
“We replicated the distinct voltage waveforms and time trajectories of these pulses using mathematical equations,” says Anastassiou. Then they built computer models using data sets from experiments in mice.
These experiments measure certain things in the cells — like their size, shape, and structure, or how they respond to changes. Each cell model combines all these elements and can help reveal how they connect.
“A large part of what we know about the cellular makeup of the brain comes from gene expression — ‘omics’ studies,” Anastassiou says. “On the other hand, most of what we know about human cognition comes from measuring and monitoring cellular voltage dynamics in the living, breathing animal brain.”
Computer models can reconcile two critical pieces of information: the cellular makeup (building blocks of brain cells) and the patterns observed during brain activity. With the computer’s help, links between the data sets become clear. This could help pave the way from association to causation, the researchers say — a crucial step when looking at disorders.
What Can Computers Tell Us About the Human Brain?
One of the exciting potential uses of the brain cell models would be to test all kinds of theories about brain disorders that would be difficult or impossible to create through experiments in the lab.
Beyond that, the work can lead to new insights about the brain: how similar or different brain cells are, what connects or makes them different, and what that means across a spectrum of properties.
Computers and mathematics are telling stories about the brain, and Anastassiou says for him, the fascination comes from the simplicity of the outcome and the richness of their impacts.
“I have always been fascinated by the question of how mathematical equations represent living, computing, biological cells — particularly so for the brain, the epicenter of what makes us human,” he says.
Costas Anastassiou, PhD, research scientist, Department of Neurosurgery, Cedars-Sinai, Los Angeles.
Cell Reports: “Single-neuron models linking electrophysiology, morphology, and transcriptomics across cortical cell types.”
Source: Read Full Article