New technique scans electrical ’brainscape’

Using hairlike microelectrodes and computer analysis, neurobiologists at Duke University Medical Center have demonstrated that they can see the detailed instant-to-instant electrical “brainscape” of neural activity across a living brain.

In their study on rats, they demonstrated that they could distinguish in unprecedented detail the patterns of brain activity — including fleeting changes in communication among brain structures — in awake animals, as they fall sleep and as they transition among different sleep stages.

The study is important, not only for its insight into the sleep process, but because neurobiologists have strong evidence that memory consolidation occurs during sleep, said the researchers.

More generally, they believe that their new analytical technique will enable unprecedented insights into function of both the healthy brain and those afflicted with neurological disease. Such insights could lead to new understanding and treatment if diseases including epilepsy, Alzheimer’s disease and schizophrenia, they said.

Led by neurobiologist Dr. Miguel Nicolelis, M.D., Ph.D., the researchers published their findings in the December 8, 2004, Journal of Neuroscience. Nicolelis is professor of neurobiology and co-director of Duke’s Center for Neuroengineering. Other co-authors were Damien Gervasoni, Shih-Chieh Lin, Sidarta Ribeiro, Ernesto Soares and Janaina Pantoja. The research was sponsored by the National Institutes of Health.

In their studies, Nicolelis and his colleagues implanted the microelectrodes, smaller than the diameter of a human hair, into regions of the brain responsible for a range of functions — including sensory processing, motor function and memory formation. They then recorded and analyzed the electrical signals from the rats as the animals went through several days of sleep-wake cycling. Their analysis could detect activity patterns that marked waking, deep “slow wave” sleep and so-called “rapid-eye movement” sleep.

Importantly, said Nicolelis, their analysis could distinguish the fleeting changes in the brain as the animals transitioned from one sleep state from the other.

“We can actually predict such changes, because at that moment, these different structures fire together for a few hundred milliseconds to create a synchronous pattern of firing that is a signature of the change from the previous state to the next,” said Nicolelis. A millisecond is one thousandth of a second. “It’s almost like two computers exchanging information over a modem, and they get synchronized in the process,” he said.

“Our analysis revealed significant functional insights into sleep,” said Nicolelis. “For example, we found that there are only a few physiologically possible transitions from state to state — just as in chemistry there are only certain chemical reactions that are possible.” For example, he said, the data distinguished the elusive transition called “intermediate sleep” between slow wave sleep and rapid-eye-movement sleep.

Importantly, said Nicolelis, the transitions they observed were the same from one animal to another, “suggesting that we have arrived at a major basic principle of how the brain actually operates.”

The technology and analysis the researchers used is an extension of that used to enable monkeys to control a robot arm using only their brain signals, which Nicolelis and his colleagues reported in 2003. “Now, however, we are recording broader brain signals — hundreds, perhaps thousands,” said Nicolelis. “By filtering and analyzing them, we can actually measure the global dynamic activity that tells us what behavioral states the animals are going through. “Such capability is broadly important because it is the first physiological measurement that can reveal the global behavior of the brain, including the broad coordination of so many areas.”

In contrast, said Nicolelis, magnetic resonance imaging and positron emission tomography — the most widely used brain-scanning techniques — can give only limited time-resolution of brain activity. Also, they give only indirect indications of brain activity by measuring blood flow as an indicator of activity.

According to Nicolelis, their detailed studies of brain activity — including a previous study reported in the June 25, 2004, issue of Science, reveal that the brain is not the passive, unchanging computer postulated by most current theory. Rather, he said, it is a dynamic, constantly adapting organ. In the Science paper, the researchers reported that the brain response of a rat to tactile stimulus to its whiskers changed according to whether the animal was actively performing a task or passively receiving input.

“Our studies suggest that perception is not just a process of analyzing an incoming signal, which is what most textbooks teach and most scientists believe,” said Nicolelis. “Rather, perception depends on the internal state of the brain at that given moment of time, and what behavior the animal is using to sample the environment. With this new technique we can monitor such states as attention and expectation and how they modulate how the animal processes that incoming information.”

According to Nicolelis, the new results “support a global theory of brain function that holds that all these processes are extremely dynamic. And now with this analytic technique we can measure these dynamics. It gives us a new language of how to describe continuous brain function. “One of the Holy Grails of neurobiology has been the neural ’code’ by which the brain processes information. Now we can say that there is no such thing as a single neural code, because the code is continuously changing according to the internal state of the brain, and according to the strategy the animal selects to search the environment.”

Also, said Nicolelis, such analyses will influence neurobiology to advance beyond the current theory that the single neuron is the basic computational unit of the brain. “A single neuron is too noisy to act as a reliable unit of neuronal function,” he said. “But an ensemble of neurons resolves that noise and makes neuronal output stable.”

The technique will be important for studying both the different states of the healthy brain and the pathology of neurological disorders, said Nicolelis. “For example, we can analyze brain activity in sleep-deprived animals, those in the process of learning or transgenic animals with changes in their brain circuitry,” he said. “We can quantify the differences in the global dynamics of such brains from those of normal animals.

“Also, there exist many mouse models of neurological disorders, and we can use this technique to explore how the brain functions in those models. For example, we can understand the global dynamic structure of the brain in Parkinson’s disease, Alzheimer’s disease or schizophrenia. Knowing the details of what goes wrong in such disorders is a critical step in treating them,” said Nicolelis.