The University of Massachusetts Amherst
University of Massachusetts Amherst

Search Google Appliance


Xu Leads Circa $1-million NSF Study of Linkage Between Neural Activity and Complex Animal Behaviors

Professor Guangyu Xu

Guangyu Xu

The National Science Foundation (NSF) has awarded a four-year grant of $953,300 to a research team led by Professor Guangyu Xu of the Electrical and Computer Engineering Department at the University of Massachusetts Amherst to study the neural activity underlying complex animal behaviors. The research being funded by the NSF will allow scientists to find precise connections within and between different regions of the brain and trace the origin of animal behaviors down to cellular levels.

This award will contribute to NSF’s significant investments in the Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative.

Xu is working with Geng-Lin Li of the Biology Department and David Moorman of Psychological & Brain Sciences Psychological & Brain Sciencesthe Psychological and Brain Sciences Department at UMass Amherst, as well as Professor Ethan Meyers of Hampshire College.

The research objective of this proposal is to combine high-precision optoelectronic neural probes (probes that use solid state devices to source and detect light) with real-time neural decoding to feedback “optogenetic control over animal behavior,” meaning the use of light to control cells in living tissues that have been genetically modified. The research team says that “Such closed-loop neural interface will establish a generalizable technology platform to study complex animal behaviors using optogenetic tools and real-time learning.”

The researchers add that the proposed work will open up ample research opportunities to connect hardware engineering, cognitive neuroscience, and data science. “These high-density probes,” their proposal states, “will allow researchers to characterize and control neural activity underlying animal behaviors with high precision, and they will be miniaturized to reduce tissue damage for deep-brain access.”

As Xu explains, "Unlike current techniques, which largely permit bulk, synchronous activation or inhibition, our new techniques will allow us to use realistic stimulation/inhibition parameters, based on recorded activity, to probe the neural correlate of complex behavior with a previously unrealized fidelity. Once realized, this technology will be widely shared with and highly valuable to neuroscience investigators across a range of research domains."

According to the NSF proposal, neural-interfaces allow researchers to sense and control the activity of diverse neural circuits, permitting quantitative studies of the link between brain activity and animal behavior. To date, high-density microelectrodes have been developed to conduct electrical stimulation and recordings of neural activity with high spatiotemporal resolution.

“However,” notes the proposal, “even these state-of-the-art microelectrodes have limited capacity to modulate cell activities in a bidirectional manner (i.e. stimulating and silencing), or to record multiple signals from the same cell simultaneously.”

As Xu and company observe, over the past decade, optical neurointerfacing — a technique that uses light to control and monitor neurocircuitry — has emerged as a viable alternative approach without the limitations of microelectrode technology.

“While extensive research has focused on its hardware development,” say the researchers, “to date optical neurointerfacing has not made its full impact in tracing complex animal behaviors down to ensembles of individual neurons and moreover establishing a real-time closed-loop brain-interface (i.e., recording and controlling neural activity).”

The research team says that neural decoding will employ effective algorithms that allow real-time learning and training among experimental animals while they are performing complex decision-making tasks.

“This new technology will allow scientists to study the heterogeneity of cell activity in a dense cell population and to find precise connections both within and between different regions of the brain,” say the researchers. “Our real-time decoding of neural circuits underlying animal behaviors by iteratively trained algorithms will allow us to deliver feedback to the animal with optogenetic stimulus pattern to control its behavior and form the closed-loop, brain-computer-interface.”

The researchers conclude that “Our technology will grant access to cell populations deep in the brain and offer multiplexed cellular recording for high-content analysis. The accessibility of such data would enable quantitative studies in the behavioral neuroscience beyond what is feasible today.” (September 2018)