Research

Imagine traveling deep down into the brain by steering an electrode the size of a human hair. Your goal: hit a brain region just as small.

That’s what we do.

Because of the minute size of the locus coeruleus and its location deep in the brain, our experiments are like carefully charting a ship’s course in the dark, rock-filled Baltic Sea. We are expert navigators, using electrophysiological landmarks to find our way deep into the brainstem to this precise destination.

Find out more about our locus coeruleus work.

We train rats to run when they see one visual stimulus, and to sit still when they see another visual stimulus. We observed that rats sometimes mistakenly start running, and then realize their mistake and stop running. We are studying brain activity during such changes-of-mind. We have licenced this behavioral technology platform to 3D Neuro and welcome opportunities to collaborate and disseminate this method. (See image 1)

When we catch ourselves from making a mistake, we usually feel badly in a noticeably visceral way: our heart rate changes, we flush, and perhaps make a facial expression (e.g., a grimace). We are investigating this mind-body connection by recording brain activity, electrocardiogram (ECG), and facial expressions when rats catch themselves in a mistake. (See image 2)

We also collaborate with machine learning labs to perform real-time decoding of running speed from brain activity, which has applications for brain-machine interfaces. (See image 3)

Using a head-fixed treadmill setup, we introduce rats to visual and audio cues, each linked to specific behaviors. Initially, rats learn a rule based on visual cues—run to one cue, do not run to the other cue — all while ignoring sounds. Then, we switch the game, requiring them to learn to use auditory cues and disregard visual ones. We apply computational tools like Principal Component Analysis (PCA) and Hierarchical Hidden Markov Models to uncover the hidden behavioral strategies that rats use to solve this task. 

We are testing the hypothesis that neuromodulatory systems promote the reorganization of cortical neuronal networks that enable the rat to explore the auditory world after becoming biased toward using the visual world to obtain rewards. We use pupillometry to record pupil size as a non-invasive measure of neuromodulatory system activity. Using 32-electrode, whole-cortex EEG recordings of field potentials, signal processing methods, and information theory, we study how interactions between brain regions change as rats discover new rules requiring switching to a sensory modality irrelevant to their earlier rule. We are also testing how computational models of learning - that form the “engine under the hood” of AI - can match the natural, biological intelligence of rats in this task.

We created a new rat behavior paradigm showcasing how the subjective perception of brightness illusions causes constriction of the eye’s pupil. We used 32-electrode EEG to record primary visual cortex neuronal activity changes before pupil constriction.

We collaborated in subsequent work led by Prof. Masataka Watanabe and his lab at the University of Tokyo to replicate the effect of brightness illusions on pupil size in mice. We analyzed primary visual cortex single neuron activity to show those single visual cortex neurons, which are tuned to real line orientations, are also responsive to identical orientations of illusory lines. Using optogenetic inhibition, we demonstrate that this response depends on top-down input from higher visual regions.