Research

An important question in neuroscience is how a brain, which is composed of neurons operating at a fast timescale of milliseconds, manages to retain information and generate behaviors with longer timescales of seconds and minutes. My current research at Portugues lab @TUM aims to understand neural circuit mechanisms underlying longer-timescale behaviors using the larval zebrafish (Danio rerio) as a model system. The larval zebrafish is one of the smallest vertebrate model organisms used in neuroscience with the body length of 4 mm and only about 100k neurons in their brain (as compared to 70M in mice and 90B in humans). In addition to their small size, the optical accessibility of their brains, as well as their genetic tractability make the zebrafish larvae amenable to various cutting edge experimental approaches, such as brain-wide calcium imaging, optogenetics, and EM-based circuit reconstructions. Currently, I am studying how the fish brain keeps track of its orientation in the environment by compining two-photon microscopy and immersive virtual reality setups. I believe that obtaining a detailed, mechanistic understanding of fish brains is a promissing avenue towards better understanding of our own minds.

During my PhD, I studied visual feature detection and their behavioral functions in the fruit fly Drosophila at Clark lab @Yale. My aspiration there was to bridge "why" flies need to see certain things, "what" computation their brains perform to see what they need to see, and "how" such computations are actually achieved by circuits of neurons, which together consitute canonical criteria of understanding in neuroscience. You can find my thesis here to check for yourself if I lived up to this lofty goal.

During my undergrad and master's, I studied perception in human subjects at Yotsumoto lab @UTokyo, where I was initiated into the world of research.

You can find my Google Scholar profile here.

Publications

Fish works

• Comming Soon
• Hopefully!

Fly works

• Tanaka, Zhou, Agrochao, Badwan, Au, Matos, & Clark (2023) Neural mechanisms to incorporate visual counterevidence in self-movement estimation. Curr. Biol.. [link] [preprint]
• Imagine you are sitting in a train stopped at a station. There is also another train stopped on the opposite track. As the other train slowly starts to move, for a split second, you confuse their movement for your own. As this (hopefully familiar) example illustrates, movements of yourself and other objects can sometimes result in similar visual consequences. So, how can the brain disambiguate these scenarios? This study demonstrated that the fruit flies interpret wide-field visual motion NOT as the consequence of their own movements (or rotation to be preceise), when there is any stationary visual patterns in the view. This makes a logical sense: when an observer rotates, everything in the view must move in synchrony. By contraposition, anything appearing to be not moving suggest that the observer is not rotating. We also explored how detection of stationary patterns is implemented in the visual system of the flies.
• Mano, Choi, Tanaka, Creamer, Matos, Shomar, Badwan, Clandinin, & Clark (2023) Long timescale anti-directional rotation in Drosophila optomotor behavior. eLife. [link]
• When presented with wide-field visual motion rotating about them, animals ranging from flies to humans just cannot help following it with their eyes or with the whole body, a reflex called the optomotor response. This behavior is believed to stabilize the gaze and posture, and its characteristics and mechansims have been under intensive study in various species for almost a century. But sometimes, they do just the opposite! This study identified when and how this paradoxical "anti-optomotor" response happens in the fruit flies. I contributed a part of the calcium imaging data for this study.
• Tanaka & Clark (2022) Neural mechanisms to exploit positional geometry for collision avoidance. Curr. Biol. [link] [preprint]
• Imagine you are driving towards an intersection without a traffic signal, and you see another car about to reach the intersection from the side. If you are a safe driver, you would slow and let them pass (I hope you do). Actually, it has been known that flies do exactly the same thing: when they see an object which is about to pass in front of them (which is destined to appear to be moving in the back-to-front direction from their perspective), they slow or stop walking. In this study, we identified a specific neuron type called LPLC1 is necessary and sufficient for this direction-selective slowing behavior. Using techniques such as neurotransmitter imaging and connectomic analyses, we deomsntrated that LPLC1 achieves its selectivity for small objects moving back-to-front by combining outputs of direction-selective and size-selective neurons. We also uncovered a downstream pathway that signals potential collisions to the central brain.
• Tanaka & Clark (2022) Identifying inputs to visual projection neurons in Drosophila lobula by analyzing connectomic data. eNeuro. [link] [preprint]
• The release of the "hemibrain" connectome by Janelia Reseach Campus in January 2020 was a big game-changer in the field of fly neuroscience. However, unfortunately for vision researchers like myself, the hemibrain connectome did not include early visual neuropils, and thus input neurons to the lobula (which was the only visual neuropil almost fully included in the dataset) were fragmented and unlabeled. This paper summarizes my effort during the COVID lockdown to categorize these neruon fragments into cell types with a hiearchical clustering approach based on connectivity and synapse distributions.
• Tanaka & Clark (2020) Object-Displacement-Sensitive Visual Neurons Drive Freezing in Drosophila. Curr. Biol. [link]
• The visual system of Drosophila is equipped with a suite of visual projection neurons (VPNs) that are supposedly tuned to behaviorally relevant entities (e. g. conspecifics, predators), whose activity can drive specific behavioral programs. This paper focused on one of these VPNs, called LC11. As of 2016, LC11 was known to have an exquisite tuning to small objects, but their behavioral functions and circuit mechanisms remained unknown. Here, we first showed that LC11 to be necessary for a short-timescale freezing behavior in flies, caused by moving small dots. We then constrained the circuit mechanism by which LC11 detects objects, by combining connectomic analyses as well as neurotransmitter imaging. Based on these data, we built a computational model that captures LC11's visual tuning well.
• Creamer, Mano, Tanaka, & Clark (2019) A flexible geometry for panoramic visual and optogenetic stimulation during behavior and physiology. Journal of Neuroscience Methods. [link]
• A good visual stimulation setup is a prerequisite for a good study of visual behaviors. The standard solution among the fly researchers has been to either use custom-build LED arrays, or to simply repurpose a PC monitor. This paper proposed a simple, low-cost solution to create a immersive virtual reality by combining a small optical projector with a couple of planar mirrors. The small footprint of this geometry allowed us to highly parallelize behavioral setups, which enabled me to efficiently compare various viusal stimuli as well as flies with various different genotypes. I performed a proof-of-principle experiment to show that one can perform optogenetic stimulations with the light from the projectors simultaneously with visual stimuli on this setup.

Human works

• Tanaka & Yotsumoto (2017) Passage of Time Judgments Is Relative to Temporal Expectation. Front. Psych. [link]
• There has been a growing interest among psychologists in the phenomenon of "passage of time judgements (POTJ)", which refers to our sense of how fast or slow time seems to be passing, as often expressed in the statements like "time flies when you are having fun". While some psychologists have drawn a strong distinction between POTJ and memory of duration (alluding to the distinctions between time-consciousness and objective time by phenomenologists), in this paper, I tried to argue that POTJ is basically made based on deviations between remembered and expected durations with several simple experiments.
• Tanaka & Yotsumoto (2016) Networks extending across dorsal and ventral visual pathways correlate with trajectory perception. J. Vis. [link]
• This paper examined neural activities of human subjects while they observed a visual illusion called "wriggling motion trajecotry illusion" with an fMRI.

Fly AND Human works (!)

• Agrochao*, Tanaka*, Salazar-Gatzimas & Clark (2020) Mechanism for analogous illusory motion perception in flies and humans. PNAS. (*: equal contributions) See also a WIRED coverage of this paper. [link]
• Peripheral drift illusions, where repeated patterns of luminance gradients give the illusory sense of movement, is one of the most striking kinds of visual illusions. While many theories have been proposed to explain this illusion, definitive, mechanistic explanations have been lacking. In this study, we found that flies also percieve motion in repeated gradation patterns just like ourseleves, and discovered that the illusion ultimately originates from the imbalanced contributions of bright and dark edges for motion detection. Our results demonstrate that this popular illusion is a manifestation of a strategy to efficiently detect motion by exploiting the asymmetric distribution of light and dark in our visual environment, convergently evolved in vertebrates and insects. For this study, I performed human psychophysics experiments to show that the mechanistic model of the illusion derived from the experiments in flies approximately applies to humans as well, among other things.