Research Projects

Computational processes that contribute to complex animal behavior

Loren Frank Lab | UCSF | Systems/computational neuroscience

Mentored by David Kastner, I helped to develop a novel experimental paradigm for studying individual animal variability in the learning of a spatial alternation task. A unique feature of this task was the interleaving of multiple task variants, which allowed us to explore the dynamic and continual learning process. We also sought to develop a computational model of animals’ learning, which demands a large amount of data. To achieve this, we built an automated system for collecting high-throughput behavioral data for a long period of time (> 100 days). We compared and fitted the animals’ behavior to that of the model as a way of generating and testing hypotheses for how the animals might be cognitively achieving the task.

In this accelerated video, one of the animals visited several reward wells on the automated track. When the animal finished the task, the track door automatically opened. After the animal went back to his box, next animal was allowed to come out of his box and carry out the tasks.

Molecular underpinnings of feature selectivity in sensory and spatial representations

Richard Huganir Lab | Johns Hopkins University | Molecular & cellular neuroscience

I worked in the Huganir lab for 2 years during my undergraduate studies at Johns Hopkins, supported by the Provost’s Undergraduate Research Award. I studied the molecular underpinnings of neural feature selectivity. We hypothesized that feature selectivity may be causally influenced by calcium permeable AMPA receptors (CP-AMPARs). By introducing these receptors into feature selective neurons (e.g. hippocampal place cells), we found that the neurons became significantly less selective to their inputs. In contrast, removing CP-AMPARs from neurons increased their feature selectivity. These results suggest that CP-AMPARs are necessary and sufficient for maintaining low feature selectivity and regulating their expression might play a broader role in sensory representation.

Model and measure the connections between cardiomyocytes and neurons with a precision engineered system

Yasuhiko Jimbo Lab | The University of Tokyo | Bioengineering

During one summer, I joined the Jimbo lab as an Amgen Scholar. Our goal was to understand the functional connections between cardiac cells and neurons. I built a precision engineered system for measuring the electrical activities of the co-cultured cardiac and neuronal cells with microelectrode arrays. I found that the inter-beat interval of cardiac cells was significantly shortened in response to stimulating and increasing the spiking rate of the co-cultured neurons. This revealed that the neurons had innervated the cardiac cells, establishing functional connections between these two cell types. At the end of the project, I received Amgen’s “best presentation award” of the year. In the following year, I presented a poster at the WCBR conference and co-authored a paper based on this work.

In this video, the electrical signal trace (in white) was plotted and overlaid on top of a video of contracting heart cells.