听
听
Embark On an Adventure in Computational Modeling
It can all be traced back to the physiology course during my high school years, when it was my first exposure to the concept of biological systems. As I was studying the hormone feedback loops that maintain the homeostasis of human body, a na茂ve thought was seeded in my mind: what if we could represent these biological components in a digital form and encode their interactions so that we can create an artificial living system? Such a thought hovered in my head until I began my undergraduate studies in biology. The curriculum, expanded from molecular biology to ecology, addressed to me the fact that biological systems are multiplex and remain incompletely understood. Concurrently, it came along with another fact that excited me even more: the principles underlying those diverse behaviours are generally applicable among different systems and across multiple scales.
I was eager to gain more experience on the mathematical modelling of biological systems. While reconstructing a complete biological system can be overwhelming, I could start simply. During my participation in the International Genetically Engineered Machine (IGEM) Competition, I explored the possibility of designing a plasmid so that its regulators control the expression of their reporters in a 鈥減aper-rock-scissors鈥� manner. Meanwhile, my final report of the ecology course reviewed how to use game theory to describe cooperative and competitive behaviours during evolution. While exchanging at T眉bingen University, I delved into the study of neural dynamics under the instruction of Dr. Jan Benda. I was fascinated by the elegance of using differential equations in describing the dynamics of neural activities, and such experience solidified my goal of pursuing an academic career in computational neuroscience.
I was subsequently thrilled to be accepted into the Integrated Program in Neuroscience at 黑料不打烊 and to work on a computational modelling project in the labs of Dr. Anmar Khadra and Dr. Reza Sharif-Naeini. My MSc studies focused on neuropathic pain. Our experimental analysis showed that the parvalbumin-expressing neurons, a gatekeeper for pain signalling in the spinal cord, exhibited reduced excitability in the nerve-injured mice. Combining both computational and experimental approaches, I investigated how a various combination of ion channels gives rise to neural excitability and eventually affects the pain signalling. Our results revealed the key contribution of the small-conductance calcium-activated potassium channels and therefore pinpointed it as therapeutic target in treating neuropathic pain. My model further unveiled the dynamics of intracellular calcium in modulating the firing behaviours of parvalbumin-expressing neurons. This work became my first first-author paper published in the . More importantly,听 I was struck by the great power of computational models in untangling the complex phenomena in terms of their basic neuronal units. I always remember the excitement that arose when my first neural circuit model worked. It reminded me of that na茂ve thought in high school and affirmed myself: I am making every small, but meaningful, step towards my goal.
听
Figure 1 I was presenting a poster in sfn 2023.
Spark of Open Science and Computational Tool Development
The electrophysiological recordings of parvalbumin-expressing neurons were collected in Sharif Lab, which were shared with me for analysis. However, my initial attempt at analyzing these current-clamp recordings went a bit rough. There are few online documentations that summarize the commonly used electrical properties. While I found this online source useful, it still lacks a quantitative description of these properties. In fact, as I delved deeper into the literature, I realized a significant lack of consensus regarding the definition and quantification of electrical properties. For instance, 鈥渟pike threshold鈥� can be easily identified by eye as the 鈥渆lbow鈥� of the action potential preceding spiking. But the challenge remains in quantifying such an 鈥渆lbow鈥� accurately.
Another problem related to electrophysiological analysis was also exposed to me during my collaboration with members in Sharif Lab. Neuroscienists, especially those of no coding experience, often rely on commercial software in data analysis. Those software packages have an easy-to-use graphical user interface (GUI), but the analysis pipelines operate on a file-by-file basis, which can become tedious when dealing with large datasets.
An initial solution to both of these two problems was to develop a code that includes the various quantifications of the electrical properties and package it into a graphical user interface. As the tool improved with increasing user feedback, I sought to take a step forward: expanding its benefits others outside our lab and with similar needs. This led to the development of my toolbox, ElecFeX.
ElecFeX, stands for Electrophysiological Feature eXtraction, is an open-source MATLAB-based toolbox that (1) has an intuitive GUI, (2) provides customizable measurements for a wide range of electrophysiological features, (3) processes large-size datasets effortlessly via batch analysis, and (4) yields formatted output for further analysis. ElecFeX is versatile in analyzing recordings in different data formats (ABF, Igor binary wave [IBW], HEKA [DAT], and data without borders [NWB]) and different waveforms with a set of customizable measurement methods. Its functionality was demonstrated by implementing it on datasets collected from different brain regions and species. For instance, using the large dataset from the Allen Institute, we showed that ElecFeX collected >300 feature values from >1,000 recordings within 鈭�2.5 h; these features were sufficient to cluster neuronal subtypes in a manner consistent with previous reports. ElecFeX is therefore presented as a user-friendly toolbox that advances the analysis of electrophysiological recordings to be more accessible, reproducible, and efficient. It is now published in the .
Upon the release of ElecFeX, we occasionally received inquiries expressing interest in using it. We are also actively promoting the toolbox. We asked to post ElecFeX in the webpage of NWB community, an open science organization of large userbase; we are also right now working with the MathWorks team include ElecFeX in their feature stories. Supported by the IPN GREAT travel award, I was able to present ElecFeX in workshops and conferences such as the Society for Neuroscience Meeting. During those events, I gathered much useful feedback in making the toolbox more accessible and powerful, and it also convinced me about the great value of ElecFeX to its potential users. Recently, the toolbox was awarded for the 2024 Neuro-Irv and Helga Cooper Open Science Prize.
Figure 2: I was presenting at SIAM 2024.
Embracing a Journey of Serendipity, Passion and Resilience
As I started writing, I realized that opportunity is as significant as passion, planning and effort in determining a career. Be brave to seize the idea and opportunities of interest. Next, allocate your time and effort judiciously, as these are the choices that lead to who we will be.
Throughout my journey, I am deeply grateful to meet with smart and like-minded individuals, particularly every member in Dr. Anmar Khadra and Dr. Reza Sharif-Naeini鈥檚 lab. Every discussion and brainstorming with them are indispensable for the realization of these two projects.
听
听
听
Figure 3: photos of the Khadra lab and the Sharif Lab.
听