I studied Automatic Control and System's Engineering for both my bachelors and masters degrees. In an attempt to get biological aspirations for engineering applications, I went to the University of Sheffield for my Ph.D, studying information processing in the peripheral visual system of flies. After which, I got fascinated about how seemingly trivial biological systems far outperform man-made machines by its adaptive capabilities. Being eager to understand more and seeing myself as a computational life scientist, I am now working as a 2020 Science Research Fellow, modelling vision systems, under supervision of Prof. Alan Johnston in UCL, London.
My primary research interest is to study how sensory information is reliably and economically encoded by physiological systems. How do sensory neurons sample and process noisy input information reliably through unreliable stochastic biochemical reactions? How do they adapt continuously to better encode new stimuli? How does their structure and interconnections impact on the information encoding capabilities?
I address these questions by constructing and analysing both deterministic and stochastic computational models, which can perform as virtual neurons and replicate their input-output relationships. Depending on the particular goal, a model can be constructed with different levels of complexities, ranging from black-box mathematical descriptions to detailed compartmental structures.
The core of my PhD was to investigate the light information encoding mechanisms of fly photoreceptors. I constructed biophysically realistic fly photoreceptor models, which accurately simulate the encoding of visual information. By comparing stochastic simulations with single photoreceptor recordings, I showed how a photoreceptor uses a population of stochastically and adaptively operating units (30,000 microvilli) to sample photon information and robustly encode natural light changes.
Crucially, the population coding dynamics reduces to elementary sampling rules, which could be general in sensory neuroscience. For example, I am currently working on how spatial-temperal stochastic adaptive sampling forsters motion processing in vision systems.
Furthermore, I am also interested in investigating the role of stochasticity in biological systems. Stochastic reactions within well-defined structural compartments in biological systems could be beneficial for them to generate reliable responses. Rather than treating stochasticity simply as noise, its merits and costs become clear when modelled and analysed within the signalling processes, for example, with Monte Carlo simulation techniques.
Dr. Samuel Solomon, UCL, Population encoding of motion in Visual area MT
Ben Calderhead, Imperial College London, Biological data Inference
Dr. Guy S. Bewick, University of Aberdeen, Biophysical Modelling of musle sensory endings
Dr. Robert W. Banks, University of Durham, Biophysical Modelling of musle sensory endings