Dr HaDi MaBouDi
Bees effortlessly percept and navigate in a complex, noisy, and rapidly changing environment. Recent studies reported remarkable cognitive capabilities of honeybee which is beyond the sensory processing and associate learning. For example, they can learn visual concepts which are independent from visual features. Yet, the basic principles of visual processing and concept learning in such miniature brains remain unclear. How the concepts are formed in these small brains? What mathematical principles are implemented by neural circuits? Interestingly, honeybees might employ different strategies from primates to understand visual concepts. I am most interested in the architecture of visual cognition and plastic dynamic circuit of concept formation in the bee brain. More specifically, I would like to find responses to my main questions: how does honeybee brain build an abstract model of the world given by the sparse and noisy data she observes through scenes? How do neural populations’ dynamics beyond a single neuron represent input concepts? What mathematical principles are implemented by local micro-circuitries and global network topology of the miniature bee brain?
-Cognitive and neurobiological basis of visual perception and concept learning
-Computational architecture of sensory system and plastic circuit of concept learning in insects and primates
-Statistical machine learning
Hobbies: Photography, climbing, travelling and reading
To tackle my questions, I have been working on a HFSP program entitled: "A neural circuit approach to cognition and its limits in microbrains", a joint project between Chittka’s lab in QMUL and two labs in the University of Washington, Seattle and Research Center on Animal Cognition, Toulouse. Our aim is to understand cognitive processing in the honeybee brain from behavioral, neurophysiological and computational approaches. In this project, we use behavioral and multielectrode recordings of honeybees in an active task to unravel the neural basis underpinning concept learning. In addition, We construct spiking neural network models of sensory and mushroom body of the honeybee brain that enrich us to investigate possible hypothesis on minimal neural architectures for cognition.
In my PhD project I worked on computational vision. In particular, I studied statistical models of early visual cortex based on the efficient coding. My focus was on the role of spatial phase structures of natural scenes in the object recognition by visual system, which turned out to be the topics of several manuscripts we have submitted or are currently preparing. During my PhD I gained valuable experiences in Amari’s mathematical neuroscience lab at RIKEN Brain Science Institute as well as in Hyvarinen's lab in neuroinformatics at the University of Helsinki.
Educational and employment background:2015–Present: Postdoctoral Researcher (HFSP funded), School of Biological and Chemical Sciences, Queen Mary University of London: "A neural circuit approach to cognition and its limits in microbrains".
2010–2015: PhD in Cognitive Neurosciences, School of Cognitive Sciences, Institute for Research in Fundamental Sciences (IPM), Tehran, Iran (Thesis title: Hierarchical generative models of spatial phase processing in the visual cortex; supervised by Prof. Hamid Soltanian-Zadeh & Prof. Shun-ichi Amari).
2008–2010: Researcher, Advanced Information & Communication Technology Research Center (AICTC), Sharif University of Technology, Tehran, Iran.
2007–2009: Instructor, Tehran Azad University, Tehran, Iran (Courses taught: Probability & Statistics, Linear Algebra, Discrete Mathematics, Discrete Date Structure, Calculus I & II and Differential equations).
2004–2007: MSc in Mathematics (Quantum Topology), School of Mathematics, Statistics and Computer Sciences, University of Tehran, Iran; supervised by Prof. Ahmad Shafiei Deh Abad.
2000–2004: BSc in Mathematics, Department of Mathematics, Iran University of Science and Technology (IUST), Iran.
MaBouDi H, Shimazaki H, Amari S,
Soltanian-Zadeh H. Representation of higher-order statistical
structures in natural scenes via spatial phase distributions.
Vision Res 2015; doi: 10.1016/j.visres.2015.06.009 (In
MaBouDi H, Shimazaki H, Soltanian-Zadeh H, Amari S. Bimodal distributions of local phase variables in natural images. The VSS Annual Meeting, St. Pete Beach, Florida, 18 May 2014.
MaBouDi H, Shimazaki H, Abouzari M, Amari S, Soltanian-Zadeh H. Statistical inference for directed phase coupling in neural oscillators. COSYNE Abstracts 2014; Salt Lake City, USA: 109-110.
MaBouDi H, Abouzari M. The role of higher-order statistical dependency of natural images in visual perception. Perception 2012; 41(S): 257.