Ahmat Farhat
Dioscuri Centre for Topological Data Analysis
Institute of Mathematics, Polish Academy of Sciences
wygłosi prezentację zatytułowaną:
Applications of Topological Data Analysis in Neuroscience
In the field of neuroscience, the shape of neurons determines their function, and therefore is a crucial concept for neuroscientists to understand. Traditional methods of characterizing morphology often rely on the so-called morpho-metrics, involving, e.g., dendritic length and complexity, and branching patterns, which may not fully capture the complexity of neuronal architecture. Topological Data Analysis (TDA) is a powerful tool for studying the shape and structure of data, and it outperforms the traditional methods in revealing the complex geometric and topological patterns of neurons. This is particularly useful, since TDA allows for distinguishing neuronal morphology across cell classes, brain regions, and animal species, and allows neuroscientists to understand the function of different neurons.
In this talk, a topological Sholl descriptor will be presented (see the reference). This descriptor is a rule that associates to a given neuron a function, whose independent variable is its distance from the soma. The neuron is seen as a tree embedded in the 3D space. The soma is marked as the root of the embedded tree, reconstructed from the sequence of microscopic images. These Sholl descriptors capture how a particular morphological feature changes as distance from the soma increases. We will also explore standard morphometric features, like dendritic length, volume, average number of branches, and the like, to determine which neuronal features of pyramidal cells change and which are stable during development of the monkey visual cortex. Notably, the presented methods complement and advance many of the existing techniques currently used to study neuronal morphology.
Reference:
Khalil R, Kallel S, Farhat A, Dlotko P. Topological Sholl descriptors for neuronal clustering and classification. PLoS Comput Biol. 2022 Jun 22;18(6):e1010229. doi: 10.1371/journal.pcbi.1010229. PMID: 35731804; PMCID: PMC9255741.
Serdecznie zapraszamy !
