Translation. Region: Russian Federation –
Source: Peter the Great St. Petersburg Polytechnic University –
An important disclaimer is at the bottom of this article.
Polytechnic University researchers have developed a new digital methodology for assessing the shape of synapses in brain neurons. The developed metrics allow for not just estimating size, but also describing their complex three-dimensional shape with high mathematical precision. This discovery will help researchers more quickly and accurately assess the effectiveness of substances that could become the basis for future drug treatments for various diseases, such as Alzheimer's disease. The results of the study were published in the prestigious scientific journal Bioinformatics.
In the most general sense, dendritic spines of neurons in the brain can be considered structures responsible for memory and learning in humans. These membrane projections on neurons are a component of the synapse and receive signals from other neurons.
In developmental brain diseases or severe neurodegenerative diseases, changes in the shape of spines are observed. Synapses change shape, degenerate, and connections between neurons deteriorate. One of the factors influencing the negative change in spine shape, and consequently their functioning, is the accumulation of beta-amyloid oligomers (so-called amyloid plaques, an altered form of the beta-amyloid protein), which begins long before the first clinical symptoms of Alzheimer's disease appear.
Researchers have traditionally classified spines into several types based on their shape (mushroom-shaped, thin, stumpy, etc.) using visual or semiautomated classification, or described them using simple numerical parameters (length, volume, head width, angles). Scientists at the St. Petersburg Polytechnic University have developed new numerical metrics that describe spine shape much more accurately.
We used the mathematical apparatus of spherical harmonics and Zernike moments. These methods have proven themselves in engineering for analyzing complex shapes. The novelty of our work lies in the fact that we are the first to apply three-dimensional mathematical shape descriptors to microscopic images of spines. Harmonics allow us to decompose a complex three-dimensional object into a sum of basic three-dimensional shapes with specific coefficients, and even reassemble them back into this shape with high accuracy using these coefficients. Zernike moments describe the object's shadow in different projections, which also very accurately characterizes its structure. Our proposed algorithm allows us to capture the highly complex, multifaceted shape of spines as if using a scanner," noted Daria Smirnova, a programmer at the Laboratory of Biomedical Image and Data Analysis at the Institute of Biomedical Systems and Biotechnology at SPbPU.
To test the effectiveness of the new tool, the scientists compared the spine shapes of healthy neurons and neurons in a brain model of Alzheimer's disease. Previous methods for assessing spine shape only showed a decrease in spine size during the disease. The new method, however, additionally revealed statistically significant shape redistributions across five different clusters. For example, amyloid toxicity increased the prevalence of elongated and atypical spines, which are difficult to classify traditionally but play an important role in understanding the mechanisms of neurodegeneration.
The value of this new method lies in its ability to more accurately analyze the response of damaged neuronal tissue to various chemicals, including experimental therapies for neurodegenerative diseases. This means we have a tool that allows us to see previously inaccessible subtle changes in spine shape. This is important in the search for a cure for Alzheimer's disease: our tool will allow researchers to more fully and accurately record the restoration of the shape of damaged spines under the influence of the test substance. Furthermore, in the future, this technology will enable the creation of a realistic 3D model of neurons, which can be used to train neural networks and virtually test medical hypotheses, saving time and money on complex biological experiments, noted Ekaterina Pchitskaya, Head of the Laboratory of Biomedical Image and Data Analysis at the Institute of Biomedical Systems and Biotechnology at SPbPU.
The research team's immediate plans include refining the method for characterizing very thin and elongated spines and integrating the development into the open-source software tool SpineTool, making it accessible to neuroscientists worldwide.
The study was supported by a grant from the Ministry of Science and Higher Education of the Russian Federation (FSEG-2024-0025) and a postgraduate research fellowship from the Idea Center for Advanced Interdisciplinary Research.
Please note: This information is raw content obtained directly from the source. It represents an accurate account of the source's assertions and does not necessarily reflect the position of MIL-OSI or its clients.