Info

Charting super-colorful brain wiring using an AI's super-human eye

We are delighted to announce the recent research findings of the Professor Takeshi Imai‘s research group, affiliated with the Medical and Environmental Unit/Neuroscience Group within the Integrated Initiative for Designing Future Society at Kyushu University.
 
Our brain functions are realized by complex computations performed by a vast number of neurons. To comprehend the information processing mechanism in the brain, it is essential to decipher the connections among these neurons. However, given the complex entangling of neuronal wirings, it has traditionally been challenging to analyze multiple neuron connections simultaneously. In this study, we have pioneered a new technique to automatically analyze the neural circuit connections using multicolor labeling.

The team, comprising Professor Takeshi Imai, Assistant Professor Markus Lewi (during the research), Assistant Professor Satoshi Fujimoto, and graduate student Toshikazu Baba, initially succeeded in multi-color labeling of neurons using a blend of seven different fluorescent proteins. Previously, labeling with around three types of fluorescent proteins was the standard, but using seven types of dyes exponentially increased the potential color combinations. However, discerning seven different color combinations using the human eye is highly challenging. To address this, we developed a program to identify colors, extending it to cover seven primary colors. We also created a unique program, dCrawler, that can sort multidimensional data. Additionally, using dCrawler, we developed the QDyeFinder program, which classifies neurite color information and automatically identifies neurites with matching color combinations. Using these tools, we have successfully automated the analysis of numerous neurite connections based solely on color information. As such, we’ve dramatically improved the analysis of neural circuit wirings through ‘super-multicolor’ labeling and automatic primary color representation analysis.

This achievement is anticipated to contribute to advancing connectomics – the field of research devoted to mapping the ‘wiring’ links of neural circuits. These research findings were published in the renowned British scientific journal, Nature Communications, on June 25th, 2024 (Japanese time at 6 pm).

Related Links