MYCOLab Ai Logo - gold and army green

Micro-CT and deep learning: Modern techniques and applications in insect morphology and neuroscience

Summary

Modern scanning technology called micro-CT can create detailed 3D pictures of tiny insects and their brains without damaging them. Artificial intelligence using deep learning can automatically analyze these massive image files much faster than humans could. Scientists are combining these two technologies to map insect brains and sensory systems in unprecedented detail, potentially revealing how insects sense and process information from their environment.

Background

Advances in micro-computed tomography (µCT) and artificial intelligence have revolutionized biological imaging and data analysis. µCT provides non-destructive, high-resolution 3D imaging of organisms, while deep learning enables automated analysis of large image datasets. These complementary technologies offer significant potential for studying insect anatomy and neural structures.

Objective

This review examines the applications of µCT and deep learning in insect morphology and neuroscience, with focus on how these technologies can be combined to investigate insect sensory systems, from external receptor structures through neuronal pathways to brain architecture.

Results

The review documents multiple successful applications of µCT in insect neuroscience, including brain imaging of Drosophila, honey bees, and other insects at resolutions enabling visualization of neuropiles and large neuronal structures. Deep learning networks have successfully automated brain segmentation in insects with high accuracy (Dice coefficients 0.9-0.99) and reduced analysis time from hours to minutes.

Conclusion

Combining state-of-the-art µCT with AI-supported 3D analysis provides researchers with powerful tools for non-invasive, rapid imaging and analysis of insect morphology and neural systems. Future development of standardized training datasets, neuron-specific contrast enhancers, and advanced network architectures will enable complete mapping of insect sensory systems and neural circuits.
  • Published in:Frontiers in Insect Science,
  • Study Type:Review,
  • Source: PMID: 38469492, DOI: 10.3389/finsc.2023.1016277
Scroll to Top