A convolutional neural network model detecting lasting behavioral changes in mice with kanamycin-induced unilateral...

Area of Science:

• Neuroscience
• Ototoxicology
• Animal Behavior

Background:

• Aminoglycoside ototoxicity can cause acute inner ear dysfunction.
• Neural compensation typically resolves physical abnormalities like nystagmus within days.
• Long-term behavioral changes and compensatory mechanisms remain less understood.

Purpose of the Study:

• To investigate long-term exploratory behaviors in mice following unilateral kanamycin-induced ototoxicity.
• To evaluate the utility of deep learning for analyzing behavioral compensation over an extended period.
• To assess the potential of this method for therapeutic intervention efficacy.

Main Methods:

• Unilateral inner ear injection of kanamycin in freely moving mice.
• Home cage behavioral recording and deep learning-based object detection (excluding tail).
• Convolutional neural network classification for result validation.
• Analysis of movement patterns, turning frequency, and specific physical signs (tail suspension/twisting).

Main Results:

• Kanamycin-injected mice showed increased total distance moved by postoperative day 3.
• Injured mice exhibited persistent turning towards the healthy side for up to 17 days.
• A characteristic rapid rotation with dorsal bending was observed for up to 14 days.

Conclusions:

• Deep learning analysis provides a robust method for evaluating long-term neuronal compensatory processes after ototoxicity.
• The observed persistent behavioral alterations highlight the complex adaptation following inner ear injury.
• This analytical strategy is valuable for assessing the effectiveness of treatments for ototoxicity.