MUPET

MUPET identifies and characterizes ultrasonic vocalizations (USVs) in rodents using automated signal processing and unsupervised machine-learning to quantify syllable types and repertoire structure.

Key Features:

• Automated Signal Processing: Employs automated signal processing techniques to detect and extract ultrasonic vocalizations (USVs) from audio recordings and handle large datasets.
• Unsupervised Learning: Utilizes an unsupervised learning methodology and machine learning techniques to identify and categorize syllable types without prior labeling.
• Syllable Analysis and Comparison: Measures, learns, and compares different syllable types to characterize repertoire complexity across behavioral contexts.
• Time-Stamped Syllable Events: Automatically timestamps syllable events to enable precise temporal analysis of vocalization patterns.
• High-throughput Analysis: Supports high-throughput analyses of USV datasets.
• Data-driven Repertoire Characterization: Implements a data-driven approach to characterize USV repertoire complexity.
• Open-access MATLAB Implementation: Distributed as open-access MATLAB software.

Scientific Applications:

• Behavioral Neuroscience: Enables analysis of rodent communication to study vocalization patterns within behavioral neuroscience.
• Mouse Genetic Reference Panel Analysis: Facilitates analysis of USVs from a large mouse genetic reference panel to investigate genetic influences on vocal repertoire.
• Cross-condition Dataset Comparison: Allows comparison of USVs across open-source datasets recorded under different social conditions to examine context-dependent vocal behavior.
• Correlation with Behavior and Genetics: Supports investigation of correlations between vocalizations, specific behavioral states, and genetic variations.
• Cross-species Applicability: Adaptable for use with species beyond rodents.

Methodology:

Performs automated signal processing and unsupervised machine-learning analyses to detect, timestamp, measure, learn, and compare syllable types from ultrasonic audio recordings.