Machine learning-based classification of binary dynamic fluorescence signals reveals muscle stem cell fate transitions in response to pro-regenerative niche factors