Sample skewness as a statistical measurement of neuronal tuning sharpness

Area of Science:

• Neuroscience
• Computational Neuroscience
• Statistical Analysis

Background:

• Quantifying neural tuning curve sharpness is challenging due to complex and diverse response functions.
• Existing parametric methods struggle with non-uniform tuning curves, limiting broad applicability.

Purpose of the Study:

• To propose and validate sample skewness as a robust, nonparametric measure for neural tuning curve sharpness.
• To demonstrate skewness's effectiveness across various tuning curve complexities and its invariance to mean and variance.

Main Methods:

• Statistical measurement of sample skewness applied to mean firing rates of tuning curves.
• Comparison of skewness with other nonparametric measures of tuning curve sharpness and selectivity.
• Formal definition of sharpness using shape-based information gain and correlation with skewness.

Main Results:

• Skewness provides a robust, nonparametric measure of tuning curve sharpness, applicable to diverse and complex tuning functions.
• It is invariant to the mean and variance of the tuning curve, allowing broad application.
• Skewness effectively captures sharpness in multimodal tuning curves and correlates with information gain measures.

Conclusions:

• Sample skewness is a versatile and powerful tool for quantifying neural tuning sharpness, especially in complex scenarios.
• Its model-independent nature enhances statistical power by enabling application to all recorded neurons.
• Skewness offers a superior method for analyzing sharpening phenomena driven by recurrent neural interactions.