High resolution behavioral and neural activity representation using a geometrical approach

Available tools for recording neuronal activity are limited and reductive due to massive data arising from high-frequency measurements. We have developed a method that utilizes variance within the physiological activity and includes all data points per measurement. Data is expressed geometrically in a physiologically meaningful manner, to represent a precise and detailed view of the recorded neural activity. The recorded raw data from any pair of electrodes was plotted and following a covariance calculation, an eigenvalues and chi-square distribution were used to define the ellipse which bounds 95% of the raw data. We validated our method by correlating specific behavioral observation and physiological activity with behavioral tasks that require similar body movement but potentially involve significantly different neuronal activity. Graphical representation of telemetrically recorded data generates a scatter plot with distinct elliptic geometrical properties that clearly and significantly correlated with behavior. Our reproducible approach improves on existing methods by allowing a dynamic, accurate and comprehensive correlate using an intuitive output. Long-term, it may serve as the basis for advanced machine learning applications and animal-based artificial intelligence models aimed at predicting or characterizing behavior.