Input-output relationship in social communications characterized by spike train analysis

Area of Science:

• Human communication dynamics
• Computational social science
• Information theory

Background:

• Human communication occurs across diverse channels, including short messages, phone calls, and emails.
• Understanding the temporal dynamics of these communication patterns is crucial for analyzing social interactions.
• Neuronal spike train analysis offers novel techniques for characterizing temporal fluctuations in event sequences.

Purpose of the Study:

• To investigate the dynamical properties of human communication across different channels.
• To characterize temporal fluctuations in interevent times using methods from neuronal spike train analysis.
• To identify factors influencing correlations in communication activity.

Main Methods:

• Analysis of local variation (LV) for incoming and outgoing event sequences.
• Examination of response-time distributions for different communication channels.
• Development of a point process model to simulate communication dynamics.

Main Results:

• Local variation (LV) values for incoming and outgoing event sequences are positively correlated for short messages but uncorrelated for phone calls and emails.
• Response-time scales and amplitudes differ significantly across short messages, phone calls, and emails.
• Numerical simulations indicate that rapid responses and refractory periods are key to positive LV correlations.

Conclusions:

• The temporal dynamics of human communication are channel-dependent.
• Response patterns, specifically speed and user refractory effects, significantly influence communication correlations.
• Point process models can effectively capture and explain observed communication dynamics.