Even the most recent advances in Systems Biology and Integrative Physiology (Joyner, 2011) continue to focus on the “vertical integration” from sub-cellular and cellular level to tissues and single organs, and on the the signaling and feedback mechanisms across space and time scales that facilitate this “vertical integration” (Fig A). Surprisingly, there is a clear absence of knowledge and research effort in the direction of “horizontal integration” of organ interactions (Fig B). This horizontal integration is essential to produce health, and understanding the underlying mechanisms is crucial to fill in the gap between Systems Biology/Integrative Physiology and Epidemiology. A major focus of our research is this new direction of horizontal integration, Network Physiology (NatureCommunications-2012).

Identifying and quantifying organ interactions is a major challenge due to several levels of complexity inherent to the dynamics of organ systems.

Level 1: Individual systems: each organ is a multi-component system with its own regulatory mechanisms, leading to complex emerging behavior characterized by noisy, intermittent, scale-invariant and nonlinear output signals; organ systems have different types of output dynamics oscillatory, stochastic, or even mixed; systems operate on different time scales from msec to hours.

Level 2: Pair-wise coupling: coupling links between systems are non-linear and vary in time; moreover, as we have recently discovered, certain pairs of organ systems communicate through several forms of coupling that switch on/off and can simultaneously coexist.

Level 3: Networked interactions: global network dynamics of the entire organism are not simply the sum of the behaviors of individual systems; minor changes in the relative strength of local interactions can strongly influence emergent behaviors at the organism level, even when the topology of network interactions among these systems remains unchanged.

We have introduced innovative approaches to analyze physiologic data by adapting concepts from modern statistical physics, nonlinear dynamics and applied mathematics. We particularly tailored these approaches to extract hidden information embedded in complex non-stationary, noisy and nonlinear signals. We have successfully applied these methods to individual physiological systems, including cardiac, respiratory, locomotion, brain, considering sleep-stage transitions and circadian rhythms. These data-driven approaches enabled us to discover basic laws of physiologic regulation we published in a series of papers, including Nature-1996, PRL-1998, Nature-1999, PRL-2001, Nature-2002, PNAS-2004, PNAS-2007, PNAS-2009, PNAS-2012. Our team has also pioneered methods to investigate pair-wise coupling between physiological systems including: (i) cross-correlation of instantaneous phase increments method — we uncovered new aspects of cerebral autoregulation and its breakdown with stroke (PRL-2004, PRE-2006); (ii) cross-correlation method based on local and global detrending —to probe coupling between nonstationary stochastic signals with trends (EPJB-2009); (iii) an automated phase synchronization technique—instrumental to uncover a range of patterns of synchronous behavior between the cardiac and respiratory systems, with dramatic phase transitions across sleep stages (PNAS-2012). This experience uniquely positions us to develop new data science methodologies adequate to probe the dynamics and coupling between multiple physiological systems.