The July 2021 Editorial inspired us to add another “Godot” to the list. As the Author pointed out1 – changing clinical guidelines in the traditionally biomarker-aversive field of psychiatry is not an easy step-indeed. That so well applies to depression research, too.
The knowledge from decades-long research in complex systems dynamics offers tools for extracting information from electrophysiological signals (ECG, EEG, etc.). These tools provide high accuracy of detection of irregularities by quantifying subtle changes in signal patterns, using nonlinear measures, like different forms of statistical entropy (Shannon entropy, approximate entropy, sample entropy, multiscale entropy, etc.) or fractal dimension measures (Higuchi fractal dimension, detrended fluctuation analysis-DFA). Nonlinear parameters, like different forms of statistical entropy or fractal measures, calculated from electrophysiological signals (e.g., EEG or ECG), are demonstrated to be predictive of many psychiatric disorders and their phases. Beside diagnostics, complex system analysis can be used for monitoring therapy results (or forecasting responders to medication or other modalities of therapy like repetitive transcranial magnetic stimulation). Based on this analytic approach it is possible not only to accurately confirm depression, but also delineate between phases of disease (episode vs remission, like in2), differentiate between subtypes (melancholic vs non-melancholic depression), comorbidities, or even detect existing suicide risk.3 Knowing those additional information early in the process can help in effectively choosing the therapy that increases the probability that the patient would recover and avoid relapses. Pincus4 stresses the importance of dynamics of the systems, which requires a quantifier that is sensitive to the order of events in time series, for example, approximate entropy (ApEn). There is a lot of research demonstrating that nonlinear measures are much more accurate and reliable than the conventional ones in analyzing history sensitive systems.5 Widely used Fourier transform that is embedded in any software in any operating machine made to record electrophysiological signals, is proven to be redundant to fractal analysis6 and it is known to be not sensitive to detect early changes in the signal unlike other fractal and nonlinear methods.7
Perhaps especially urgent is detecting cardiovascular diseases (CVD) in people suffering from depression. The connection between these two diseases that carries a high mortality risk8 has long been known9 and yet monitoring heart function in depressive patients is far from clinical routine. The data can be easily obtained by novel portable ECG monitoring devices that are approved as medical-grade signal quality equivalent to holter, but are much more practical and comfortable to use by the patient her-/himself, leading to early detection of risks and potentially to personalized medicine at its very best. The data can then be processed by a combination of nonlinear analytics and advanced statistical procedures (to control, for example, for comorbidities, subtypes and other confounding factors10). Even better, the analysis can be empowered with machine learning applications11 that are widely in use due to high power of computation and cloud computing.
This process is neither costly nor invasive, so, why wait to save lives?