Computational Physiology and Biomedical Instruments Group, Bioengineering and Robotics Research Center 'E. Piaggio' and Department of Information Engineering, School of Engineering, University of Pisa, Largo Lucio Lazzarino 1-56122, Pisa, Italy.
Physiol Meas. 2018 Mar 29;39(3):034004. doi: 10.1088/1361-6579/aaaeac.
Depression is one of the leading causes of disability worldwide. Most previous studies have focused on major depression, and studies on subclinical depression, such as those on so-called dysphoria, have been overlooked. Indeed, dysphoria is associated with a high prevalence of somatic disorders, and a reduction of quality of life and life expectancy. In current clinical practice, dysphoria is assessed using psychometric questionnaires and structured interviews only, without taking into account objective pathophysiological indices. To address this problem, in this study we investigated heartbeat linear and nonlinear dynamics to derive objective autonomic nervous system biomarkers of dysphoria.
Sixty undergraduate students participated in the study: according to clinical evaluation, 24 of them were dysphoric. Extensive group-wise statistics was performed to characterize the pathological and control groups. Moreover, a recursive feature elimination algorithm based on a K-NN classifier was carried out for the automatic recognition of dysphoria at a single-subject level.
The results showed that the most significant group-wise differences referred to increased heartbeat complexity (particularly for fractal dimension, sample entropy and recurrence plot analysis) with regards to the healthy controls, confirming dysfunctional nonlinear sympatho-vagal dynamics in mood disorders. Furthermore, a balanced accuracy of 79.17% was achieved in automatically distinguishing dysphoric patients from controls, with the most informative power attributed to nonlinear, spectral and polyspectral quantifiers of cardiovascular variability.
This study experimentally supports the assessment of dysphoria as a defined clinical condition with specific characteristics which are different both from healthy, fully euthymic controls and from full-blown major depression.
抑郁症是全球导致残疾的主要原因之一。大多数先前的研究都集中在重度抑郁症上，而对亚临床抑郁症（如所谓的心境恶劣）的研究则被忽视了。事实上，心境恶劣与较高的躯体障碍患病率、生活质量和预期寿命降低有关。在当前的临床实践中，心境恶劣是通过心理计量问卷和结构化访谈来评估的，而没有考虑到客观的生理病理指标。为了解决这个问题，本研究调查了心跳的线性和非线性动力学，以得出心境恶劣的客观自主神经系统生物标志物。
60 名本科生参加了这项研究：根据临床评估，其中 24 名学生患有心境恶劣。进行了广泛的组间统计，以描述病理性组和对照组。此外，还基于 K-NN 分类器进行了递归特征消除算法，以实现对单个个体心境恶劣的自动识别。
结果表明，最显著的组间差异涉及到心跳复杂性的增加（特别是分形维数、样本熵和递归图分析），这与健康对照组相比，证实了情绪障碍中非线性交感神经-迷走神经功能障碍。此外，自动区分心境恶劣患者和对照组的平衡准确率达到 79.17%，其中心血管变异性的非线性、谱和多谱量化指标最具信息量。
这项研究从实验上支持了将心境恶劣作为一种具有特定特征的明确临床病症的评估，这些特征与健康、完全心境正常的对照组以及完全严重的抑郁症不同。
