Skip to main content

Big-Data-Ansätze in der Psychiatrie: Beispiele aus der Depressionsforschung

Big data approaches in psychiatry: examples in depression research



Die Erforschung und Therapie von Depression ist erschwert durch heterogene ätiologische Mechanismen und diverse Komorbiditäten. Durch den Big-Data-Trend in der Psychiatrie können Forschung und Therapie zunehmend auf den individuellen Patienten ausgerichtet werden. Diese veränderte Zielstellung erfordert spezielle Analyseverfahren.

Ziel der Arbeit

Daher sollen die Möglichkeiten und Herausforderungen von Big-Data-Ansätzen in der Depressionsforschung näher beleuchtet werden.

Material und Methode

Anhand von Beispielen werden die Möglichkeiten von Big-Data-Ansätzen in der Depressionsforschung illustriert. Dabei werden moderne maschinelle Lernverfahren mit den traditionellen statistischen Methoden hinsichtlich ihres Anwendungspotenzials in der Depressionsforschung verglichen.


Maschinelle Lernverfahren eignen sich besonders zur Analyse detaillierter Beobachtungsdaten, zur Vorhersage einzelner Datenpunkte und mehrerer klinischer Variablen sowie zur Beschreibung von Endophänotypen. Der Transfer der Ergebnisse in die klinische Praxis stellt eine aktuelle Herausforderung von Big-Data-Ansätzen in der Depressionsforschung dar.


Big-Data-Ansätze könnten ermöglichen, biologische Subtypen innerhalb der Depression zu identifizieren und Vorhersagen für einzelne Patienten zu treffen. Damit bergen sie enormes Potenzial für die Prävention, Diagnose, Therapie und Verlaufsprognose von Depressionen.



The exploration and therapy of depression is aggravated by heterogeneous etiological mechanisms and various comorbidities. With the growing trend towards big data in psychiatry, research and therapy can increasingly target the individual patient. This novel objective requires special methods of analysis.


The possibilities and challenges of the application of big data approaches in depression are examined in closer detail.

Material and methods

Examples are given to illustrate the possibilities of big data approaches in depression research. Modern machine learning methods are compared to traditional statistical methods in terms of their potential in applications to depression.


Big data approaches are particularly suited to the analysis of detailed observational data, the prediction of single data points or several clinical variables and the identification of endophenotypes. A current challenge lies in the transfer of results into the clinical treatment of patients with depression.


Big data approaches enable biological subtypes in depression to be identified and predictions in individual patients to be made. They have enormous potential for prevention, early diagnosis, treatment choice and prognosis of depression as well as for treatment development.

This is a preview of subscription content, access via your institution.


  1. 1.

    Arbabshirani MR, Plis S, Sui J et al (2017) Single subject prediction of brain disorders in neuroimaging: promises and pitfalls. Neuroimage 145:137–165

    Article  PubMed  Google Scholar 

  2. 2.

    Bengio Y, Courville A, Vincent P (2013) Representation learning: a review and new perspectives. IEEE Trans Pattern Anal Mach Intell 35:1798–1828

    Article  PubMed  Google Scholar 

  3. 3.

    Bhaumik R, Jenkins LM, Gowins JR et al (2016) Multivariate pattern analysis strategies in detection of remitted major depressive disorder using resting state functional connectivity. Neuroimage Clin 2:390–398

    Google Scholar 

  4. 4.

    Breiman L (2001) Statistical modeling: the two cultures. Stat Sci 16:199–231

    Article  Google Scholar 

  5. 5.

    Breiman L, Friedman JH (1997) Predicting multivariate responses in multiple linear regression. J R Stat Soc Series B Stat Methodol 59:3–54

    Article  Google Scholar 

  6. 6.

    Bzdok D, Yeo BTT (2017) Inference in the age of big data: future perspectives on neuroscience. Neuroimage 14:549–564

    Article  Google Scholar 

  7. 7.

    Caruana R (1998) Multitask learning. In: Thrun S, Pratt L (Hrsg) Learning to learn. Springer, Boston, S 95–133

    Chapter  Google Scholar 

  8. 8.

    Chekroud AM, Zotti RJ, Shehzad Z et al (2016) Cross-trial prediction of treatment outcome in depression: a machine learning approach. Lancet Psychiatry 3:243–250

    Article  PubMed  Google Scholar 

  9. 9.

    Cuthbert BN, Insel TR (2013) Toward the future of psychiatric diagnosis: the seven pillars of RDoC. BMC Med 11:126

    Article  PubMed  PubMed Central  Google Scholar 

  10. 10.

    Drysdale AT, Grosenick L, Downar J et al (2017) Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nat Med 23:28–38

    Article  PubMed  CAS  Google Scholar 

  11. 11.

    Eyre HA, Singh AB, Reynolds C (2016) Tech giants enter mental health. World Psychiatry 15:21–22

    Article  PubMed  PubMed Central  Google Scholar 

  12. 12.

    Gabrieli JD, Ghosh SS, Whitfield-Gabrieli S (2015) Prediction as a humanitarian and pragmatic contribution from human cognitive neuroscience. Neuron 85:11–26

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  13. 13.

    Goodfellow IJ, Bengio Y, Courville A (2016) Deep learning. MIT Press, USA

    Google Scholar 

  14. 14.

    Hastie T, Tibshirani R, Friedman J (2001) The elements of statistical learning. Springer, Heidelberg

    Book  Google Scholar 

  15. 15.

    Huys QJM, Maia TV, Frank MJ (2016) Computational psychiatry as a bridge from neuroscience to clinical applications. Nat Neurosci 19:404–413

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  16. 16.

    Insel T, Cuthbert B, Garvey M et al (2010) Research domain criteria (RDoC): toward a new classification framework for research on mental disorders. Am J Psychiatry 167:748–751

    Article  PubMed  Google Scholar 

  17. 17.

    Insel TR, Cuthbert BN (2015) Brain disorders? Precisely. Science 348:499–500

    Article  PubMed  CAS  Google Scholar 

  18. 18.

    Just MA, Pan L, Cherkassky VL et al (2017) Machine learning of neural representations of suicide and emotion concepts identifies suicidal youth. Nat Hum Behav.

    PubMed  PubMed Central  Article  Google Scholar 

  19. 19.

    Kessler RC, van Loo HM, Wardenaar KJ et al (2016) Testing a machine-learning algorithm to predict the persistence and severity of major depressive disorder from baseline self-reports. Mol Psychiatry 21:1366–1371

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  20. 20.

    Manyika J, Chui M, Brown B et al (2011) Big data: the next frontier for innovation, competition, and productivity. Technical report. McKinsey Global Institute, Düsseldorf

    Google Scholar 

  21. 21.

    Mumtaz W, Ali SSA, Yasin MAM et al (2017) A machine learning framework involving EEG-based functional connectivity to diagnose major depressive disorder (MDD). Med Biol Eng Comput.

    PubMed  Article  Google Scholar 

  22. 22.

    Østergaard SD, Jensen SOW, Bech P (2011) The heterogeneity of the depressive syndrome: when numbers get serious. Acta Psychiatr Scand 124:495–496

    Article  PubMed  Google Scholar 

  23. 23.

    Passos IC, Mwangi B, Cao B et al (2016) Identifying a clinical signature of suicidality among patients with mood disorders: a pilot study using a machine learning approach. J Affect Disord 15:109–116

    Article  Google Scholar 

  24. 24.

    Perna G, Nemeroff CB (2017) Personalized medicine in psychiatry: back to the future. Per Med Psychiatry 1:1

    Google Scholar 

  25. 25.

    Schnyer DM, Clasen PC, Gonzalez C et al (2017) Evaluating the diagnostic utility of applying a machine learning algorithm to diffusion tensor MRI measures in individuals with major depressive disorder. Psychiatry Res 30:1–9

    Article  Google Scholar 

  26. 26.

    Shalev-Shwartz S, Ben-David S (2014) Understanding machine learning: from theory to algorithms. Cambridge University Press, Cambridge

    Book  Google Scholar 

  27. 27.

    Stephan KE, Schlagenhauf F, Huys QJM et al (2017) Computational neuroimaging strategies for single patient predictions. Neuroimage 145:180–199

    Article  PubMed  CAS  Google Scholar 

  28. 28.

    Wasserstein RL, Lazar NA (2016) The ASA’s statement on p‑values: context, process, and purpose. Am Stat 70:129–133

    Article  Google Scholar 

  29. 29.

    Woo C‑W, Chang LJ, Lindquist MA et al (2017) Building better biomarkers: brain models in translational neuroimaging. Nat Neurosci 20:365–377

    Article  PubMed  PubMed Central  CAS  Google Scholar 

  30. 30.

    Woo CW, Wager TD (2015) Neuroimaging-based biomarker discovery and validation. Pain 156:1379–1381

    Article  PubMed  PubMed Central  Google Scholar 

  31. 31.

    Zhang X, Mormino EC, Sun N et al (2016) Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer’s disease. Proceedings of the National Academy of Sciences, S E6535–E6544

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to D. Bzdok.

Ethics declarations


D. Bzdok, T. M. Karrer, U. Habel und F. Schneider geben an, dass kein Interessenkonflikt besteht.

Dieser Beitrag beinhaltet keine von den Autoren durchgeführten Studien an Menschen oder Tieren. Für die aufgeführten Studien gelten die jeweils dort angegebenen ethischen Richtlinien.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bzdok, D., Karrer, T.M., Habel, U. et al. Big-Data-Ansätze in der Psychiatrie: Beispiele aus der Depressionsforschung. Nervenarzt 89, 869–874 (2018).

Download citation


  • Biologische Subtypen
  • Personalisierte Medizin
  • Verlaufsprognose
  • Maschinelles Lernen
  • Endophänotypen


  • Biological subtypes
  • Personalized medicine
  • Prognosis
  • Machine learning
  • Endophenotypes