ArticlePDF Available
RESEARCH HIGHLIGHT OPEN
Moving towards precision psychiatry: the hard nut of
depression
Juergen Dukart
1,2
, Leon D. Lotter
1,2,3
and Simon B. Eickhoff
1,2
Signal Transduction and Targeted Therapy (2024) 9:310 ; https://doi.org/10.1038/s41392-024-02023-8
In a recent study published in Nature, Lynch et al.
1
reported a
consistent spatial expansion of the resting state salience network
in subjects with major depression. The expansion was already
present in children who later developed depression and remained
stable over time in patients with depression, indicating a trait-like
behavior of the observed endophenotype.
Research on major depression is symbolic of the challenges
faced in the study of most, if not all, psychiatric disorders. Whilst
various therapeutic options exist, about a third of patients with
major depression do not respond to existing treatments, and no
validated biomarkers exist for early diagnosis, prognosis, or
monitoring of the disease. Naturally, for a brain disease,
neuroimaging is often the method of choice in studies aiming
to identify biomarkers for depression with, to date, very modest
results. Brain alterations in depression were either reported to be
of small effect size and therefore not clinically meaningful, or they
consistently failed to replicate in later independent cohorts.
Similarly, neuroimaging-based machine learning models have
continuously yielded close to chance-level accuracies for differ-
entiating between major depression patients and healthy controls
in large independent cohorts.
2
Instead of looking at depression-related brain alterations using
classical region- or voxel-wise statistics, Lynch et al. adopted a
network-centered perspective comparing the topology and spatial
extent of well-established resting state connectivity patterns
between patients and healthy controls. Using this approach, the
authors report the intriguing nding of a spatially expanded
salience network with very large effect sizes observed in the
deeply-phenotyped high-density longitudinal discovery cohort.
The authors then replicate this nding in two independent
cohorts with still large, although substantially reduced, effect sizes.
The salience network is well established in the literature and likely
operates under the substantial inuence of noradrenergic and
mesolimbic dopaminergic inputs. Functionally, it is implicated in
conscious integration of autonomic feedback with internal goals
as well as social-emotional regulation.
3
These functions align well
with the symptomatology observed in depression, adding face-
validity to the ndings by Lynch et al.
The adopted topological approach also illustrates the necessity
of moving away from classical voxel- or region-wise inferential
statistics towards more complex methods such as taking into
account the spatial topology of respective neuroimaging mea-
sures. Despite the promising ndings by Lynch et al., the
biological factors that shape network topology, as well as the
pathophysiological mechanisms underlying the observed
expansion of the salience network in major depression, remain
elusive. Understanding both is essential for development of
accurate diagnostic tools and improved interventions. To address
the limited insight into the underlying pathophysiological
mechanisms, the applied topology methods can now be
combined with publicly available gene expression and neuro-
transmitter atlases as well as with other neurophysiological
modalities.
4
Integrating this multimodal information would greatly
facilitate the interpretation of the pathophysiological mechanisms
underlying the respective ndings and allow for more guided
hypothesis generation regarding potential interventions (Fig. 1).
Considering the trait-like behavior of the identied network
expansion, normative modeling may be a further promising
avenue to gain a better understanding of the variability of this
endophenotype in the general population. In this regard, follow-
up research will also have to explore the specicity of the
observed salience network expansion to major depression. Both
are important prerequisites for a biomarker to be considered for
clinical applications. For the rst, healthy control populations
recruited into clinical studies are often highly selective. Such
preselection may have narrowed the actual variability in network
expansion in the control cohort and thereby inated the observed
effect sizes. For the second, many of the endophenotypes
reported in the literature for one psychiatric disorder are often
later rediscovered in other psychiatric disorders. Understanding
the specicity of the observed expansion effects is therefore
important for assessing the actual diagnostic or predictive value of
these endophenotypes. Similarly, a crucial part of future research
with respect to the observed network expansion needs to be
directed towards understanding the high variability in clinical
symptomatology and treatment response observed in major
depression. Understanding if and how the observed expansion
aligns with potential depression subtypes and relates to treatment
effects at the individual level will be essential for moving the
ndings into clinical applications.
On a more cautious note, the authors trained a machine
learning-based classication framework using the extracted net-
work expansion features. They report rather high accuracies of
about 78% for differentiation between depression cases and
healthy controls, with salience network expansion unsurprisingly
providing the largest contribution. Despite these promising
ndings, it is important to note that the search for machine
learning biomarkers in psychiatry has been continuously ham-
pered by failures to replicate and generalize. For example, several
promising machine learning models for schizophrenia diagnosis
Received: 11 October 2024 Revised: 16 October 2024 Accepted: 18 October 2024
1
Institute of Neurosciences and Medicine, Brain & Behaviour (INM-7), Research Centre Juelich, Juelich, Germany;
2
Institute of Systems Neuroscience, Medical Faculty, Heinrich
Heine University, Düsseldorf, Germany and
3
Max Planck School of Cognition, Leipzig, Germany
Correspondence: Juergen Dukart (j.dukart@fz-juelich.de)
www.nature.com/sigtrans
Signal Transduction and Targeted Therapy
©The Author(s) 2024
1234567890();,:
initially performed well in large multi-centric datasets, but all failed
to replicate in truly independent cohorts.
5
Small sample sizes,
study-specic confounding effects such as differences in demo-
graphics or acquisition parameters, and processing and analyses
choices contributed to limited generalization of previous machine
learning models. Several of these limitations including the limited
sample size and specic processing and analysis choices certainly
also apply to the study by Lynch et al. In this respect, whilst
optimism is warranted, a prospective and independent replication
of the ndings is imperative for further pursuit into clinical
translation.
The innovative approach introduced by Lynch et al. certainly
provides a promising path to study psychiatric disorders and, more
specically, to translate neuroimaging measures into clinically
relevant biomarkers for major depression. In this regard, the
demonstrated applicability to individual patients indicates its high
potential for moving the eld towards true precision psychiatry.
ACKNOWLEDGEMENTS
J.D. and S.B.E. were supported by the Ministry for Culture and Research NRW
(Kooperationsplattformen 2022). L.D.L. was supported by the Federal Ministry of
Education and Research (BMBF) and the Max Planck Society (MPG).
AUTHOR CONTRIBUTIONS
J.D. designed and wrote the draft manuscript. L.D.L. and S.B.E. revised the manuscript.
L.D.L. prepared the gure. All authors have read and approved the article.
FUNDING
Open Access funding enabled and organized by Projekt DEAL.
ADDITIONAL INFORMATION
Competing interests: The authors declare no competing interests.
REFERENCES
1. Lynch, C. J. et al. Frontostriatal salience network expansion in individuals in
depression. Nature 633, 624633 (2024).
2. Winter, N. R. et al. A systematic evaluation of machine learningbased biomarkers
for major depressive disorder. JAMA Psychiatry 81, 386395 (2024).
3. Seeley, W. W. The salience network: a neural system for perceiving and responding
to homeostatic demands. J. Neurosci. 39, 98789882 (2019).
4. Lotter, L. D. et al. Regional patterns of human cortex development correlate with
underlying neurobiology. Nat. Commun. 15, 7987 (2024).
5. Chekroud, A. M. et al. Illusory generalizability of clinical prediction models. Science
383, 164167 (2024).
Fig. 1 Conceptual overview of the next steps for translation of the ndings by Lynch et al.
1
into clinical applications for major depression. To
enable clinical translation, the observed endophenotypes need to be integrated with other neurophysiological modalities to gain insights into
the pathomechanisms underlying the network expansion observed in major depression. Normative modeling and subtyping studies are
further required to gain insights into the variability and reliability of expansion endophenotypes in the general as well as in the major
depression population. Both needs to be combined with prospective clinical studies to validate the ndings as potential biomarkers for
diagnosis and prognosis of major depression
Moving towards precision psychiatry: the hard nut of depression
Dukart et al.
2
Signal Transduction and Targeted Therapy (2024) 9:310
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if changes were made. The images or other third party
material in this article are included in the articles Creative Commons licence, unless
indicated otherwise in a credit line to the material. If material is not included in the
articles Creative Commons licence and your intended use is not permitted by statutor y
regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this licence, visit http://
creativecommons.org/licenses/by/4.0/.
© The Author(s) 2024
Moving towards precision psychiatry: the hard nut of depression
Dukart et al.
3
Signal Transduction and Targeted Therapy (2024) 9:310
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Human brain morphology undergoes complex changes over the lifespan. Despite recent progress in tracking brain development via normative models, current knowledge of underlying biological mechanisms is highly limited. We demonstrate that human cortical thickness development and aging trajectories unfold along patterns of molecular and cellular brain organization, traceable from population-level to individual developmental trajectories. During childhood and adolescence, cortex-wide spatial distributions of dopaminergic receptors, inhibitory neurons, glial cell populations, and brain-metabolic features explain up to 50% of the variance associated with a lifespan model of regional cortical thickness trajectories. In contrast, modeled cortical thickness change patterns during adulthood are best explained by cholinergic and glutamatergic neurotransmitter receptor and transporter distributions. These relationships are supported by developmental gene expression trajectories and translate to individual longitudinal data from over 8000 adolescents, explaining up to 59% of developmental change at cohort- and 18% at single-subject level. Integrating neurobiological brain atlases with normative modeling and population neuroimaging provides a biologically meaningful path to understand brain development and aging in living humans.
Article
Full-text available
Decades of neuroimaging studies have shown modest differences in brain structure and connectivity in depression, hindering mechanistic insights or the identification of risk factors for disease onset¹. Furthermore, whereas depression is episodic, few longitudinal neuroimaging studies exist, limiting understanding of mechanisms that drive mood-state transitions. The emerging field of precision functional mapping has used densely sampled longitudinal neuroimaging data to show behaviourally meaningful differences in brain network topography and connectivity between and in healthy individuals2–4, but this approach has not been applied in depression. Here, using precision functional mapping and several samples of deeply sampled individuals, we found that the frontostriatal salience network is expanded nearly twofold in the cortex of most individuals with depression. This effect was replicable in several samples and caused primarily by network border shifts, with three distinct modes of encroachment occurring in different individuals. Salience network expansion was stable over time, unaffected by mood state and detectable in children before the onset of depression later in adolescence. Longitudinal analyses of individuals scanned up to 62 times over 1.5 years identified connectivity changes in frontostriatal circuits that tracked fluctuations in specific symptoms and predicted future anhedonia symptoms. Together, these findings identify a trait-like brain network topology that may confer risk for depression and mood-state-dependent connectivity changes in frontostriatal circuits that predict the emergence and remission of depressive symptoms over time.
Article
Full-text available
The term "salience network" refers to a suite of brain regions whose cortical hubs are the anterior cingulate and ventral anterior insular (i.e. frontoinsular) cortices. This network, which also includes nodes in the amygdala, hypothalamus, ventral striatum, thalamus, and specific brainstem nuclei, co-activates in response to diverse experimental tasks and conditions, suggesting a domain-general function. In the 12 years since its initial description, the salience network has been extensively studied, using diverse methods, concepts, and mammalian species, including healthy and diseased humans across the lifespan. Despite this large and growing body of research, the essential functions of the salience network remain uncertain. In this paper, which makes no attempt to comprehensively review this literature, I describe the circumstances surrounding the initial discovery, conceptualization, and naming of the salience network, highlighting aspects that may be unfamiliar to many readers. I then discuss some of the key advances provided by subsequent research and conclude by posing a few of the questions that remain to be explored.
Article
It is widely hoped that statistical models can improve decision-making related to medical treatments. Because of the cost and scarcity of medical outcomes data, this hope is typically based on investigators observing a model’s success in one or two datasets or clinical contexts. We scrutinized this optimism by examining how well a machine learning model performed across several independent clinical trials of antipsychotic medication for schizophrenia. Models predicted patient outcomes with high accuracy within the trial in which the model was developed but performed no better than chance when applied out-of-sample. Pooling data across trials to predict outcomes in the trial left out did not improve predictions. These results suggest that models predicting treatment outcomes in schizophrenia are highly context-dependent and may have limited generalizability.
Article
Importance Biological psychiatry aims to understand mental disorders in terms of altered neurobiological pathways. However, for one of the most prevalent and disabling mental disorders, major depressive disorder (MDD), no informative biomarkers have been identified. Objective To evaluate whether machine learning (ML) can identify a multivariate biomarker for MDD. Design, Setting, and Participants This study used data from the Marburg-Münster Affective Disorders Cohort Study, a case-control clinical neuroimaging study. Patients with acute or lifetime MDD and healthy controls aged 18 to 65 years were recruited from primary care and the general population in Münster and Marburg, Germany, from September 11, 2014, to September 26, 2018. The Münster Neuroimaging Cohort (MNC) was used as an independent partial replication sample. Data were analyzed from April 2022 to June 2023. Exposure Patients with MDD and healthy controls. Main Outcome and Measure Diagnostic classification accuracy was quantified on an individual level using an extensive ML-based multivariate approach across a comprehensive range of neuroimaging modalities, including structural and functional magnetic resonance imaging and diffusion tensor imaging as well as a polygenic risk score for depression. Results Of 1801 included participants, 1162 (64.5%) were female, and the mean (SD) age was 36.1 (13.1) years. There were a total of 856 patients with MDD (47.5%) and 945 healthy controls (52.5%). The MNC replication sample included 1198 individuals (362 with MDD [30.1%] and 836 healthy controls [69.9%]). Training and testing a total of 4 million ML models, mean (SD) accuracies for diagnostic classification ranged between 48.1% (3.6%) and 62.0% (4.8%). Integrating neuroimaging modalities and stratifying individuals based on age, sex, treatment, or remission status does not enhance model performance. Findings were replicated within study sites and also observed in structural magnetic resonance imaging within MNC. Under simulated conditions of perfect reliability, performance did not significantly improve. Analyzing model errors suggests that symptom severity could be a potential focus for identifying MDD subgroups. Conclusion and Relevance Despite the improved predictive capability of multivariate compared with univariate neuroimaging markers, no informative individual-level MDD biomarker—even under extensive ML optimization in a large sample of diagnosed patients—could be identified.