Introduction

Major depressive disorder (MDD) is one of the most common psychiatric disorders with a lifetime prevalence of up to 21% [1, 2]. MDD is among the leading causes of burden of disease [3] and treatments for it have mixed efficacy. About half of the patients with MDD do not respond to first-line medication and up to a third of the patients remain treatment-resistant even after multiple therapeutic attempts [4,5,6]. This poor treatment response may be due to the large heterogeneity of symptoms that can be present in MDD [7, 8].

One of the frequent symptom domains of MDD is psychomotor disturbance (PmD). It occurs in up to 70% of patients and is associated with higher depression severity and poorer treatment response to antidepressants [9,10,11,12,13]. PmD presents in two dimensions: psychomotor retardation (PmR) and psychomotor agitation (PmA). PmR is characterized by a reduction of movement and activity, e.g., postural slum**, lower volume of voice, reduced facial expressions, or slowed gait. PmA represents an increased amount or amplitude of movements, as well as a shift toward a more erratic nature of motion, e.g., restlessness. Despite their seemingly opposite nature, PmR and PmA can occur simultaneously [14]. Moreover, PmD tends to persist in remitted patients [15]. To date, the pathophysiology of PmD remains elusive, and there are no treatments available for this specific feature of MDD.

MDD has been linked to various structural and functional brain alterations. Structural alterations encompass a global decrease of gray matter volume, particularly in cerebellum, limbic network, and multiple frontal areas [16]; as well as a cortical thinning in frontal, temporal, and limbic regions [17]. In addition, patients with MDD present with widespread lower fractional anisotropy, including the corona radiata, corpus callosum (genu and body but not splenium), external capsule, anterior limb of internal capsule, sagittal stratum, fronto-occipital fasciculus, cingulate part of the cingulum bundle, and stria terminalis [18]. Furthermore, functional alterations have been reported in intrinsic resting-state networks such as the default mode network or the salience network [19,20,21].

However, little is currently known on neural correlates of PmD in depression, specifically. Lower physical activity is linked to altered white matter integrity in motor circuits in MDD [22, 23], and to altered perfusion in orbitofrontal, rostral frontal cortex, and supplementary motor area (SMA) where higher cerebral blood flow is associated with reduced physical activity [24], while patients with MDD and PmR reportedly present lower cerebral blood flow in the primary motor cortex than patients without PmR [67]. Similarly, increased thalamo-cortical connectivity in schizophrenia was linked to hypokinetic motor abnormalities [68]. In contrast, PmA was characterized by prominent higher pallido-cortical connectivity focusing on M1 and S1. While higher thalamocortical connectivity was also present in PmA, it was restricted to parietal cortices. However, our findings argue against a simple transdiagnostic continuum in which PmR results from reduced thalamo-cortical connectivity and PmA from increased thalamo-cortical connectivity as suggested by Northoff and colleagues [66]. Albeit very useful, their model requires more testing and updating. Finally, the data of the current study may guide efforts to apply neuromodulation to aberrant motor networks in MDD. Both M1 and SMA could be valuable targets for transcranial magnetic stimulation. A prior study found that inhibitory rTMS on the SMA ameliorated PmR in a transdiagnostic sample including MDD [69].

Limitations

The present study has some limitations. As mentioned above, this study was not designed to assess PmD specifically. Therefore, a relevant number of patients with subtle PmD might have been considered as having no PmD. Our sample showed moderate total depression and PmD severity and a more severely depressed population might include a larger proportion of patients with PmD. Thus, generalizability to more severely depressed patients may be limited and our results warrant replication in such a sample. Even if the current study is providing a large overview of the brain alterations associated with PmD in MDD, the understanding of these complex conditions would benefit from adding further imaging modalities, such as structural connectivity, effective connectivity, task-based connectivity and activations. Finally, to capture the trait vs. state effects, longitudinal analyses would be required in a group of patients with PmD.

Conclusion

The present study is the first one to provide an exploration of the neural correlates of PmD in MDD using classical resting-state functional connectivity as well as network measures based on graph theory. PmA and PmR showed specific connectivity signatures within the motor network while we found evidence for compensatory increased cortico-cortical connectivity that may prevent PmD in current depression. This compensatory connectivity change disturbs network topology. In contrast, remitted patients with PmD show altered network topology of the motor network in the absence of single connection alterations, suggesting reduced adaptability of the motor network in patients with PmD.